Comparison of Random Forest, k-Nearest Neighbor, and Support Vector Machine Classifiers for Land Cover Classification Using Sentinel-2 Imagery

Analysing Process and Probability of Built-Up Expansion Using Machine Learning and Fuzzy Logic in English Bazar, West Bengal

The study sought to investigate the process of built-up expansion and the probability of built-up expansion in the English Bazar Block of West Bengal, India, using multitemporal Landsat satellite images and an integrated machine learning algorithm and fuzzy logic model. The land use and land cover (LULC) classification were prepared using a support vector machine (SVM) classifier for 2001, 2011, and 2021. The landscape fragmentation technique using the landscape fragmentation tool (extension for ArcGIS software) and frequency approach were proposed to model the process of built-up expansion. To create the built-up expansion probability model, the dominance, diversity, and connectivity index of the built-up areas for each year were created and then integrated with fuzzy logic. The results showed that, during 2001–2021, the built-up areas increased by 21.67%, while vegetation and water bodies decreased by 9.28 and 4.63%, respectively. The accuracy of the LULC maps for 2001, 2011, and 2021 was 90.05, 93.67, and 96.24%, respectively. According to the built-up expansion model, 9.62% of the new built-up areas was created in recent decades. The built-up expansion probability model predicted that 21.46% of regions would be converted into built-up areas. This study will assist decision-makers in proposing management strategies for systematic urban growth that do not damage the environment.

Integrating the Strength of Multi-Date Sentinel-1 and -2 Datasets for Detecting Mango (Mangifera indica L.) Orchards in a Semi-Arid Environment in Zimbabwe

Generating tree-specific crop maps within heterogeneous landscapes requires imagery of fine spatial and temporal resolutions to discriminate among the rapid transitions in tree phenological and spectral features. The availability of freely accessible satellite data of relatively high spatial and temporal resolutions offers an unprecedented opportunity for wide-area land use and land cover (LULC) mapping, including tree crop (e.g., mango; Mangifera indica L.) detection. We evaluated the utility of combining Sentinel-1 (S1) and Sentinel-2 (S2) derived variables (n = 81) for mapping mango orchard occurrence in Zimbabwe using machine learning classifiers, i.e., support vector machine and random forest. Field data were collected on mango orchards and other LULC classes. Fewer variables were selected from ‘All’ combined S1 and S2 variables using three commonly utilized variable selection methods, i.e., relief filter, guided regularized random forest, and variance inflation factor. Several classification experiments (n = 8) were conducted using 60% of field datasets and combinations of ‘All’ and fewer selected variables and were compared using the remaining 40% of the field dataset and the area underclass approach. The results showed that a combination of random forest and relief filter selected variables outperformed (F1 score > 70%) all other variable combination experiments. Notwithstanding, the differences among the mapping results were not significant (p ≤ 0.05). Specifically, the mapping accuracy of the mango orchards was more than 80% for each of the eight classification experiments. Results revealed that mango orchards occupied approximately 18% of the spatial extent of the study area. The S1 variables were constantly selected compared with the S2-derived variables across the three variable selection approaches used in this study. It is concluded that the use of multi-modal satellite imagery and robust machine learning classifiers can accurately detect mango orchards and other LULC classes in semi-arid environments. The results can be used for guiding and upscaling biological control options for managing mango insect pests such as the devastating invasive fruit fly Bactrocera dorsalis (Hendel) (Diptera: Tephritidae).

Large-Scale Surface Water Mapping Based on Landsat and Sentinel-1 Images

Surface water is a highly dynamical object on the earth’s surface. At present, satellite remote sensing is the most effective way to accurately depict the temporal and spatial variation characteristics of surface water on a large scale. In this study, a region-adaptive random forest algorithm is designed on the Google Earth Engine (GEE) for automatic surface water mapping by using data from multi-sensors such as Landsat 7 ETM+, Landsat 8 OLI, and Sentinel-1 SAR images as source data, and China as a case study region. The visual comparison of the mapping results with the original images under different landform areas shows that the extracted water body boundary is consistent with the water range in the image. The cross-validation with the JRC GSW validation samples shows a very high precision that the average producer’s accuracy and average user’s accuracy of water is 0.933 and 0.998, respectively. The average overall accuracy and average kappa is 0.966 and 0.931, respectively. The independent verification results of lakes with different areas also prove the high accuracy for our method, with a maximum average error of 3.299%. These results show that the method is an ideal way for large-scale surface water mapping with a high spatial–temporal resolution.

Monitoring Annual Land Use/Land Cover Change in the Tucson Metropolitan Area with Google Earth Engine (1986–2020)

The Tucson metropolitan area, located in the Sonoran Desert of southeastern Arizona (USA), is affected by both massive population growth and rapid climate change, resulting in important land use and land cover (LULC) changes. As its fragile arid ecosystem and scarce resources are increasingly under pressure, there is a crucial need to monitor such landscape transformations. For such ends, we propose a method to compute yearly 30 m resolution LULC maps of the region from 1986 to 2020, using a combination of Landsat imagery, derived transformation and indices, texture analysis and other ancillary data fed to a Random Forest classifier. The entire process was hosted in the Google Earth Engine with tremendous computing capacities that allowed us to process a large amount of data and to achieve high overall classification accuracy for each year, ranging from 86.7 to 96.3%. Conservative post-processing techniques were also used to mitigate the persistent confusions between the numerous isolated houses in the region and their desert surroundings and to smooth year-specific LULC changes in order to identify general trends. We then show that policies to lessen urban sprawl in the area had little effects and we provide an automated tool to continue monitoring such dynamics in the future.

Land Use and Land Cover Mapping Using Sentinel-2, Landsat-8 Satellite Images, and Google Earth Engine: A Comparison of Two Composition Methods

Accurate and real-time land use/land cover (LULC) maps are important to provide precise information for dynamic monitoring, planning, and management of the Earth. With the advent of cloud computing platforms, time series feature extraction techniques, and machine learning classifiers, new opportunities are arising in more accurate and large-scale LULC mapping. In this study, we aimed at finding out how two composition methods and spectral–temporal metrics extracted from satellite time series can affect the ability of a machine learning classifier to produce accurate LULC maps. We used the Google Earth Engine (GEE) cloud computing platform to create cloud-free Sentinel-2 (S-2) and Landsat-8 (L-8) time series over the Tehran Province (Iran) as of 2020. Two composition methods, namely, seasonal composites and percentiles metrics, were used to define four datasets based on satellite time series, vegetation indices, and topographic layers. The random forest classifier was used in LULC classification and for identifying the most important variables. Accuracy assessment results showed that the S-2 outperformed the L-8 spectral–temporal metrics at the overall and class level. Moreover, the comparison of composition methods indicated that seasonal composites outperformed percentile metrics in both S-2 and L-8 time series. At the class level, the improved performance of seasonal composites was related to their ability to provide better information about the phenological variation of different LULC classes. Finally, we conclude that this methodology can produce LULC maps based on cloud computing GEE in an accurate and fast way and can be used in large-scale LULC mapping.

Earth Observation Mission of a 6U CubeSat with a 5-Meter Resolution for Wildfire Image Classification Using Convolution Neural Network Approach

The KITSUNE satellite is a 6-unit CubeSat platform with the main mission of 5-m-class Earth observation in low Earth orbit (LEO), and the payload is developed with a 31.4 MP commercial off-the-shelf sensor, customized optics, and a camera controller board. Even though the payload is designed for Earth observation and to capture man-made patterns on the ground as the main mission, a secondary mission is planned for the classification of wildfire images by the convolution neural network (CNN) approach. Therefore, KITSUNE will be the first CubeSat to employ CNN to classify wildfire images in LEO. In this study, a deep-learning approach is utilized onboard the satellite in order to reduce the downlink data by pre-processing instead of the traditional method of performing the image processing at the ground station. The pre-trained CNN models generated in Colab are saved in RPi CM3+, in which, an uplink command will execute the image classification algorithm and append the results on the captured image data. The on-ground testing indicated that it could achieve an overall accuracy of 98% and an F1 score of a 97% success rate in classifying the wildfire events running on the satellite system using the MiniVGGNet network. Meanwhile, the LeNet and ShallowNet models were also compared and implemented on the CubeSat with 95% and 92% F1 scores, respectively. Overall, this study demonstrated the capability of small satellites to perform CNN onboard in orbit. Finally, the KITSUNE satellite is deployed from ISS on March 2022.

Mortality Prediction of COVID-19 Patients Using Radiomic and Neural Network Features Extracted from a Wide Chest X-ray Sample Size: A Robust Approach for Different Medical Imbalanced Scenarios

Aim: The aim of this study was to develop robust prognostic models for mortality prediction of COVID-19 patients, applicable to different sets of real scenarios, using radiomic and neural network features extracted from chest X-rays (CXRs) with a certified and commercially available software. Methods: 1816 patients from 5 different hospitals in the Province of Reggio Emilia were included in the study. Overall, 201 radiomic features and 16 neural network features were extracted from each COVID-19 patient’s radiography. The initial dataset was balanced to train the classifiers with the same number of dead and survived patients, randomly selected. The pipeline had three main parts: balancing procedure; three-step feature selection; and mortality prediction with radiomic features through three machine learning (ML) classification models: AdaBoost (ADA), Quadratic Discriminant Analysis (QDA) and Random Forest (RF). Five evaluation metrics were computed on the test samples. The performance for death prediction was validated on both a balanced dataset (Case 1) and an imbalanced dataset (Case 2). Results: accuracy (ACC), area under the ROC-curve (AUC) and sensitivity (SENS) for the best classifier were, respectively, 0.72 ± 0.01, 0.82 ± 0.02 and 0.84 ± 0.04 for Case 1 and 0.70 ± 0.04, 0.79 ± 0.03 and 0.76 ± 0.06 for Case 2. These results show that the prediction of COVID-19 mortality is robust in a different set of scenarios. Conclusions: Our large and varied dataset made it possible to train ML algorithms to predict COVID-19 mortality using radiomic and neural network features of CXRs.

MFDNN: multi-channel feature deep neural network algorithm to identify COVID19 chest X-ray images

The use of chest X-ray images (CXI) to detect Severe Acute Respiratory Syndrome Coronavirus 2 (SARS CoV-2) caused by Coronavirus Disease 2019 (COVID19) is life-saving important for both patients and doctors. This research proposes a multi-channel feature deep neural network (MFDNN) algorithm to screen people infected with COVID19. The algorithm integrates data over-sampling technology and MFDNN model to carry out the training. The oversampling technique reduces the deviation of the prior probability of the MFDNN algorithm on unbalanced data. Multi-channel feature fusion technology improves the efficiency of feature extraction and the accuracy of model diagnosis. In the experiment, Compared with traditional deep learning models (VGG19, GoogLeNet, Resnet50, Desnet201), the MFDNN model obtains an average test accuracy of 93.19% in all data. Furthermore, in each type of screening, the precision, recall, and F1 Score of the MFDNN model are also better than traditional deep learning networks. Furthermore, through ablation experiments, we proved that a multi-channel convolutional neural network (CNN) is superior to single-channel CNN, additional layer and PSN module, and indirectly proved the sufficiency and necessity of each step of the MFDNN classification method. Finally, our experimental code will be placed at https://github.com/panliangrui/covid19.

Semi-automatic classification for rapid delineation of the geohazard-prone areas using Sentinel-2 satellite imagery

Abstract

The study of land use land cover has become increasingly significant with the availability of remote sensing data. The main objective of this study is to delineate geohazard-prone areas using semi-automatic classification technique and Sentinel-2 satellite imagery in Bhutan. An open-source, semi-automatic classification plugin tools in QGIS software enabled efficient and rapid conduct of land cover classification. Band sets 2-8, 8A, and 11-12 are utilized and the virtual colour composites have been used for the clustering and creation of training samples or regions of interest. An iterative self-organizing data analysis technique is used for clustering and the image is classified by a minimum distance algorithm in the unsupervised classification. The Random Forest (RF) classifier is used for the supervised classification. The unsupervised classification shows an overall accuracy of 85.47% (Kappa coefficient = 0.71) and the RF classifier resulted in an accuracy of 92.62% (Kappa coefficient = 0.86). A comparison of the classification shows a higher overall accuracy of the RF classifier with an improvement of 7.15%. The study highlights 35.59% (512,100 m2) of the study area under the geohazard-prone area. The study also overlaid the major landslide polygons to roughly validate the landslide hazards.

Article highlights

Monitoring Complex Integrated Crop–Livestock Systems at Regional Scale in Brazil: A Big Earth Observation Data Approach

Due to different combinations of agriculture, livestock and forestry managed by rotation, succession and intercropping practices, integrated agriculture production systems such as integrated crop–livestock systems (iCL) constitute a very complex target and a challenge for automatic mapping of cropping practices based on remote sensing data. The overall objective of this study was to develop a classification strategy for the annual mapping of integrated Crop–Livestock systems (iCL) at a regional scale. This strategy was designed and tested in the six agro-climatic regions of Mato Grosso, the largest Brazilian soybean producer state, using MODIS satellite time-series images acquired between 2012 and 2019, ground data with heterogeneous distribution in space and time and a Random Forest classifier. The results showed that: 1. the use of unbalanced training samples with a class composition close to the real one was the right classifier training strategy; 2. the use of a single training database (pooling samples from different years and regions) to classify each region and year individually proved to be robust enough to provide similar classification accuracies in comparison to those based on the use of a database acquired for each region and for each year. The final hierarchical classification overall accuracy was 0.89 for Level 1, the cropping pattern level (single and double crops DC); 0.84 for Level 2, the DC category level (integrated system iCL soy-pasture/brachiaria, soy-cotton and soy-cereal); 0.77 for Level 3, the iCL level (iCL1 soy-pasture and iCL2 soy-pasture mixed with corn). The F-scores for DC, iCL and iCL1 cropping systems presented high accuracy (0.89, 0.85 and 0.84), while iCL2 was more difficult to classify (0.63). This approach will next be applied across the entire Brazilian soybean corridor, leading to an operational tool for monitoring the adoption of sustainable intensification practices recognized by Brazil’s Agriculture Low Carbon Plan (ABC PLAN).

Alterations in common marmoset gut microbiome associated with duodenal strictures

Chronic gastrointestinal (GI) diseases are the most common diseases in captive common marmosets (Callithrix jacchus). Despite standardized housing, diet and husbandry, a recently described gastrointestinal syndrome characterized by duodenal ulcers and strictures was observed in a subset of marmosets sourced from the New England Primate Research Center. As changes in the gut microbiome have been associated with GI diseases, the gut microbiome of 52 healthy, non-stricture marmosets (153 samples) were compared to the gut microbiome of 21 captive marmosets diagnosed with a duodenal ulcer/stricture (57 samples). No significant changes were observed using alpha diversity metrics, and while the community structure was significantly different when comparing beta diversity between healthy and stricture cases, the results were inconclusive due to differences observed in the dispersion of both datasets. Differences in the abundance of individual taxa using ANCOM, as stricture-associated dysbiosis was characterized by Anaerobiospirillum loss and Clostridium perfringens increases. To identify microbial and serum biomarkers that could help classify stricture cases, we developed models using machine learning algorithms (random forest, classification and regression trees, support vector machines and k-nearest neighbors) to classify microbiome, serum chemistry or complete blood count (CBC) data. Random forest (RF) models were the most accurate models and correctly classified strictures using either 9 ASVs (amplicon sequence variants), 4 serum chemistry tests or 6 CBC tests. Based on the RF model and ANCOM results, C. perfringens was identified as a potential causative agent associated with the development of strictures. Clostridium perfringens was also isolated by microbiological culture in 4 of 9 duodenum samples from marmosets with histologically confirmed strictures. Due to the enrichment of C. perfringens in situ, we analyzed frozen duodenal tissues using both 16S microbiome profiling and RNAseq. Microbiome analysis of the duodenal tissues of 29 marmosets from the MIT colony confirmed an increased abundance of Clostridium in stricture cases. Comparison of the duodenal gene expression from stricture and non-stricture marmosets found enrichment of genes associated with intestinal absorption, and lipid metabolism, localization, and transport in stricture cases. Using machine learning, we identified increased abundance of C. perfringens, as a potential causative agent of GI disease and intestinal strictures in marmosets.

Startups and Consumer Purchase Behavior: Application of Support Vector Machine Algorithm

This study evaluated the impact of startup technology innovations and customer relationship management (CRM) performance on customer participation, value co-creation, and consumer purchase behavior (CPB). This analytical study empirically tested the proposed hypotheses using structural equation modeling (SEM) and SmartPLS 3 techniques. Moreover, we used a support vector machine (SVM) algorithm to verify the model’s accuracy. SVM algorithm uses four different kernels to check the accuracy criterion, and we checked all of them. This research used the convenience sampling approach in gathering the data. We used the conventional bias test method. A total of 466 respondents were completed. Technological innovations of startups and CRM have a positive and significant effect on customer participation. Customer participation significantly affects the value of pleasure, economic value, and relationship value. Based on the importance-performance map analysis (IPMA) matrix results, “customer participation” with a score of 0.782 had the highest importance. If customers increase their participation performance by one unit during the COVID-19 epidemic, its overall CPB increases by 0.782. In addition, our results showed that the lowest performance is related to the technological innovations of startups, which indicates an excellent opportunity for development in this area. SVM results showed that polynomial kernel, to a high degree, is the best kernel that confirms the model’s accuracy.

Mapping Land Cover Types for Highland Andean Ecosystems in Peru Using Google Earth Engine

Highland Andean ecosystems sustain high levels of floral and faunal biodiversity in areas with diverse topography and provide varied ecosystem services, including the supply of water to cities and downstream agricultural valleys. Google (™) has developed a product specifically designed for mapping purposes (Earth Engine), which enables users to harness the computing power of a cloud-based solution in near-real time for land cover change mapping and monitoring. We explore the feasibility of using this platform for mapping land cover types in topographically complex terrain with highly mixed vegetation types (Nor Yauyos Cochas Landscape Reserve located in the central Andes of Peru) using classification machine learning (ML) algorithms in combination with different sets of remote sensing data. The algorithms were trained using 3601 sampling pixels of (a) normalized spectral bands between the visible and near infrared spectrum of the Landsat 8 OLI sensor for the 2018 period, (b) spectral indices of vegetation, soil, water, snow, burned areas and bare ground and (c) topographic-derived indices (elevation, slope and aspect). Six ML algorithms were tested, including CART, random forest, gradient tree boosting, minimum distance, naïve Bayes and support vector machine. The results reveal that ML algorithms produce accurate classifications when spectral bands are used in conjunction with topographic indices, resulting in better discrimination among classes with similar spectral signatures such as pajonal (tussock grass-dominated cover) and short grasses or rocky groups, and moraines, agricultural and forested areas. The model with the highest explanatory power was obtained from the combination of spectral bands and topographic indices using the random forest algorithm (Kappa = 0.81). Our study presents a first approach of its kind in topographically complex Cordilleran terrain and we show that GEE is particularly useful in large-scale land cover mapping and monitoring in mountainous ecosystems subject to rapid changes and conversions, with replicability and scalability to other areas with similar characteristics.

Modeling Forest Canopy Cover: A Synergistic Use of Sentinel-2, Aerial Photogrammetry Data, and Machine Learning

Forest canopy cover (FCC) is an important ecological parameter of forest ecosystems, and is correlated with forest characteristics, including plant growth, regeneration, biodiversity, light regimes, and hydrological properties. Here, we present an approach of combining Sentinel-2 data, high-resolution aerial images, and machine learning (ML) algorithms to model FCC in the Hyrcanian mixed temperate forest, Northern Iran. Sentinel-2 multispectral bands and vegetation indices were used as variables for modeling and mapping FCC based on UAV ground truth to a wider spatial extent. Random forest (RF), support-vector machine (SVM), elastic net (ENET), and extreme gradient boosting (XGBoost) were the ML algorithms used to learn and generalize on the remotely sensed variables. Evaluation of variable importance indicated that vegetation indices including NDVI, NDVI-A, NDRE, and NDI45 were the dominant predictors in most of the models. Model accuracy estimation results showed that among the tested models, RF (R2 = 0.67, RMSE = 18.87%, MAE = 15.35%) and ENET (R2 = 0.63, RMSE = 20.04%, MAE = 16.44%) showed the best and the worst performance, respectively. In conclusion, it was possible to prove the suitability of integrating UAV-obtained RGB images, Sentinel-2 data, and ML models for the estimation of FCC, intended for precise and fast mapping at landscape-level scale.

An Object-Based Genetic Programming Approach for Cropland Field Extraction

Cropland fields are the basic spatial units for agricultural management, and information about their distribution is critical for analyzing agricultural investments and management. However, the extraction of cropland fields of smallholder farms is a challenging task because of their irregular shapes and diverse spectrum. In this paper, we proposed a new object-based Genetic Programming (GP) approach to extract cropland fields. The proposed approach used the multiresolution segmentation (MRS) method to acquire objects from a very high resolution (VHR) image, and extracted spectral, shape and texture features as inputs for GP. Then GP was used to automatically evolve the optimal classifier to extract cropland fields. The results show that the proposed approach has obtained high accuracy in two areas with different landscape complexities. Further analysis show that the GP approach significantly outperforms five commonly used classifiers, including K-Nearest Neighbor (KNN), Decision Tree (DT), Naïve Bayes (NB), Support Vector Machine (SVM), and Random Forest (RF). By using different numbers of training samples, GP can maintain high accuracy with any volume of samples compared to other classifiers.

Can a Hierarchical Classification of Sentinel-2 Data Improve Land Cover Mapping?

Monitoring of land cover plays an important role in effective environmental management, assessment of natural resources, environmental protection, urban planning and sustainable development. Increasing demand for accurate and repeatable information on land cover and land cover changes causes rapid development of the advanced, machine learning algorithms dedicated to land cover mapping using satellite images. Free and open access to Sentinel-2 data, characterized with high spatial and temporal resolution, increased the potential to map and to monitor land surface with high accuracy and frequency. Despite a considerable number of approaches towards land cover classification based on satellite data, there is still a challenge to clearly separate complex land cover classes, for example grasslands, arable land and wetlands. The aim of this study is to examine, whether a hierarchal classification of Sentinel-2 data can improve the accuracy of land cover mapping and delineation of complex land cover classes. The study is conducted in the Lodz Province, in central Poland. The pixel-based land cover classification is carried out using the machine learning Random Forest (RF) algorithm, based on a time series of Sentinel-2 imagery acquired in 2020. The following nine land cover classes are mapped: sealed surfaces, woodland broadleaved, woodland coniferous, shrubs, permanent herbaceous (grassy cover), periodically herbaceous (i.e., arable land), mosses, non-vegetated (bare soil) and water bodies. The land cover classification is conducted following two approaches: (1) flat, where all land cover classes are classified together, and (2) hierarchical, where the stratification is applied to first separate the most stable land cover classes and then classifying the most problematic once. The national databases served as the source of the reference sampling plots for the classification process. The process of selection and verification of the reference sampling plots is performed automatically. To assess the stability of the classification models the classification processes are performed iteratively. The results of this study confirmed that the hierarchical approach gave more accurate results compared to the commonly used flat approach. The median of the overall accuracy (OA) of the hierarchical classification was higher by 3–9 percentage points compared to the flat one. Of interest, the OA of the hierarchical classification reached 0.93–0.99, whereas the flat approach reached 0.90. Individual classes are also better classified in the hierarchical approach.

A wildfire growth prediction and evaluation approach using Landsat and MODIS data

The increasing wildfire occurrence due to global climate changes urged the improvement of present wildfire growth prediction and evaluation methods. This study aimed to propose novel solutions to their two primary limitations, including the lack of robust fuel classification method and the low spatial resolution of wildfire growth accuracy assessment while ensuring wide applicability using open data satellite missions and software. The first objective was to create a robust two-step fuel model classification method consisted of the supervised machine learning classification of generalized land cover classes in the 1st level and their individual unsupervised classification to vegetation subtypes in the 2nd level. The second objective was creating a wildfire prediction accuracy assessment method using MODIS 250 m images, which overcome the limitations of low spatial resolution while preserving sub-daily temporal resolution. The wildfire on the Korčula island in Croatia was analyzed in the study, being specific for its long duration from 18 to 24 July 2015. The wildfire ignition occurred in the isolated area, which prolonged the response time from emergency agencies. Random Forest (RF) with input Landsat 8 spectral bands and indices resulted in the highest classification accuracy in the 1st classification level with an overall agreement of 83.6%. The vegetation subclasses from the 2nd classification level were matched to the 13 standard fuel models for the input in FARSITE software. The predicted wildfire evaluation showed the highest mean accuracy of 0.906 for the first two days, which decreased to 0.722 in the latter stages of the active wildfire caused by overprediction. The proposed two-step fuel model classification presented a cost-efficient solution to the fuel map creation in any part of the world, with a disadvantage of no in-situ ground truth identification and accuracy assessment for 2nd classification level. The evaluation of wildfire growth prediction with 250 m images enabled high spatial and temporal resolution of the assessment, while its limitations of wildfire overprediction and the negative effects of wildfire smoke in MODIS images should be addressed in future research.

The Accuracy of Winter Wheat Identification at Different Growth Stages Using Remote Sensing

The aim of this study was to explore the differences in the accuracy of winter wheat identification using remote sensing data at different growth stages using the same methods. Part of northern Henan Province, China was taken as the study area, and the winter wheat growth cycle was divided into five periods (seeding‒tillering, overwintering, reviving, jointing‒heading, and flowering‒maturing) based on monitoring data obtained from agrometeorological stations. With the help of the Google Earth Engine (GEE) platform, the separability between winter wheat and other land cover types was analyzed and compared using the Jeffries‒Matusita (J‒M) distance method. Spectral features, vegetation index, water index, building index, texture features, and terrain features were generated from Sentinel-2 remote sensing images at different growth periods, and then were used to establish a random forest classification and extraction model. A deep U-Net semantic segmentation model based on the red, green, blue, and near-infrared bands of Sentinel-2 imagery was also established. By combining models with field data, the identification of winter wheat was carried out and the difference between the accuracy of the identification in the five growth periods was analyzed. The experimental results show that, using the random forest classification method, the best separability between winter wheat and the other land cover types was achieved during the jointing‒heading period: the overall identification accuracy for the winter wheat was then highest at 96.90% and the kappa coefficient was 0.96. Using the deep-learning classification method, it was also found that the semantic segmentation accuracy of winter wheat and the model performance were best during the jointing‒heading period: a precision, recall, F1 score, accuracy, and IoU of 0.94, 0.93, 0.93, and 0.88, respectively, were achieved for this period. Based on municipal statistical data for winter wheat, the accuracy of the extraction of the winter wheat area using the two methods was 96.72% and 88.44%, respectively. Both methods show that the jointing‒heading period is the best period for identifying winter wheat using remote sensing and that the identification made during this period is reliable. The results of this study provide a scientific basis for accurately obtaining the area planted with winter wheat and for further studies into winter wheat growth monitoring and yield estimation.

Land Cover Changes in Selected Areas Next to Lagoons Located on the Southern Coast of the Baltic Sea, 1984–2021

The aim of the study is the evaluation of land cover changes in selected areas next to three lagoons (the Curonian Lagoon, the Vistula Lagoon and the Szczecin Lagoon) located on the southern coast of the Baltic Sea (in Lithuania, Russia, Poland and Germany) from 1984 to 2021. The changes are evaluated using multispectral (visible light—RGB and near infrared—NIR) satellite images from the Landsat 5 and Sentinel-2 sensors. Due to their high importance for ecosystem services, two main land cover types are evaluated, i.e., forest area and inland water reservoirs. The classification of the images is performed using a random forest algorithm. Areas of water bodies and forests are evaluated for the years 1984 and 2021. During period 1984–2021, positive changes in land cover are observed in all three regions included in the study. In almost all parts, with the exception of the Polish part of the area located next to the Szczecin Lagoon, of these regions, an increase in forest area is observed. The increase ranges from 0.1% (Poland, area next to the Vistula Lagoon) to 1.2% (Germany, area next to the Szczecin Lagoon). The area of inland water reservoirs has not changed significantly in the long term. Despite the global warming, no reduction in the area of these water reservoirs is observed, even new seminatural reservoirs have been created in some parts of the study area.

Cotton Classification Method at the County Scale Based on Multi-Features and Random Forest Feature Selection Algorithm and Classifier

Accurate cotton maps are crucial for monitoring cotton growth and precision management. The paper proposed a county-scale cotton mapping method by using random forest (RF) feature selection algorithm and classifier based on selecting multi-features, including spectral, vegetation indices, and texture features. The contribution of texture features to cotton classification accuracy was also explored in addition to spectral features and vegetation index. In addition, the optimal classification time, feature importance, and the best classifier on the cotton extraction accuracy were evaluated. The results showed that the texture feature named the gray level co-occurrence matrix (GLCM) is effective for improving classification accuracy, ranking second in contribution among all studied spectral, VI, and texture features. Among the three classifiers, the RF showed higher accuracy and better stability than support vector machines (SVM) and artificial neural networks (ANN). The average overall accuracy (OA) of the classification combining multiple features was 93.36%, 7.33% higher than the average OA of the single-time spectrum, and 2.05% higher than the average OA of the multi-time spectrum. The classification accuracy after feature selection by RF can still reach 92.12%, showing high accuracy and efficiency. Combining multiple features and random forest methods may be a promising county-scale cotton classification method.

SVM optimization using a grid search algorithm to identify robusta coffee bean images based on circularity and eccentricity

Coffee variety is one of the main factors affecting the quality and price of coffee, so it is important to recognize coffee varieties. This study aims to optimize the recognition of robusta coffee beans based on circularity and eccentricity image features using a support vector machine (SVM) and Grid search algorithm. The methods used included image acquisition, preprocessing, feature extraction, classification, and evaluation. Circularity and eccentricity are used in the feature extraction process, while the grid search algorithm is used to optimize SVM parameters in the classification process for four different kernels. This study produced the best classification model with the highest accuracy of 94% for the RBF and Polynomial kernels.

Monitoring Urban Expansion and Loss of Agriculture on the North Coast of West Java Province, Indonesia, Using Google Earth Engine and Intensity Analysis

Uncontrolled urban expansion resulting from urbanization has a disastrous impact on agricultural land. This situation is being experienced by the densely populated and fertile island Java in Indonesia. Remote sensing technologies have developed rapidly in recent years, including the creation of Google Earth Engine (GEE). Intensity analysis (IA) is increasingly being used to systematically and substantially analyze land-use/land-cover (LULC) change. As yet, however, no study of land conversion from agriculture to urban areas in Indonesia has adopted GEE and IA approaches simultaneously. Therefore, this study aims to monitor urban penetration to agricultural land in the north coastal region of West Java Province by applying both methods to two time intervals: 2003–2013 and 2013–2020. Landsat data and a robust random forest (RF) classifier available in GEE were chosen for producing LULC maps. Monitoring LULC change using GEE and IA has demonstrated reliable findings. The overall accuracy of Landsat image classification results for 2003, 2013, and 2020 were 88%, 87%, and 88%, respectively. IA outputs at interval levels for all categories showed that the annual change-of-area rate was higher during 2013–2020 than during 2003–2013. At the category level, IA results showed that the area of agricultural land experienced net losses in both periods, with net loss in 2013–2020 being 2.3 times greater than that in 2003–2013 (∼1,850 ha per year). In contrast, the built-up area made net gains in both periods, reaching almost twice as much in the second period as in the first (∼2,030 ha per year). The transition-level IA performed proved that agricultural land had been the primary target for the expansion of built-up areas. The most extensive spatial distribution of land conversion from agriculture to built-up area was concentrated in the regencies of Bekasi, Karawang, and Cirebon. These findings are intended to provide stakeholders with enrichment in terms of available literature and with valuable inputs useful for identifying better urban and regional planning policies in Indonesia and similar regions.

Comparison of Random Forest and Support Vector Machine Classifiers for Regional Land Cover Mapping Using Coarse Resolution FY-3C Images

The type of algorithm employed to classify remote sensing imageries plays a great role in affecting the accuracy. In recent decades, machine learning (ML) has received great attention due to its robustness in remote sensing image classification. In this regard, random forest (RF) and support vector machine (SVM) are two of the most widely used ML algorithms to generate land cover (LC) maps from satellite imageries. Although several comparisons have been conducted between these two algorithms, the findings are contradicting. Moreover, the comparisons were made on local-scale LC map generation either from high or medium resolution images using various software, but not Python. In this paper, we compared the performance of these two algorithms for large area LC mapping of parts of Africa using coarse resolution imageries in the Python platform by the employing Scikit-Learn (sklearn) library. We employed a big dataset, 297 metrics, comprised of systematically selected 9-month composite FegnYun-3C (FY-3C) satellite images with 1 km resolution. Several experiments were performed using a range of values to determine the best values for the two most important parameters of each classifier, the number of trees and the number of variables, for RF, and penalty value and gamma for SVM, and to obtain the best model of each algorithm. Our results showed that RF outperformed SVM yielding 0.86 (OA) and 0.83 (k), which are 1–2% and 3% higher than the best SVM model, respectively. In addition, RF performed better in mixed class classification; however, it performed almost the same when classifying relatively pure classes with distinct spectral variation, i.e., consisting of less mixed pixels. Furthermore, RF is more efficient in handling large input datasets where the SVM fails. Hence, RF is a more robust ML algorithm especially for heterogeneous large area mapping using coarse resolution images. Finally, default parameter values in the sklearn library work well for satellite image classification with minor/or no adjustment for these algorithms.

Designing a European-Wide Crop Type Mapping Approach Based on Machine Learning Algorithms Using LUCAS Field Survey and Sentinel-2 Data

One of the most challenging aspects of obtaining detailed and accurate land-use and land-cover (LULC) maps is the availability of representative field data for training and validation. In this manuscript, we evaluate the use of the Eurostat Land Use and Coverage Area frame Survey (LUCAS) 2018 data to generate a detailed LULC map with 19 crop type classes and two broad categories for woodland and shrubland, and grassland. The field data were used in combination with Copernicus Sentinel-2 (S2) satellite data covering Europe. First, spatially and temporally consistent S2 image composites of (1) spectral reflectances, (2) a selection of spectral indices, and (3) several bio-geophysical indicators were created for the year 2018. From the large number of features, the most important were selected for classification using two machine-learning algorithms (support vector machine and random forest). Results indicated that the 19 crop type classes and the two broad categories could be classified with an overall accuracy (OA) of 77.6%, using independent data for validation. Our analysis of three methods to select optimum training data showed that by selecting the most spectrally different pixels for training data, the best OA could be achieved, and this already using only 11% of the total training data. Comparing our results to a similar study using Sentinel-1 (S1) data indicated that S2 can achieve slightly better results, although the spatial coverage was slightly reduced due to gaps in S2 data. Further analysis is ongoing to leverage synergies between optical and microwave data.

Prediction of AI-Based Personal Thermal Comfort in a Car Using Machine-Learning Algorithm

Defining a passenger’s thermal comfort in a car cabin is difficult because of the narrow environment and various parameters. Although passenger comfort is predicted using a thermal-comfort scale in the overall cabin or a local area, the scale’s range of passenger comfort may differ owing to psychological factors and individual preferences. Among the many factors affecting such comfort levels, the temperature of the seat is one of the direct and significant environmental factors. Therefore, it is necessary to predict the cabin environment and seat-related personal thermal comfort. Accordingly, machine learning is used in this research to predict whether a passenger’s seat-heating-operation pattern can be predicted in a winter environment. The experiment measures the ambient factor and collects data on passenger heating-operation patterns using a device in an actual winter environment. The temperature is set as the input parameter in the measured data and the operation pattern is used as the output parameter. Based on the parameters, the predictive accuracy of the heating-operation pattern is investigated using machine learning. The algorithms used in the machine-learning train are Tree, SVM, and kNN. In addition, the predictive accuracy is tested using SVM and kNN, which shows a high validation accuracy based on the prediction results of the algorithm. In this research, the parameters predicting the personal thermal comfort of three passengers are investigated as a combination of input parameters, according to the passengers. As a result, the predictive accuracy of the operation pattern according to the tested input parameter is 0.96, showing the highest accuracy. Considering each passenger, the predictive accuracy has a maximum deviation of 30%. However, we verify that it indicates the level of accuracy in predicting a passenger’s heating-operation pattern. Accordingly, the possibility of operating a heating seat without a switch operation is confirmed through machine learning. The primary-stage research result reveals whether it is possible to predict objective personal thermal comfort using the passenger seat’s heating-operation pattern. Based on the results of this research, it is expected to be utilized for system construction based on the AI prediction of operation patterns according to the passenger through machine learning.

Monitoring System for the Management of the Common Agricultural Policy Using Machine Learning and Remote Sensing

The European Commission promotes new technologies and data generated by the Copernicus Programme. These technologies are intended to improve the management of the Common Agricultural Policy aid, implement new monitoring controls to replace on-the-spot checks, and apply up to 100% of the applications continuously for an agricultural year. This paper presents a generic methodology developed for implementing monitoring controls. To achieve this, the dataset provided by the Sentinel-2 time series is transformed into information through the combination of classifications with machine learning using random forest and remote sensing-based biophysical indices. This work focuses on monitoring the helpline associated with rice cultivation, using 13 Sentinel-2 images whose grouping and characteristics change depending on the event or landmark being sought. Moreover, the functionality to check, before harvesting the crop, that the area declared is equal to the area cultivated is added. The 2020 results are around 96% for most of the metrics analysed, demonstrating the potential of Sentinel-2 for controlling subsidies, particularly for rice. After the quality assessment, the hit rate is 98%. The methodology is transformed into a tool for regular use to improve decision making by determining which declarants comply with the crop-specific aid obligations, contributing to optimising the administrations’ resources and a fairer distribution of funds.

Mapping bare ground in New Zealand hill-country agriculture and forestry for soil erosion risk assessment: An automated satellite remote-sensing method

Removing vegetation cover from hill-slope land increases risk for soil erosion and delivery of sediment to waterways. In New Zealand's productive landscapes, clear-fell harvesting of forestry blocks and winter forage grazing by agricultural livestock are two significant causes of vegetation removal. Bare ground exposed by these activities varies annually and seasonally in location and spatial extent. Modelling soil erosion therefore requires temporally and spatially explicit mapping of this bare ground. We have developed an automated mapping method using time-series satellite imagery, thereby enabling wide-area coverage and ease of updating. The temporal analysis identifies land use along with the period of vegetation removal. It produces results per land parcel (in vector format) for use in a Geographic Information System. We present a description of our method, national maps and statistics of bare ground extent in New Zealand's hill-country forestry and winter forage grazing land in 2018, and an assessment of accuracy. The attributes of the mapped land parcels are designed for input into a soil erosion estimation model such as the New Zealand Universal Soil Loss Equation.

Recognition of Maize Phenology in Sentinel Images with Machine Learning

The scarcity of water for agricultural use is a serious problem that has increased due to intense droughts, poor management, and deficiencies in the distribution and application of the resource. The monitoring of crops through satellite image processing and the application of machine learning algorithms are technological strategies with which developed countries tend to implement better public policies regarding the efficient use of water. The purpose of this research was to determine the main indicators and characteristics that allow us to discriminate the phenological stages of maize crops (Zea mays L.) in Sentinel 2 satellite images through supervised classification models. The training data were obtained by monitoring cultivated plots during an agricultural cycle. Indicators and characteristics were extracted from 41 Sentinel 2 images acquired during the monitoring dates. With these images, indicators of texture, vegetation, and colour were calculated to train three supervised classifiers: linear discriminant (LD), support vector machine (SVM), and k-nearest neighbours (kNN) models. It was found that 45 of the 86 characteristics extracted contributed to maximizing the accuracy by stage of development and the overall accuracy of the trained classification models. The characteristics of the Moran's I local indicator of spatial association (LISA) improved the accuracy of the classifiers when applied to the L*a*b* colour model and to the near-infrared (NIR) band. The local binary pattern (LBP) increased the accuracy of the classification when applied to the red, green, blue (RGB) and NIR bands. The colour ratios, leaf area index (LAI), RGB colour model, L*a*b* colour space, LISA, and LBP extracted the most important intrinsic characteristics of maize crops with regard to classifying the phenological stages of the maize cultivation. The quadratic SVM model was the best classifier of maize crop phenology, with an overall accuracy of 82.3%.

Modeling of nitrogen solubility in normal alkanes using machine learning methods compared with cubic and PC-SAFT equations of state

Accurate prediction of the solubility of gases in hydrocarbons is a crucial factor in designing enhanced oil recovery (EOR) operations by gas injection as well as separation, and chemical reaction processes in a petroleum refinery. In this work, nitrogen (N₂) solubility in normal alkanes as the major constituents of crude oil was modeled using five representative machine learning (ML) models namely gradient boosting with categorical features support (CatBoost), random forest, light gradient boosting machine (LightGBM), k-nearest neighbors (k-NN), and extreme gradient boosting (XGBoost). A large solubility databank containing 1982 data points was utilized to establish the models for predicting N₂ solubility in normal alkanes as a function of pressure, temperature, and molecular weight of normal alkanes over broad ranges of operating pressure (0.0212–69.12 MPa) and temperature (91–703 K). The molecular weight range of normal alkanes was from 16 to 507 g/mol. Also, five equations of state (EOSs) including Redlich–Kwong (RK), Soave–Redlich–Kwong (SRK), Zudkevitch–Joffe (ZJ), Peng–Robinson (PR), and perturbed-chain statistical associating fluid theory (PC-SAFT) were used comparatively with the ML models to estimate N₂ solubility in normal alkanes. Results revealed that the CatBoost model is the most precise model in this work with a root mean square error of 0.0147 and coefficient of determination of 0.9943. ZJ EOS also provided the best estimates for the N₂ solubility in normal alkanes among the EOSs. Lastly, the results of relevancy factor analysis indicated that pressure has the greatest influence on N₂ solubility in normal alkanes and the N₂ solubility increases with increasing the molecular weight of normal alkanes.

A Novel Feature Selection Technique to Better Predict Climate Change Stage of Change

Indications of people’s environmental concern are linked to transport decisions and can provide great support for policymaking on climate change. This study aims to better predict individual climate change stage of change (CC-SoC) based on different features of transport-related behavior, General Ecological Behavior, New Environmental Paradigm, and socio-demographic characteristics. Together these sources result in over 100 possible features that indicate someone’s level of environmental concern. Such a large number of features may create several analytical problems, such as overfitting, accuracy reduction, and high computational costs. To this end, a new feature selection technique, named the Coyote Optimization Algorithm-Quadratic Discriminant Analysis (COA-QDA), is first proposed to find the optimal features to predict CC-SoC with the highest accuracy. Different conventional feature selection methods (Lasso, Elastic Net, Random Forest Feature Selection, Extra Trees, and Principal Component Analysis Feature Selection) are employed to compare with the COA-QDA. Afterward, eight classification techniques are applied to solve the prediction problem. Finally, a sensitivity analysis is performed to determine the most important features affecting the prediction of CC-SoC. The results indicate that COA-QDA outperforms conventional feature selection methods by increasing average testing data accuracy from 0.7% to 5.6%. Logistic Regression surpasses other classifiers with the highest prediction accuracy.

COVID-19 anomaly detection and classification method based on supervised machine learning of chest X-ray images

The term COVID-19 is an abbreviation of Coronavirus 2019, which is considered a global pandemic that threatens the lives of millions of people. Early detection of the disease offers ample opportunity of recovery and prevention of spreading. This paper proposes a method for classification and early detection of COVID-19 through image processing using X-ray images. A set of procedures are applied, including preprocessing (image noise removal, image thresholding, and morphological operation), Region of Interest (ROI) detection and segmentation, feature extraction, (Local binary pattern (LBP), Histogram of Gradient (HOG), and Haralick texture features) and classification (K-Nearest Neighbor (KNN) and Support Vector Machine (SVM)). The combinations of the feature extraction operators and classifiers results in six models, namely LBP-KNN, HOG-KNN, Haralick-KNN, LBP-SVM, HOG-SVM, and Haralick-SVM. The six models are tested based on test samples of 5,000 images with the percentage of training of 5-folds cross-validation. The evaluation results show high diagnosis accuracy from 89.2% up to 98.66%. The LBP-KNN model outperforms the other models in which it achieves an average accuracy of 98.66%, a sensitivity of 97.76%, specificity of 100%, and precision of 100%. The proposed method for early detection and classification of COVID-19 through image processing using X-ray images is proven to be usable in which it provides an end-to-end structure without the need for manual feature extraction and manual selection methods.

Differentiation of Organic Cocoa Beans and Conventional Ones by Using Handheld NIR Spectroscopy and Multivariate Classification Techniques

The global market for organic cocoa beans continues to show sturdy growth. A low-cost handheld NIR spectrometer (900-1700 nm) combined with multivariate classification algorithms was used for rapid differentiation analysis of organic cocoa beans' integrity. In this research, organic and conventionally cultivated cocoa beans were collected from different locations in Ghana and scanned nondestructively with a handheld spectrometer. Different preprocessing treatments were employed. Principal component analysis (PCA) and classification analysis, RF (random forest), KNN (K-nearest neighbours), LDA (linear discriminant analysis), and PLS-DA (partial least squares-discriminant analysis) were performed comparatively to build classification models. The performance of the models was evaluated by accuracy, specificity, sensitivity, and efficiency. Second derivative preprocessing together with PLS-DA algorithm was superior to the rest of the algorithms with a classification accuracy of 100.00% in both the calibration set and prediction set. Second derivative algorithm was found to be the best preprocessing tool. The identification rates for the calibration set and prediction set were 96.15% and 98.08%, respectively, for RF, 91.35% and 92.31% for KNN, and 90.38% and 98.08% for LDA. Generally, the results showed that a handheld NIR spectrometer coupled with an appropriate multivariate algorithm could be used in situ for the differentiation of organic cocoa beans from conventional ones to ensure food integrity along the cocoa bean value chain.

Disaster and Pandemic Management Using Machine Learning: A Survey

This article provides a literature review of state-of-the-art machine learning (ML) algorithms for disaster and pandemic management. Most nations are concerned about disasters and pandemics, which, in general, are highly unlikely events. To date, various technologies, such as IoT, object sensing, UAV, 5G, and cellular networks, smartphone-based system, and satellite-based systems have been used for disaster and pandemic management. ML algorithms can handle multidimensional, large volumes of data that occur naturally in environments related to disaster and pandemic management and are particularly well suited for important related tasks, such as recognition and classification. ML algorithms are useful for predicting disasters and assisting in disaster management tasks, such as determining crowd evacuation routes, analyzing social media posts, and handling the post-disaster situation. ML algorithms also find great application in pandemic management scenarios, such as predicting pandemics, monitoring pandemic spread, disease diagnosis, etc. This article first presents a tutorial on ML algorithms. It then presents a detailed review of several ML algorithms and how we can combine these algorithms with other technologies to address disaster and pandemic management. It also discusses various challenges, open issues and, directions for future research.

Support Vector Machine Classifiers Show High Generalizability in Automatic Fall Detection in Older Adults

Falls are a major cause of morbidity and mortality in neurological disorders. Technical means of detecting falls are of high interest as they enable rapid notification of caregivers and emergency services. Such approaches must reliably differentiate between normal daily activities and fall events. A promising technique might be based on the classification of movements based on accelerometer signals by machine-learning algorithms, but the generalizability of classifiers trained on laboratory data to real-world datasets is a common issue. Here, three machine-learning algorithms including Support Vector Machine (SVM), k-Nearest Neighbors (kNN), and Random Forest (RF) were trained to detect fall events. We used a dataset containing intentional falls (SisFall) to train the classifier and validated the approach on a different dataset which included real-world accidental fall events of elderly people (FARSEEING). The results suggested that the linear SVM was the most suitable classifier in this cross-dataset validation approach and reliably distinguished a fall event from normal everyday activity at an accuracy of 93% and similarly high sensitivity and specificity. Thus, classifiers based on linear SVM might be useful for automatic fall detection in real-world applications.

The Utility of Sentinel-2 Spectral Data in Quantifying Above-Ground Carbon Stock in an Urban Reforested Landscape

The transformation of the natural landscape into an impervious surface due to urbanization has often been considered an important driver of environmental change, affecting essential urban ecological processes and ecosystem services. Continuous forest degradation and deforestation due to urbanization have led to an increase in atmospheric carbon emissions, risks, and impacts associated with climate change within urban landscapes and beyond them. Hence, urban reforestation has become a reliable long-term alternative for carbon sink and climate change mitigation. However, there is an urgent need for spatially accurate and concise quantification of these forest carbon stocks in order to understand and effectively monitor the accumulation and progress on such ecosystem services. Hence, this study sought to examine the prospect of Sentinel-2 spectral data in quantifying carbon stock in a reforested urban landscape using the random forest ensemble. Results show that Sentinel-2 spectral data estimated reforested forest carbon stock to an RMSE between 0.378 and 0.466 t·ha−1 and R2 of 79.82 and 77.96% using calibration and validation datasets. Based on random forest variable selection and backward elimination approaches, the red-edge normalized difference vegetation index, enhanced vegetation index, modified simple ratio index, and normalized difference vegetation index were the best subset of predictor variables of carbon stock. These findings demonstrate the value and prospects of Sentinel-2 spectral data for predicting carbon stock in reforested urban landscapes. This information is critical for adopting informed management policies and plans for optimizing urban reforested landscapes carbon sequestration capacity and improving their climate change mitigation potential.

An integrated framework with machine learning and radiomics for accurate and rapid early diagnosis of COVID-19 from Chest X-ray

The objective of the research article is to propose and validate a combination of machine learning and radiomics features to detect COVID-19 early and rapidly from chest X-ray (CXR) in presence of other viral/bacterial pneumonia and at different severity levels of diseases. It is vital to assess the performance of any diagnosis method on an independent data set and at very early stage of the disease when the disease severity of is very low. In such cases, most of the diagnosis methods fail. A total of 378 CXR images containing both normal lung and pneumonia (both COVID-19 and others lung conditions) were collected from publically available data set. 71 radiomics features for each lung segment were chosen from 100 extracted features based on Z-score heatmap and one way ANOVA test that can detect COVID-19. Three best performing classical machine learning algorithms during the training phase - 1) fine Gaussian support vector machine (SVM), 2) fine k-nearest neighbor (KNN) and 3) ensemble bagged model (EBM) trees were chosen for further evaluation on an independent test data set. The independent test data set consists of 115 COVID-19 CXR images collected from a local hospital and 100 CXR images collected from publically available data set containing normal lung and viral/bacterial pneumonia. Severity was scored between 0 to 4 by two experienced radiologists for each lung with pneumonia (both COVID-19 and non COVID-19) for the test data set. Ensemble Bagging Model Trees (EBM) with the selected radiomics features is the most suitable to distinguish between COVID-19 and other lung infections with an overall sensitivity of 87.8% and specificity of 97% (95.2% accuracy and 0.9228 area under curve) and is robust across severity levels. The method also can detect COVID-19 from CXR when two experienced radiologists were unable to detect any abnormality in the lung CXR (represented by severity score of 0). Once the CXR is acquired and lung is segmented, it takes less than two minutes for extracting radiomics features and providing diagnosis result. Since the proposed method does not require any manual intervention (e.g., sample collection etc.), it can be straightway integrated with standard X-ray reporting system to be used as an efficient, cost-effective and rapid early diagnosis device.

Mapping and Monitoring of Land Cover/Land Use (LCLU) Changes in the Crozon Peninsula (Brittany, France) from 2007 to 2018 by Machine Learning Algorithms (Support Vector Machine, Random Forest, and Convolutional Neural Network) and by Post-classification Comparison (PCC)

Land cover/land use (LCLU) is currently a very important topic, especially for coastal areas that connect the land and the coast and tend to change frequently. LCLU plays a crucial role in land and territory planning and management tasks. This study aims to complement information on the types and rates of LCLU multiannual changes with the distributions, rates, and consequences of these changes in the Crozon Peninsula, a highly fragmented coastal area. To evaluate the multiannual change detection (CD) capabilities using high-resolution (HR) satellite imagery, we implemented three remote sensing algorithms: a support vector machine (SVM), a random forest (RF) combined with geographic object-based image analysis techniques (GEOBIA), and a convolutional neural network (CNN), with SPOT 5 and Sentinel 2 data from 2007 and 2018. Accurate and timely CD is the most important aspect of this process. Although all algorithms were indicated as efficient in our study, with accuracy indices between 70% and 90%, the CNN had significantly higher accuracy than the SVM and RF, up to 90%. The inclusion of the CNN significantly improved the classification performance (5–10% increase in the overall accuracy) compared with the SVM and RF classifiers applied in our study. The CNN eliminated some of the confusion that characterizes a coastal area. Through the study of CD results by post-classification comparison (PCC), multiple changes in LCLU could be observed between 2007 and 2018: both the cultivated and non-vegetated areas increased, accompanied by high deforestation, which could be explained by the high rate of urbanization in the peninsula.

Dual Attention Feature Fusion and Adaptive Context for Accurate Segmentation of Very High-Resolution Remote Sensing Images

Land cover classification of high-resolution remote sensing images aims to obtain pixel-level land cover understanding, which is often modeled as semantic segmentation of remote sensing images. In recent years, convolutional network (CNN)-based land cover classification methods have achieved great advancement. However, previous methods fail to generate fine segmentation results, especially for the object boundary pixels. In order to obtain boundary-preserving predictions, we first propose to incorporate spatially adapting contextual cues. In this way, objects with similar appearance can be effectively distinguished with the extracted global contextual cues, which are very helpful to identify pixels near object boundaries. On this basis, low-level spatial details and high-level semantic cues are effectively fused with the help of our proposed dual attention mechanism. Concretely, when fusing multi-level features, we utilize the dual attention feature fusion module based on both spatial and channel attention mechanisms to relieve the influence of the large gap, and further improve the segmentation accuracy of pixels near object boundaries. Extensive experiments were carried out on the ISPRS 2D Semantic Labeling Vaihingen data and GaoFen-2 data to demonstrate the effectiveness of our proposed method. Our method achieves better performance compared with other state-of-the-art methods.

Poststroke Grasp Ability Assessment using an Intelligent Data Glove based on Action Research Arm Test: Development, Algorithms, and Experiments

Growing impact of poststroke upper extremity (UE) functional limitations entails newer dimensions in assessment methodologies. This has compelled researchers to think way beyond traditional stroke assessment scales during the out-patient rehabilitation phase. In concurrence with this, sensor-driven quantitative evaluation of poststroke UE functional limitations has become a fertile field of research. Here, we have emphasized an instrumented wearable for systematic monitoring of stroke patients with right-hemiparesis for evaluating their grasp abilities deploying intelligent algorithms. An instrumented glove housing 6 flex sensors, 3 force sensors, and a motion processing unit was developed to administer 19 activities of Action Research Arm Test (ARAT) while experimenting on 20 voluntarily participating subjects. After necessary signal conditioning, meaningful features were extracted, and subsequently the most appropriate ones were selected using the ReliefF algorithm. An optimally tuned support vector classifier was employed to classify patients with different degrees of disability and an accuracy of 92% was achieved supported by a high area under the receiver operating characteristic score. Furthermore, selected features could provide additional information that revealed the causes of grasp limitations. This would assist physicians in planning more effective poststroke rehabilitation strategies. Results of the one-way ANOVA test conducted on actual and predicted ARAT scores of the subjects indicated remarkable prospects of the proposed glove-based method in poststroke grasp ability assessment and rehabilitation.

Reviewing the Potential of Sentinel-2 in Assessing the Drought

This paper systematically reviews the potential of the Sentinel-2 (A and B) in assessing drought. Research findings, including the IPCC reports, highlighted the increasing trend in drought over the decades and the need for a better understanding and assessment of this phenomenon. Continuous monitoring of the Earth’s surface is an efficient method for predicting and identifying the early warnings of drought, which enables us to prepare and plan the mitigation procedures. Considering the spatial, temporal, and spectral characteristics, the freely available Sentinel-2 data products are a promising option in this area of research, compared to Landsat and MODIS. This paper evaluates the recent developments in this field induced by the launch of Sentinel-2, as well as the comparison with other existing data products. The objective of this paper is to evaluate the potential of Sentinel-2 in assessing drought through vegetation characteristics, soil moisture, evapotranspiration, surface water including wetland, and land use and land cover analysis. Furthermore, this review also addresses and compares various data fusion methods and downscaling methods applied to Sentinel-2 for retrieving the major bio-geophysical variables used in the analysis of drought. Additionally, the limitations of Sentinel-2 in its direct applicability to drought studies are also evaluated.

Recognizing Physical Activities for Spinal Cord Injury Rehabilitation Using Wearable Sensors

The research area of activity recognition is fast growing with diverse applications. However, advances in this field have not yet been used to monitor the rehabilitation of individuals with spinal cord injury. Noteworthily, relying on patient surveys to assess adherence can undermine the outcomes of rehabilitation. Therefore, this paper presents and implements a systematic activity recognition method to recognize physical activities applied by subjects during rehabilitation for spinal cord injury. In the method, raw sensor data are divided into fragments using a dynamic segmentation technique, providing higher recognition performance compared to the sliding window, which is a commonly used approach. To develop the method and build a predictive model, a machine learning approach was adopted. The proposed method was evaluated on a dataset obtained from a single wrist-worn accelerometer. The results demonstrated the effectiveness of the proposed method in recognizing all of the activities that were examined, and it achieved an overall accuracy of 96.86%.

Semantic Boosting: Enhancing Deep Learning Based LULC Classification

The classification of land use and land cover (LULC) is a well-studied task within the domain of remote sensing and geographic information science. It traditionally relies on remotely sensed imagery and therefore models land cover classes with respect to their electromagnetic reflectances, aggregated in pixels. This paper introduces a methodology which enables the inclusion of geographical object semantics (from vector data) into the LULC classification procedure. As such, information on the types of geographic objects (e.g., Shop, Church, Peak, etc.) can improve LULC classification accuracy. In this paper, we demonstrate how semantics can be fused with imagery to classify LULC. Three experiments were performed to explore and highlight the impact and potential of semantics for this task. In each experiment CORINE LULC data was used as a ground truth and predicted using imagery from Sentinel-2 and semantics from LinkedGeoData using deep learning. Our results reveal that LULC can be classified from semantics only and that fusing semantics with imagery—Semantic Boosting—improved the classification with significantly higher LULC accuracies. The results show that some LULC classes are better predicted using only semantics, others with just imagery, and importantly much of the improvement was due to the ability to separate similar land use classes. A number of key considerations are discussed.

Estimation of Urban Land-Use Efficiency for Sustainable Development by Integrating over 30-Year Landsat Imagery with Population Data: A Case Study of Ha Long, Vietnam

Humans are moving into urban areas at an accelerated pace. An increasing urban population fuels urban expansion and reduces nearby agricultural lands and natural environments such as forests, swamps, other water-pervious areas. Unsustainable development creates a disproportion between the growth of urban areas and the growth in urban population. The UN SDG indicator 11.3.1 specifically addresses the issue of the measurement of land-use efficiency. While the metric and methodology to estimate the indicator are straightforward, it faces problems of data unavailability and inconsistency. Vietnam has a record of tremendous economic growth that has translated into more urban settlements of size. Consequently, rural population movement into urban areas has led to many urban sustainable planning and development challenges. In the absence of previous work on estimating land-use efficiency in Vietnamese cities, this study makes the first attempt to examine land-use efficiency in Ha Long, one of the country’s fast-growing cities in recent decades. We mapped land use from high-resolution Landsat imagery (30 m) spanning multi-decadal observations from 1986 to 2020. An advanced machine learning approach, the Support Vector Machine algorithm, was applied to estimate the built-up area, which, by integration with census data, is essential for calculating SDG indicator 11.3.1. This study shows that the land-use efficiency metric was positive but small at the beginning of the considered period but increased in 2000–2020. These results suggest that before 2000, the urban land consumption rate in Ha Long was lower than the population growth rate, implying denser urban land use. The situation changed to the opposite when the urban land consumption rate exceeded the population growth rate in the past two decades. The study’s approach is applicable to regional and district levels to provide comparative analyses between cities or parts of a region or districts of the city. These analyses are valuable tools for assessing the impact of local urban and municipal planning policies on urban development.

Crash severity analysis of vulnerable road users using machine learning

Road crash fatality is a universal problem of the transportation system. A massive death toll caused annually due to road crash incidents, and among them, vulnerable road users (VRU) are endangered with high crash severity. This paper focuses on employing machine learning-based classification approaches for modelling injury severity of vulnerable road users-pedestrian, bicyclist, and motorcyclist. Specifically, this study aims to analyse critical features associated with different VRU groups-for pedestrian, bicyclist, motorcyclist and all VRU groups together. The critical factor of crash severity outcomes for these VRU groups is estimated in identifying the similarities and differences across different important features associated with different VRU groups. The crash data for the study is sourced from the state of Queensland in Australia for the years 2013 through 2019. The supervised machine learning algorithms considered for the empirical analysis includes the K-Nearest Neighbour (KNN), Support Vector Machine (SVM) and Random Forest (RF). In these models, 17 distinct road crash parameters are considered as input features to train models, which originate from road user characteristics, weather and environment, vehicle and driver condition, period, road characteristics and regions, traffic, and speed jurisdiction. These classification models are separately trained and tested for individual and unified VRU to assess crash severity levels. Afterwards, model performances are compared with each other to justify the best classifier where Random Forest classification models for all VRU modes are found to be comparatively robust in test accuracy: (motorcyclist: 72.30%, bicyclist: 64.45%, pedestrian: 67.23%, unified VRU: 68.57%). Based on the Random Forest model, the road crash features are ranked and compared according to their impact on crash severity classification. Furthermore, a model-based partial dependency of each road crash parameters on the severity levels is plotted and compared for each individual and unified VRU. This clarifies the tendency of road crash parameters to vary with different VRU crash severity. Based on the outcome of the comparative analysis, motorcyclists are found to be more likely exposed to higher crash severity, followed by pedestrians and bicyclists.

Impervious Surfaces Mapping at City Scale by Fusion of Radar and Optical Data through a Random Forest Classifier

Urbanization increases the amount of impervious surfaces, making accurate information on spatial and temporal expansion trends essential; the challenge is to develop a cost- and labor-effective technique that is compatible with the assessment of multiple geographical locations in developing countries. Several studies have identified the potential of remote sensing and multiple source information in impervious surface quantification. Therefore, this study aims to fuse datasets from the Sentinel 1 and 2 Satellites to map the impervious surfaces of nine Pakistani cities and estimate their growth rates from 2016 to 2020 utilizing the random forest algorithm. All bands in the optical and radar images were resampled to 10 m resolution, projected to same coordinate system and geometrically aligned to stack into a single product. The models were then trained, and classifications were validated with land cover samples from Google Earth’s high-resolution images. Overall accuracies of classified maps ranged from 85% to 98% with the resultant quantities showing a strong linear relationship (R-squared value of 0.998) with the Copernicus Global Land Services data. There was up to 9% increase in accuracy and up to 12 % increase in kappa coefficient from the fused data with respect to optical alone. A McNemar test confirmed the superiority of fused data. Finally, the cities had growth rates ranging from 0.5% to 2.5%, with an average of 1.8%. The information obtained can alert urban planners and environmentalists to assess impervious surface impacts in the cities.

Mapping Maize Area in Heterogeneous Agricultural Landscape with Multi-Temporal Sentinel-1 and Sentinel-2 Images Based on Random Forest

Accurate estimation of crop area is essential to adjusting the regional crop planting structure and the rational planning of water resources. However, it is quite challenging to map crops accurately by high-resolution remote sensing images because of the ecological gradient and ecological convergence between crops and non-crops. The purpose of this study is to explore the combining application of high-resolution multi-temporal Sentinel-1 (S1) radar backscatter and Sentinel-2 (S2) optical reflectance images for maize mapping in highly complex and heterogeneous landscapes in the middle reaches of Heihe River, northwest China. We proposed a new two-step method of vegetation extraction and followed by maize extraction, that is, extract the vegetation-covered areas first to reduce the inter-class variance by using a Random Forest (RF) classifier based on S2 data, and then extract the maize distribution in the vegetation area by using another RF classifier based on S1 and/or S2 data. The results demonstrate that the vegetation extraction classifier successfully identified vegetation-covered regions with an overall accuracy above 96% in the study area, and the accuracy of the maize extraction classifier constructed by the combined multi-temporal S1 and S2 images is significantly improved compared with that S1 (alone) or S2 (alone), with an overall accuracy of 87.63%, F1_Score of 0.86, and Kappa coefficient of 0.75. In addition, with the introduction of multi-temporal S1 and/or S2 images in crop growing season, the constructed RF model is more beneficial to maize mapping.

Classification Efficacy Using K-Fold Cross-Validation and Bootstrapping Resampling Techniques on the Example of Mapping Complex Gully Systems

The availability of aerial and satellite imageries has greatly reduced the costs and time associated with gully mapping, especially in remote locations. Regardless, accurate identification of gullies from satellite images remains an open issue despite the amount of literature addressing this problem. The main objective of this work was to investigate the performance of support vector machines (SVM) and random forest (RF) algorithms in extracting gullies based on two resampling methods: bootstrapping and k-fold cross-validation (CV). In order to achieve this objective, we used PlanetScope data, acquired during the wet and dry seasons. Using the Normalized Difference Vegetation Index (NDVI) and multispectral bands, we also explored the potential of the PlanetScope image in discriminating gullies from the surrounding land cover. Results revealed that gullies had significantly different (p < 0.001) spectral profiles from any other land cover class regarding all bands of the PlanetScope image, both in the wet and dry seasons. However, NDVI was not efficient in gully discrimination. Based on the overall accuracies, RF’s performance was better with CV, particularly in the dry season, where its performance was up to 4% better than the SVM’s. Nevertheless, class level metrics (omission error: 11.8%; commission error: 19%) showed that SVM combined with CV was more successful in gully extraction in the wet season. On the contrary, RF combined with bootstrapping had relatively low omission (16.4%) and commission errors (10.4%), making it the most efficient algorithm in the dry season. The estimated gully area was 88 ± 14.4 ha in the dry season and 57.2 ± 18.8 ha in the wet season. Based on the standard error (8.2 ha), the wet season was more appropriate in gully identification than the dry season, which had a slightly higher standard error (8.6 ha). For the first time, this study sheds light on the influence of these resampling techniques on the accuracy of satellite-based gully mapping. More importantly, this study provides the basis for further investigations into the accuracy of such resampling techniques, especially when using different satellite images other than the PlanetScope data.

Ecological Status as the Basis for the Holistic Environmental Flow Assessment of a Tropical Highland River in Ethiopia

There is an increasing need globally to establish relationships among flow, ecology, and livelihoods to make informed decisions about environmental flows. This paper aimed to establish the ecological foundation for a holistic environmental flow assessment method in the Gumara River that flows into Lake Tana in Ethiopia and the Blue Nile River. First, the ecological conditions (fish, macro-invertebrate, riparian vegetation, and physicochemical) of the river system were characterized, followed by determining the hydrological condition and finally linking the ecological and hydrological components. The ecological data were collected at 30 sites along the Gumara River on March 2016 and 2020. River hydrology was estimated using the SWAT model and showed that the low flow decreased over time. Both physico-chemical and macroinvertebrate scores showed that water quality was moderate in most locations. The highest fish diversity index was in the lower reach at Wanzaye. Macroinvertebrate diversity was observed to decrease downstream. Both the fish and macroinvertebrate diversity indices were less than the expected maximum, being 3.29 and 4.5, respectively. The normalized difference vegetation index (NDVI) for 30 m and 60 m buffer distances from the river decreased during the dry season (March–May). Hence, flow conditions, water quality, and land-use change substantially influenced the abundance and diversity of fish, vegetation, and macroinvertebrate species. The pressure on the ecology is expected to increase because the construction of the proposed dam is expected to alter the flow regime. Thus, as demand for human water consumption grows, measures are needed, including quantification of environmental flow requirements and regulating river water uses to conserve the ecological status of the Gumara River and Lake Tana sub-basin.