Terrestrial Remotely Sensed Imagery in Support of Public Health: New Avenues of Research Using Object-Based Image Analysis

Temporal assessment of forest cover dynamics in response to forest fires and other environmental impacts using AI

The rapid reduction of forests due to environmental impacts such as deforestation, global warming, natural disasters such as forest fires as well as various human activities is an escalating concern. The increasing frequency and severity of forest fires are causing significant harm to the ecosystem, economy, wildlife, and human safety. During dry and hot seasons, the likelihood of forest fires also increases. It is crucial to accurately monitor and analyze the large-scale changes in the forest cover to ensure sustainable forest management. Remote sensing technology helps to precisely study such changes in forest cover over a wide area over time. This research analyzes the impact of forest fires over time, identifies hotspots, and explores the environmental factors that affect forest cover change. Sentinel-2 imagery was utilized to study changes in Brunei Darussalam's forest cover area over five years from 2017 to 2022. An object-based approach, Simple Non-Iterative Clustering (SNIC), is employed to cluster the region using NDVI values and analyze the changes per cluster. The results indicate that the area of the clusters reduced where fire incidence occurred as well as the precipitation dropped. Between 2017 and 2022, the increased forest fires and decreased precipitation levels resulted in the change in cluster areas as follows: 66.11%, 69.46%, 68.32%, 73.88%, 77.27%, and 78.70%, respectively. Additionally, hotspots in response to forest fires each year were identified in the Belait district. This study will help forest managers assess the causes of forest cover loss and develop conservation and afforestation strategies.

An integrated system for the management of environmental data to support veterinary epidemiology

Environmental and climatic fluctuations can greatly influence the dynamics of infectious diseases of veterinary concern, or interfere with the implementation of relevant control measures. Including environmental and climatic aspects in epidemiological studies could provide policy makers with new insights to assign resources for measures to prevent or limit the spread of animal diseases, particularly those with zoonotic potential. The ever-increasing number of technologies and tools permits acquiring environmental data from various sources, including ground-based sensors and Satellite Earth Observation (SEO). However, the high heterogeneity of these datasets often requires at least some basic GIS (Geographic Information Systems) and/or coding skills to use them in further analysis. Therefore, the high availability of data does not always correspond to widespread use for research purposes. The development of an integrated data pre-processing system makes it possible to obtain information that could be easily and directly used in subsequent epidemiological analyses, supporting both research activities and the management of disease outbreaks. Indeed, such an approach allows for the reduction of the time spent on searching, downloading, processing and validating environmental data, thereby optimizing available resources and reducing any possible errors directly related to data collection. Although multitudes of free services that allow obtaining SEO data exist nowadays (either raw or pre-processed through a specific coding language), the availability and quality of information can be sub-optimal when dealing with very small scale and local data. In fact, some information sets (e.g., air temperature, rainfall), usually derived from ground-based sensors (e.g., agro-meteo station), are managed, processed and redistributed by agencies operating on a local scale which are often not directly accessible by the most common free SEO services (e.g., Google Earth Engine). The EVE (Environmental data for Veterinary Epidemiology) system has been developed to acquire, pre-process and archive a set of environmental information at various scales, in order to facilitate and speed up access by epidemiologists, researchers and decision-makers, also accounting for the integration of SEO information with locally sensed data.

Mapping Restoration Activities on Dirk Hartog Island Using Remotely Piloted Aircraft Imagery

Conservation practitioners require cost-effective and repeatable remotely sensed data for assistive monitoring. This paper tests the ability of standard remotely piloted aircraft (DJI Phantom 4 Pro) imagery to discriminate between plant species in a rangeland environment. Flights were performed over two 0.3–0.4 ha exclusion plot sites, established as controls to protect vegetation from translocated animal disturbance on Dirk Hartog Island, Western Australia. Comparisons of discriminatory variables, classification potential, and optimal flight height were made between plot sites with different plant species diversity. We found reflectance bands and height variables to have high differentiation potential, whilst measures of texture were less useful for multisegmented plant canopies. Discrimination between species varied with omission errors ranging from 13 to 93%. Purposely resampling c. 5 mm imagery as captured at 20–25 m above terrain identified that a flight height of 120 m would improve capture efficiency in future surveys without hindering accuracy. Overall accuracy at a site with low species diversity (n = 4) was 70%, which is an encouraging result given the imagery is limited to visible spectral bands. With higher species diversity (n = 10), the accuracy reduced to 53%, although it is expected to improve with additional bands or grouping like species. Findings suggest that in rangeland environments with low species diversity, monitoring using a standard RPA is viable.

An Alternative Method for the Generation of Consistent Mapping to Monitoring Land Cover Change: A Case Study of Guerrero State in Mexico

Land cover is crucial for ecosystems and human activities. Therefore, monitoring land cover changes has become relevant in recent years. This study proposes an alternative method based on conventional change detection techniques combined with maximum likelihood (MaxLike) supervised classification of satellite images to generate consistent Land Use/Land Cover (LULC) maps. The novelty of this method is that the supervised classification is applied in an earlier stage of change detection exclusively to identified dynamics zones. The LULC categories of the stable zones are acquired from an initial date’s previously elaborated base map. The methodology comprised the use of Landsat images from 2011 and 2016, applying the Sun Canopy Sensor (SCS + C) topographic correction model enhanced through the classification of slopes, using derived topographic corrected images with NDVI, and employing Tasseled Cap (TC) Brightness-Greenness-Wetness indices and Principal Components (PCs). The study incorporated a comparative analysis of the consistency of the LULC mapping, which is generated based on control areas. The results show that the proposed method, although slightly laborious, is viable and fully automatable. The generated LULC map is accurate and robust and achieves a Kappa concordance index of 87.53. Furthermore, the boundary consistency was visually superior to the conventional classified map.

Detection of Standing Deadwood from Aerial Imagery Products: Two Methods for Addressing the Bare Ground Misclassification Issue

Deadwood mapping is of high relevance for studies on forest biodiversity, forest disturbance, and dynamics. As deadwood predominantly occurs in forests characterized by a high structural complexity and rugged terrain, the use of remote sensing offers numerous advantages over terrestrial inventory. However, deadwood misclassifications can occur in the presence of bare ground, displaying a similar spectral signature. In this study, we tested the potential to detect standing deadwood (h > 5 m) using orthophotos (0.5 m resolution) and digital surface models (DSM) (1 m resolution), both derived from stereo aerial image matching (0.2 m resolution and 60%/30% overlap (end/side lap)). Models were calibrated in a 600 ha mountain forest area that was rich in deadwood in various stages of decay. We employed random forest (RF) classification, followed by two approaches for addressing the deadwood-bare ground misclassification issue: (1) post-processing, with a mean neighborhood filter for “deadwood”-pixels and filtering out isolated pixels and (2) a “deadwood-uncertainty” filter, quantifying the probability of a “deadwood”-pixel to be correctly classified as a function of the environmental and spectral conditions in its neighborhood. RF model validation based on data partitioning delivered high user’s (UA) and producer’s (PA) accuracies (both > 0.9). Independent validation, however, revealed a high commission error for deadwood, mainly in areas with bare ground (UA = 0.60, PA = 0.87). Post-processing (1) and the application of the uncertainty filter (2) improved the distinction between deadwood and bare ground and led to a more balanced relation between UA and PA (UA of 0.69 and 0.74, PA of 0.79 and 0.80, under (1) and (2), respectively). Deadwood-pixels showed 90% location agreement with manually delineated reference to deadwood objects. With both alternative solutions, deadwood mapping achieved reliable results and the highest accuracies were obtained with deadwood-uncertainty filter. Since the information on surface heights was crucial for correct classification, enhancing DSM quality could substantially improve the results.

Decision Tree Algorithms for Developing Rulesets for Object-Based Land Cover Classification

Decision tree (DT) algorithms are important non-parametric tools used for land cover classification. While different DTs have been applied to Landsat land cover classification, their individual classification accuracies and performance have not been compared, especially on their effectiveness to produce accurate thresholds for developing rulesets for object-based land cover classification. Here, the focus was on comparing the performance of five DT algorithms: Tree, C5.0, Rpart, Ipred, and Party. These DT algorithms were used to classify ten land cover classes using Landsat 8 images on the Copperbelt Province of Zambia. Classification was done using object-based image analysis (OBIA) through the development of rulesets with thresholds defined by the DTs. The performance of the DT algorithms was assessed based on: (1) DT accuracy through cross-validation; (2) land cover classification accuracy of thematic maps; and (3) other structure properties such as the sizes of the tree diagrams and variable selection abilities. The results indicate that only the rulesets developed from DT algorithms with simple structures and a minimum number of variables produced high land cover classification accuracies (overall accuracy > 88%). Thus, algorithms such as Tree and Rpart produced higher classification results as compared to C5.0 and Party DT algorithms, which involve many variables in classification. This high accuracy has been attributed to the ability to minimize overfitting and the capacity to handle noise in the data during training by the Tree and Rpart DTs. The study produced new insights on the formal selection of DT algorithms for OBIA ruleset development. Therefore, the Tree and Rpart algorithms could be used for developing rulesets because they produce high land cover classification accuracies and have simple structures. As an avenue of future studies, the performance of DT algorithms can be compared with contemporary machine-learning classifiers (e.g., Random Forest and Support Vector Machine).

Characterizing the Landscape Structure of Urban Wetlands Using Terrain and Landscape Indices

Several studies have shown human impacts on urban wetlands. These impacts are mostly studied at broad scales, which may generalize and aggregate important information needed for landscape quantification or terrain analysis. This situation can weakly or inappropriately address the structure of wetland landscapes, thus affecting the assessment of the quantities and qualities of terrestrial wetland habitats. To address these issues for urban wetland dynamics, this study proposes the use of landscape and terrain indices to characterize the landscape structure of urban wetlands at a fine scale in order to assess its usefulness in contributing to wildlife sustainability. To achieve this goal, secondary terrain attribute data are integrated with an object-based satellite image classification at the wetland and watershed level. The result reveals a general swell in wetland coverage at the watershed level. Further analysis shows the size and shape complexities, and edge irregularities are increased significantly at the patch level but slightly at the watershed level. Terrain analysis further reveals a potential increase in wetness and decrease in stream power vulnerability for most of the major wetlands under study. These results suggest that terrain and landscape indices are effective in characterizing the structure of urban wetlands that supports socio-ecological sustainability.

EMMCNN: An ETPS-Based Multi-Scale and Multi-Feature Method Using CNN for High Spatial Resolution Image Land-Cover Classification

Land-cover information is significant for land-use planning, urban management, and environment monitoring. This paper presented a novel extended topology-preserving segmentation (ETPS)-based multi-scale and multi-feature method using the convolutional neural network (EMMCNN) for high spatial resolution (HSR) image land-cover classification. The EMMCNN first segmented the images into superpixels using the ETPS algorithm with false-color composition and enhancement and built parallel convolutional neural networks (CNNs) with dense connections for superpixel multi-scale deep feature learning. Then, the multi-resolution segmentation (MRS) object hand-delineated features were extracted and mapped to superpixels for complementary multi-segmentation and multi-type representation. Finally, a hybrid network was designed to consist of 1-dimension CNN and multi-layer perception (MLP) with channel-wise stacking and attention-based weighting for adaptive feature fusion and comprehensive classification. Experimental results on four real HSR GaoFen-2 datasets demonstrated the superiority of the proposed EMMCNN over several well-known classification methods in terms of accuracy and consistency, with overall accuracy averagely improved by 1.74% to 19.35% for testing images and 1.06% to 8.78% for validating images. It was found that the solution combining an appropriate number of larger scales and multi-type features is recommended for better performance. Efficient superpixel segmentation, networks with strong learning ability, optimized multi-scale and multi-feature solution, and adaptive attention-based feature fusion were key points for improving HSR image land-cover classification in this study.

Identification of a Threshold Minimum Area for Reflectance Retrieval from Thermokarst Lakes and Ponds Using Full-Pixel Data from Sentinel-2

Thermokarst waterbodies caused by permafrost thawing and degradation are ubiquitous in many subarctic and Arctic regions. They are globally important components of the biogeochemical carbon cycle and have potential feedback effects on climate. These northern waters are mostly small lakes and ponds, and although they may be mapped using very high-resolution satellites or aerial photography, these approaches are generally not suitable for monitoring purposes, due to the cost and limited availability of such images. In this study we evaluated the potential use of widely available high-resolution imagery from Sentinel-2 (S2) for the characterization of the spectral reflectance of thermokarst lakes and ponds. Specifically, we aimed to define the minimum lake area that could be reliably imaged, and to identify challenges and solutions for remote sensing of such waters in the future. The study was conducted in subarctic Canada, in the vicinity of Whapmagoostui-Kuujjuarapik (Nunavik, Québec), an area in the sporadic permafrost zone with numerous thermokarst waterbodies that vary greatly in size. Ground truthing lake reflectance data were collected using an Unmanned Aerial System (UAS) fitted with a multispectral camera that collected images at 13 cm resolution. The results were compared with reflectance from Sentinel-2 images, and the effect of lake area on the reflectance response was assessed. Our results show that Sentinel-2 imagery was suitable for waterbodies larger than 350 m2 once their boundaries were defined, which in the two test sites would allow monitoring from 11% to 30% of the waterbodies and 73% to 85% of the total lake area. Challenges for remote sensing of small lakes include the confounding effects of water reflection (both direct radiation and diffuse), wind and shadow. Given the small threshold area and frequent revisit time, Sentinel-2 provides a valuable approach towards the continuous monitoring of waterbodies, including ponds and small lakes such as those found in thermokarst landscapes. UASs provide a complementary approach for ground truthing and boundary definition.