Detection and Monitoring of Forest Fires Using Himawari-8 Geostationary Satellite Data in South Korea

Advancements in remote sensing for active fire detection: A review of datasets and methods

This study comprehensively and critically reviews active fire detection advancements in remote sensing from 1975 to the present, focusing on two main perspectives: datasets and corresponding instruments, and detection algorithms. The study highlights the increasing role of machine learning, particularly deep learning techniques, in active fire detection. Looking forward, the review outlines current challenges and future research opportunities in remote sensing for active fire detection. These include exploring data quality management and multi-modal learning, developing spatiotemporally explicit models, investigating self-supervised learning models, improving explainable and interpretable models, integrating physical-process based models with machine learning, and building digital twins to replicate wildfire dynamics and perform what-if scenario analysis. The review aims to serve as a valuable resource for informing natural resource management and enhancing environmental protection efforts through the application of remote sensing technology.

Early Stage Forest Fire Detection from Himawari-8 AHI Images Using a Modified MOD14 Algorithm Combined with Machine Learning

The early detection and rapid extinguishing of forest fires are effective in reducing their spread. Based on the MODIS Thermal Anomaly (MOD14) algorithm, we propose an early stage fire detection method from low-spatial-resolution but high-temporal-resolution images, observed by the Advanced Himawari Imager (AHI) onboard the geostationary meteorological satellite Himawari-8. In order to not miss early stage forest fire pixels with low temperature, we omit the potential fire pixel detection from the MOD14 algorithm and parameterize four contextual conditions included in the MOD14 algorithm as features. The proposed method detects fire pixels from forest areas using a random forest classifier taking these contextual parameters, nine AHI band values, solar zenith angle, and five meteorological values as inputs. To evaluate the proposed method, we trained the random forest classifier using an early stage forest fire data set generated by a time-reversal approach with MOD14 products and time-series AHI images in Australia. The results demonstrate that the proposed method with all parameters can detect fire pixels with about 90% precision and recall, and that the contribution of contextual parameters is particularly significant in the random forest classifier. The proposed method is applicable to other geostationary and polar-orbiting satellite sensors, and it is expected to be used as an effective method for forest fire detection.

Fire Monitoring Algorithm and Its Application on the Geo-Kompsat-2A Geostationary Meteorological Satellite

Geo-Kompsat-2A (GK-2A) is the third new-generation geostationary meteorological satellite that orbits Asia and monitors China and its surrounding areas, following the Himawari-8 and Fengyun-4A satellites. The nadir point positioning and satellite channel parameters of the GK-2A are better than those of the Himawari-8 and FY-4A, which are more conducive to fire monitoring in China. In this study, a new fire detection algorithm is proposed based on GK-2A satellite data. That is, considering the large solar zenith angle correction for reflectance and the proportion information of background pixels in the existing spatial threshold method, fires under the different underlying surface types and solar radiation states can be automatically identified. Moreover, the accuracy of the Himawari-8 fire monitoring algorithm and the present algorithm of GK-2A is compared and analyzed through the ground truth fire spot data. The results show that compared with the original fire monitoring algorithm with fixed parameter thresholds, the brightness temperature difference of this algorithm is reduced by 0.55 K, and the correction coefficient is reduced by 0.6 times, the fire can be found earlier, and the monitoring sensitivity is improved. According to the practical fire case, the present fire monitoring algorithm of GK-2A has better monitoring accuracy than the fire monitoring algorithm of Himawari-8. The present fire monitoring algorithm of GK-2A can meet the fire monitoring requirements under different sun angles, different cloud cover ratios and vegetation ratios with good versatility.

Forest Fire Monitoring and Positioning Improvement at Subpixel Level: Application to Himawari-8 Fire Products

Forest fires are among the biggest threats to forest ecosystems and forest resources, and can lead to ecological disasters and social crises. Therefore, it is imperative to detect and extinguish forest fires in time to reduce their negative impacts. Satellite remote sensing, especially meteorological satellites, has been a useful tool for forest-fire detection and monitoring because of its high temporal resolution over large areas. Researchers monitor forest fires directly at pixel level, which usually presents a mixture of forest and fire, but the low spatial resolution of such mixed pixels cannot accurately locate the exact position of the fire, and the optimal time window for fire suppression can thus be missed. In order to improve the positioning accuracy of the origin of forest fire (OriFF), we proposed a mixed-pixel unmixing integrated with pixel-swapping algorithm (MPU-PSA) model to monitor the OriFFs in time. We then applied the model to the Japanese Himawari-8 Geostationary Meteorological Satellite data to obtain forest-fire products at subpixel level. In this study, the ground truth data were provided by the Department of Emergency Management of Hunan Province, China. To validate the positioning accuracy of MPU-PSA for OriFFs, we applied the model to the Himawari-8 satellite data and then compared the derived fire results with fifteen reference forest-fire events that occurred in Hunan Province, China. The results show that the extracted forest-fire locations using the proposed method, referred to as forest fire locations at subpixel (FFLS) level, were far closer to the actual OriFFs than those from the modified Himawari-8 Wild Fire Product (M-HWFP). This improvement will help to reduce false fire claims in the Himawari-8 Wild Fire Product (HWFP). We conducted a comparative study of M-HWFP and FFLS products using three accuracy-evaluation indexes, i.e., Euclidean distance, RMSE, and MAE. The mean distances between M-HWFP fire locations and OriFFs and between FFLS fire locations and OriFFs were 3362.21 m and 1294.00 m, respectively. The mean RMSEs of the M-HWFP and FFLS products are 1225.52 m and 474.93 m, respectively. The mean MAEs of the M-HWFP and FFLS products are 992.12 m and 387.13 m, respectively. We concluded that the newly proposed MPU-PSA method can extract forest-fire locations at subpixel level, providing higher positioning accuracy of forest fires for their suppression.

Active Fire Detection from Landsat-8 Imagery Using Deep Multiple Kernel Learning

Active fires are devastating natural disasters that cause socio-economical damage across the globe. The detection and mapping of these disasters require efficient tools, scientific methods, and reliable observations. Satellite images have been widely used for active fire detection (AFD) during the past years due to their nearly global coverage. However, accurate AFD and mapping in satellite imagery is still a challenging task in the remote sensing community, which mainly uses traditional methods. Deep learning (DL) methods have recently yielded outstanding results in remote sensing applications. Nevertheless, less attention has been given to them for AFD in satellite imagery. This study presented a deep convolutional neural network (CNN) “MultiScale-Net” for AFD in Landsat-8 datasets at the pixel level. The proposed network had two main characteristics: (1) several convolution kernels with multiple sizes, and (2) dilated convolution layers (DCLs) with various dilation rates. Moreover, this paper suggested an innovative Active Fire Index (AFI) for AFD. AFI was added to the network inputs consisting of the SWIR2, SWIR1, and Blue bands to improve the performance of the MultiScale-Net. In an ablation analysis, three different scenarios were designed for multi-size kernels, dilation rates, and input variables individually, resulting in 27 distinct models. The quantitative results indicated that the model with AFI-SWIR2-SWIR1-Blue as the input variables, using multiple kernels of sizes 3 × 3, 5 × 5, and 7 × 7 simultaneously, and a dilation rate of 2, achieved the highest F1-score and IoU of 91.62% and 84.54%, respectively. Stacking AFI with the three Landsat-8 bands led to fewer false negative (FN) pixels. Furthermore, our qualitative assessment revealed that these models could detect single fire pixels detached from the large fire zones by taking advantage of multi-size kernels. Overall, the MultiScale-Net met expectations in detecting fires of varying sizes and shapes over challenging test samples.

A Systematic Review of Applications of Machine Learning Techniques for Wildfire Management Decision Support

Wildfires threaten and kill people, destroy urban and rural property, degrade air quality, ravage forest ecosystems, and contribute to global warming. Wildfire management decision support models are thus important for avoiding or mitigating the effects of these events. In this context, this paper aims at providing a review of recent applications of machine learning methods for wildfire management decision support. The emphasis is on providing a summary of these applications with a classification according to the case study type, machine learning method, case study location, and performance metrics. The review considers documents published in the last four years, using a sample of 135 documents (review articles and research articles). It is concluded that the adoption of machine learning methods may contribute to enhancing support in different fire management phases.

Retrieval of Aerosol Optical Thickness with Custom Aerosol Model Using SKYNET Data over the Chiba Area

The retrieval of the aerosol optical thickness (AOT) from remotely-sensed data relies on the adopted aerosol model. However, the method of this technique has been rather limited because of the high variability of the surface albedo, in addition to the spatial variability in the aerosol properties over the land surfaces. To overcome unsolved problems, we proposed a method for the visibility-derived AOT estimation from SKYNET-based measurement and daytime satellite images with a custom aerosol model over the Chiba area (35.62° N, 140.10° E), which is located in the greater Tokyo metropolitan area in Japan. Different from conventionally-used aerosol models for the boundary layer, we created a custom aerosol model by using sky-radiometer observation data of aerosol volume size distribution and refractive indices, coupled with spectral response functions (SPFs) of satellite visible bands to alleviate the wide range of path-scattered radiance. We utilized the radiative transfer code 6S to implement the radiative transfer calculation based on the created custom aerosol model. The concurrent data from ground-based measurement are used in the radiative analysis, namely the temporal variation of AOT from SKYNET. The radiative estimation conducted under clear-sky conditions with minimum aerosol loading is used for the determination of the surface albedo, so that the 6S simulation yields a well-defined relation between total radiance and surface albedo. We made look-up tables (LUTs) pixel-by-pixel over the Chiba area for the custom aerosol model to retrieve the satellite AOT distribution based on the surface albedo. Therefore, such a reference of surface albedo generated from clear-sky conditions, in turn, can be employed to retrieve the spatial distribution of AOT on both clear and relatively turbid days. The value for the AOTs retrieved using the custom aerosol model is found to be stable than conventionally-used typical aerosol models, indicating that our method yields substantially better performance.

Decision Support System Development of Wildland Fire: A Systematic Mapping

Wildland fires have been a rising problem on the worldwide level, generating ecological and economic losses. Specifically, between wildland fire types, uncontrolled fires are critical due to the potential damage to the ecosystem and their effects on the soil, and, in the last decade, different technologies have been applied to fight them. Selecting a specific technology and Decision Support Systems (DSS) is fundamental, since the results and validity of this could drastically oscillate according to the different environmental and geographic factors of the terrain to be studied. Given the above, a systematic mapping was realized, with the purpose of recognizing the most-used DSS and context where they have been applied. One hundred and eighty-three studies were found that used different types of DSS to solve problems of detection, prediction, prevention, monitoring, simulation, administration, and access to routes. The concepts key to the type of solution are related to the use or development of systems or Information and Communication Technologies (ICT) in the computer science area. Although the use of BA and Big Data has increased in recent years, there are still many challenges to face, such as staff training, the friendly environment of DSS, and real-time decision-making.

Monitoring and Evaluating Restoration Vegetation Status in Mine Region Using Remote Sensing Data: Case Study in Inner Mongolia, China

The ecological restoration of mining areas is very important, and repeated field surveys are inefficient in large-scale vegetation monitoring. The coal mining industry is currently facing the challenge of the lack of appropriate methods for monitoring restoration processes. This study used an open pit coal mine in Dongsheng District, Inner Mongolia, China as an example, and used the 2011–2018 Landsat TM/ETM+ and OLI images to monitor and evaluate vegetation restoration activity of the coal mine. The average value of the monthly maximum value of vegetation index in the growing season was selected as the basic indicator for studying vegetation and bare soil changes. The growth root normalized differential vegetation index (GRNDVI) and GRNDVI anomaly method indicated that the constructed land type change factor was used to study the growth of mine vegetation and change of the range of bare land in the entire mining region. We found that westward mining activities started from 2012, and vegetation was restored in the eastern original mining region from 2013. The restoration vegetation areas from 2015 to 2016 and from 2017 to 2018 were larger than those in the other restoration years. Moreover, areas of expanded bare land from 2011 to 2012, and from 2017 to 2018 were larger than those in the other expansion years. The restoration vegetation growth changes were compared with those of the natural vegetation growth. Results showed that the restoration vegetation growth trend was considerably similar with that of the natural vegetation. Inter-annual restoration effects were analyzed by constructing the effect of the area-average factor and using vegetation growth data. Accordingly, the restoration vegetation effects were best in 2014 and 2016. Comprehensive restoration effect was analyzed using the weighted evaluation method to obtain the overall restoration effects of the coal mine. Results showed that the comprehensive restoration effect is inclined to the inferior growth state. This study conducted a preliminary evaluation of mine restoration vegetation, thereby providing a promising way for the future monitoring and evaluation of such processes.

Damage-Map Estimation Using UAV Images and Deep Learning Algorithms for Disaster Management System

Estimating the damaged area after a forest fire is important for responding to this natural catastrophe. With the support of aerial remote sensing, typically with unmanned aerial vehicles (UAVs), the aerial imagery of forest-fire areas can be easily obtained; however, retrieving the burnt area from the image is still a challenge. We implemented a new approach for segmenting burnt areas from UAV images using deep learning algorithms. First, the data were collected from a forest fire in Andong, the Republic of Korea, in April 2020. Then, the proposed two-patch-level deep-learning models were implemented. A patch-level 1 network was trained using the UNet++ architecture. The output prediction of this network was used as a position input for the second network, which used UNet. It took the reference position from the first network as its input and refined the results. Finally, the final performance of our proposed method was compared with a state-of-the-art image-segmentation algorithm to prove its robustness. Comparative research on the loss functions was also performed. Our proposed approach demonstrated its effectiveness in extracting burnt areas from UAV images and can contribute to estimating maps showing the areas damaged by forest fires.

Short-Term Forecasting of Satellite-Based Drought Indices Using Their Temporal Patterns and Numerical Model Output

Drought forecasting is essential for effectively managing drought-related damage and providing relevant drought information to decision-makers so they can make appropriate decisions in response to drought. Although there have been great efforts in drought-forecasting research, drought forecasting on a short-term scale (up to two weeks) is still difficult. In this research, drought-forecasting models on a short-term scale (8 days) were developed considering the temporal patterns of satellite-based drought indices and numerical model outputs through the synergistic use of convolutional long short term memory (ConvLSTM) and random forest (RF) approaches over a part of East Asia. Two widely used drought indices—Scaled Drought Condition Index (SDCI) and Standardized Precipitation Index (SPI)—were used as target variables. Through the combination of temporal patterns and the upcoming weather conditions (numerical model outputs), the overall performances of drought-forecasting models (ConvLSTM and RF combined) produced competitive results in terms of r (0.90 and 0.93 for validation SDCI and SPI, respectively) and nRMSE (0.11 and 0.08 for validation of SDCI and SPI, respectively). Furthermore, our short-term drought-forecasting model can be effective regardless of drought intensification or alleviation. The proposed drought-forecasting model can be operationally used, providing useful information on upcoming drought conditions with high resolution (0.05°).

Early Fire Detection Based on Aerial 360-Degree Sensors, Deep Convolution Neural Networks and Exploitation of Fire Dynamic Textures

The environmental challenges the world faces have never been greater or more complex. Global areas that are covered by forests and urban woodlands are threatened by large-scale forest fires that have increased dramatically during the last decades in Europe and worldwide, in terms of both frequency and magnitude. To this end, rapid advances in remote sensing systems including ground-based, unmanned aerial vehicle-based and satellite-based systems have been adopted for effective forest fire surveillance. In this paper, the recently introduced 360-degree sensor cameras are proposed for early fire detection, making it possible to obtain unlimited field of view captures which reduce the number of required sensors and the computational cost and make the systems more efficient. More specifically, once optical 360-degree raw data are obtained using an RGB 360-degree camera mounted on an unmanned aerial vehicle, we convert the equirectangular projection format images to stereographic images. Then, two DeepLab V3+ networks are applied to perform flame and smoke segmentation, respectively. Subsequently, a novel post-validation adaptive method is proposed exploiting the environmental appearance of each test image and reducing the false-positive rates. For evaluating the performance of the proposed system, a dataset, namely the “Fire detection 360-degree dataset”, consisting of 150 unlimited field of view images that contain both synthetic and real fire, was created. Experimental results demonstrate the great potential of the proposed system, which has achieved an F-score fire detection rate equal to 94.6%, hence reducing the number of required sensors. This indicates that the proposed method could significantly contribute to early fire detection.

Estimation of 10-Hour Fuel Moisture Content Using Meteorological Data: A Model Inter-Comparison Study

Forest fire modeling often requires estimates of fuel moisture status. Among the various fuel variables used for fire modeling studies, the 10-h fuel moisture content (10-h FMC) is a promising predictor since it can be automatically measured in real time at study sites, yielding more information for fire models. Here, the performance of 10-h FMC models based on three different approaches, including regression (MREG), machine learning algorithms (MML) with random forest and support vector machine, and a process-based model (MFSMM), were compared. In addition, whole-year models of each type were compared with their respective seasonal models to explore whether the development of separate seasonal models yielded better estimates. Meteorological conditions and 10-h FMC were measured each minute for 18 months in and near a forest site and used for constructing and examining the 10-h FMC models. In the assessments, MML showed the best performance (R2 = 0.77–0.82 and root mean squared error [RMSE] = 2.05–2.84%). The introduction of the correction coefficient into MREG improved its estimates (R2 improved from 0.56–0.58 to 0.68–0.70 and RMSE improved from 3.13–3.85% to 2.64–3.27%) by reducing the errors associated with high 10-h FMC values. MFSMM showed the worst performance (R2 = 0.41–0.43 and RMSE = 3.70–4.39%), which could possibly be attributed to the lack of radiation input from the study sites as well as the particular fuel moisture stick sensor that was used. Whole-year models and seasonal models showed almost equal performance because 10-h FMC varied in response to atmospheric moisture conditions rather than specific seasonal patterns. The adoption of a hybrid modeling approach that blends machine-learning and process-based approaches may yield better predictability and interpretability. This study provides additional evidence of the lagged response of 10-h FMC after rainfall, and suggests a new way of accounting for this response in a regression model. Our approach using comparisons among models can be utilized for other fire modeling studies, including those involving fire danger ratings.

Geolocation Accuracy Assessment of Himawari-8/AHI Imagery for Application to Terrestrial Monitoring

Recent advancements in new generation geostationary satellites have facilitated the application of their datasets to terrestrial monitoring. In this application, geolocation accuracy is an essential issue because land surfaces are generally heterogeneous. In the case of the Advanced Himawari Imager (AHI) onboard Himawari-8, geometric correction of the Himawari Standard Data provided by the Japan Meteorological Agency (JMA data) was conducted using thermal infrared band with 2 km spatial resolution. Based on JMA data, the Center for Environmental Remote Sensing (CEReS) at Chiba University applied a further geometric correction using a visible band with 500 m spatial resolution and released a dataset (CEReS data). JMA data target more general users mainly for meteorological observations, whereas CEReS data aim at terrestrial monitoring for more precise geolocation accuracy. The objectives of this study are to clarify the temporal and spatial variations of geolocation errors in these two datasets and assess their stability for unexpected large misalignment. In this study, the temporal tendencies of the relative geolocation difference between the two datasets were analyzed, and temporal fluctuations of band 3 reflectances of JMA data and CEReS data at certain fixed sites were investigated. A change in the geolocation trend and occasional shifts greater than 2 pixels were found in JMA data. With improved image navigation performance, the geolocation difference was decreased in CEReS data, suggesting the high temporal stability of CEReS data. Overall, JMA data showed an accuracy of less than 2 pixels with the spatial resolution of band 3. When large geolocation differences were observed, anomalies were also detected in the reflectance of JMA data. Nevertheless, CEReS data successfully corrected the anomalous errors and achieved higher geolocation accuracy in general. As CEReS data are processed during the daytime due to the availability of visible bands, we suggest the use of CEReS data for effective terrestrial monitoring during the daytime.

Advances in Remote Sensing-Based Disaster Monitoring and Assessment

Extreme weather/climate events have been increasing partly due to on-going climate change [...]