Fuel load is a crucial input in wildfire behavior models and a key parameter for the assessment of fire severity, fire flame length, and fuel consumption. Therefore, wildfire managers will benefit from accurate predictions of the spatiotemporal distribution of fuel load to inform strategic approaches to mitigate or prevent large-scale wildfires and respond to such incidents. Field surveys for fuel load assessment are labor-intensive, time-consuming, and as such, cannot be repeated frequently across large territories. On the contrary, remote-sensing sensors quantify fuel load in near-real time and at not only local but also regional or global scales. We reviewed the literature of the applications of remote sensing in fuel load estimation over a 12-year period, highlighting the capabilities and limitations of different remote-sensing sensors and technologies. While inherent technological constraints currently hinder optimal fuel load mapping using remote sensing, recent and anticipated developments in remote-sensing technology promise to enhance these capabilities significantly. The integration of remote-sensing technologies, along with derived products and advanced machine-learning algorithms, shows potential for enhancing fuel load predictions. Also, upcoming research initiatives aim to advance current methodologies by combining photogrammetry and uncrewed aerial vehicles (UAVs) to accurately map fuel loads at sub-meter scales. However, challenges persist in securing data for algorithm calibration and validation and in achieving the desired accuracies for surface fuels.

Improved forest management plans require a better understanding of wildfire risk and behavior to enhance the conservation of biodiversity and plan effective risk mitigation activities across the landscape. More particularly, for spatial fire hazard and risk assessing as well as fire intensity and growth modeling across a landscape, an adequate knowledge of the spatial distribution of key forest fuels attributes is required. Mapping fuel attributes is a challenging and complicated procedure because fuels are highly variable and complex. To simplify, classification schemes are used to summarize the large number of fuel attributes (e.g., height, density, continuity, arrangement, size, form, etc.) into fuel types which groups vegetation classes with a similar predicted fire behavior. Remote sensing is a cost-effective and objective technology that have been used to regularly map fuel types and have demonstrated greater success compared to traditional field surveys, especially with recent advancements in remote sensing data acquisition and fusion techniques. Thus, the main goal of this manuscript is to provide a comprehensive review of the recent remote sensing approaches used for fuel type classification. We build on findings from previous review manuscripts and focus on identifying the key challenges of different mapping approaches and the research gaps that still need to be filled in. To improve classification outcomes, more research into developing state-of-the-art deep learning algorithms with integrated remote sensing data sources is encouraged for future research. This review can be used as a guideline for practitioners, researchers, and decision-makers in the domain of fire management service.

One of the worst environmental catastrophes that endanger the Australian community is wildfire. To lessen potential fire threats, it is helpful to recognize fire occurrence patterns and identify fire susceptibility in wildfire-prone regions. The use of machine learning (ML) algorithms is acknowledged as one of the most well-known methods for addressing non-linear issues like wildfire hazards. It has always been difficult to analyze these multivariate environmental disasters because modeling can be influenced by a variety of sources of uncertainty, including the quantity and quality of training procedures and input variables. Moreover, although ML techniques show promise in this field, they are unstable for a number of reasons, including the usage of irrelevant descriptor characteristics when developing the models. Explainable AI (XAI) can assist us in acquiring insights into these constraints and, consequently, modifying the modeling approach and training data necessary. In this research, we describe how a Shapley additive explanations (SHAP) model can be utilized to interpret the results of a deep learning (DL) model that is developed for wildfire susceptibility prediction. Different contributing factors such as topographical, landcover/vegetation, and meteorological factors are fed into the model and various SHAP plots are used to identify which parameters are impacting the prediction model, their relative importance, and the reasoning behind specific decisions. The findings drawn from SHAP plots show the significant contributions made by factors such as humidity, wind speed, rainfall, elevation, slope, and normalized difference moisture index (NDMI) to the suggested model's output for wildfire susceptibility mapping. We infer that developing an explainable model would aid in comprehending the model's decision to map wildfire susceptibility, pinpoint high-contributing components in the prediction model, and consequently control fire hazards effectively.

The primary goal of this research is to see how effective cloud‐based computing services such as Google Earth Engine (GEE) platform are at classifying multitemporal satellite images and producing high‐quality land cover maps for the target year of 2020, with the possibility of using it on a larger‐scale area such as metropolitan Melbourne as a test site. To create high‐quality land cover maps, the GEE is utilized to analyze a total of 80 Landsat‐8 images. The support vector machine (SVM) approach is used to classify the images. Moreover, we use spectral bands, spectral indices, and topographic parameters to improve classification and address the limitations of existing approaches for classification with restricted input variables. Furthermore, we apply a postprocessing strategy to increase the model’s performance by removing the salt‐and‐pepper noise created by misclassified pixels in supervised classification results. The results demonstrate that given all parameters, the SVM approach achieves an overall accuracy (OA) and kappa accuracy of 88.47% and 85.34%, respectively. Following the implementation of the postprocessing technique, the OA and kappa improve to 92.90% and 90.99%, respectively. The results indicate that Landsat‐8 multitemporal data, spectral indices, topographic components, and postprocessing techniques are all important in land cover mapping. Therefore, the use of freely accessible GEE technology and multitemporal Landsat‐8 data ensures that decision makers have the resources they need to track land cover throughout the year.

Abstract Accurate vegetation analysis is crucial amid accelerating global changes and human activities. Achieving precise characterization with multi-temporal Sentinel-2 data is challenging. In this article, we present a comprehensive analysis of 2021's seasonal vegetation cover in Greater Sydney using Google Earth Engine (GEE) to process Sentinel-2 data. Using the random forest (RF) method, we performed image classification for vegetation patterns. Supplementary factors such as topographic elements, texture information, and vegetation indices enhanced the process and overcome limited input variables. Our model outperformed existing methods, offering superior insights into season-based vegetation dynamics. Multi-temporal Sentinel-2 data, topographic elements, vegetation indices, and textural factors proved to be critical for accurate analysis. Leveraging GEE and rich Sentinel-2 data, our study would benefit decision-makers involved in vegetation monitoring.

In the present work, a deep learning-based network called LeNet is applied for accurate grassland map production from Sentinel-2 data for the Greater Sydney region, Australia. First, we apply the technique to the base date Sentinel-2 data (non-seasonal) to make the vegetation maps. Then, we combine short time-series (seasonal) data and enhanced vegetation index (EVI) information to the base date imagery to improve the classification results and generate high-resolution grassland maps. The proposed model obtained an overall accuracy (OA) of 88.36% for the mono-temporal data, and 92.74% for the multi-temporal data. The experimental products proved that, by combining the short time-series images and EVI information to the base date, the classification maps' accuracy is increased by 4.38%. Moreover, the Sentinel-2 produced grassland maps are compared with the pre-existing maps such as Australian Land Use and Management (ALUM) 50 m resolution and Dynamic Land Cover Dataset (DLCD) with 250 m resolution as well as some traditional machine learning methods such as Support Vector Machine (SVM) and Random Forest (RF). The results show the effect of the LeNet network's performance and efficiency for grassland map production from short time-series data. As a result, decision-makers and urban planners can benefit from this work in terms of grassland change identification, monitoring, and planning assessment.

Building objects is one of the principal features that are essential for updating the geospatial database. Extracting building features from high-resolution imagery automatically and accurately is challenging because of the existence of some obstacles in these images, such as shadows, trees, and cars. Although deep learning approaches have shown significant improvements in the results of image segmentation in recent years, most deep neural networks still cannot achieve highly accurate results with correct segmentation map when processing high-resolution remote sensing images. Therefore, we implemented a new deep neural network named Seg–Unet method, which is a composition of Segnet and Unet techniques, to exploit building objects from high-resolution aerial imagery. Results obtained 92.73% accuracy carried on the Massachusetts building dataset. The proposed technique improved the performance to 0.44%, 1.17%, and 0.14% compared with fully convolutional neural network (FCN), Segnet, and Unet methods, respectively. Results also confirmed the superiority of the proposed method in building extraction.

Existing automated road extraction approaches concentrate on regional accuracy rather than road shape and connectivity quality. Most of these techniques produce discontinuous outputs caused by obstacles, such as shadows, buildings, and vehicles. This study proposes a shape and connectivity-preserving road identification deep learning-based architecture called SC-RoadDeepNet to overcome the discontinuous results and the quality of road shape and connectivity. The proposed model comprises a state-of-the-art deep learning-based network, namely, the recurrent residual convolutional neural network, boundary learning (BL), and a new measure based on the intersection of segmentation masks and their (morphological) skeleton called connectivity-preserving centerline Dice (CP_clDice). The recurrent residual convolutional layers accumulate low-level features for segmentation tasks, thus allowing for better feature representation. Such representation enables us to construct a UNet network with the same number of network parameters but improved segmentation effectiveness. BL also aids the model in improving the road’s boundaries by penalizing boundary misclassification and fine-tuning the road form. Furthermore, the CP_clDice method aids the model in maintaining road connectivity and obtaining accurate segmentations. We demonstrate that CP_clDice ensures connection preservation for binary segmentation, thereby allowing for efficient road network extraction at the end. The proposed model improves F1 score accuracy to 5.49%, 4.03%, 3.42%, and 2.27% compared with other comparative models, such as LinkNet, ResUNet, UNet, and VNet, respectively. Furthermore, qualitative and quantitative assessments demonstrate that the proposed SC-RoadDeepNet can improve road extraction by tackling shadow and occlusion-related interruptions. These assessments can also produce high-resolution results, particularly in the area of road network completeness.

Understanding urban dynamics, such as estimating population, urban development, and several other uses, necessitates up-to-date large-scale building maps. Since aerial imagery provides enough textural and structural details, it has been utilized as a critical data source for building detection. However, accurate mapping of building objects from aerial imagery is a challenging task. This problem is attributed due to presence of vegetation and shadows in images that present similar spectral values and transparency as a building class. To deal with the issues mentioned above, we offer a new deep-learning structure named MultiRes-UNet network, which is an improved version of the original UNet network. In the proposed network, we utilized the MultiRes block to assimilate the features learned from the data at various scales and comprise some more spatial details. Also, we suggest the incorporation of several convolutional operations along with the skip connections to mitigate the differences between the encode–decoder features. Furthermore, we integrated semantic edge information with semantic polygons to solve the issue of irregular semantic polygons and enhance the boundary of semantic polygons. We tested our network on aerial images for roof segmentation dataset, and the experimental results exhibited that the proposed network can improve the quantitative results of Intersection Over Union to 0.78% after adding semantic edges. We also used state-of-the-art comparative models such as UNet, DeeplabV3, ResNet, and FractalNet networks to show the competency of the introduced network, and the results prove the success of the introduced network for building object extraction from aerial imagery.

Terrestrial features extraction, such as roads and buildings from aerial images using an automatic system, has many usages in an extensive range of fields, including disaster management, change detection, land cover assessment, and urban planning. This task is commonly tough because of complex scenes, such as urban scenes, where buildings and road objects are surrounded by shadows, vehicles, trees, etc., which appear in heterogeneous forms with lower inter-class and higher intra-class contrasts. Moreover, such extraction is time-consuming and expensive to perform by human specialists manually. Deep convolutional models have displayed considerable performance for feature segmentation from remote sensing data in the recent years. However, for the large and continuous area of obstructions, most of these techniques still cannot detect road and building well. Hence, this work’s principal goal is to introduce two novel deep convolutional models based on UNet family for multi-object segmentation, such as roads and buildings from aerial imagery. We focused on buildings and road networks because these objects constitute a huge part of the urban areas. The presented models are called multi-level context gating UNet (MCG-UNet) and bi-directional ConvLSTM UNet model (BCL-UNet). The proposed methods have the same advantages as the UNet model, the mechanism of densely connected convolutions, bi-directional ConvLSTM, and squeeze and excitation module to produce the segmentation maps with a high resolution and maintain the boundary information even under complicated backgrounds. Additionally, we implemented a basic efficient loss function called boundary-aware loss (BAL) that allowed a network to concentrate on hard semantic segmentation regions, such as overlapping areas, small objects, sophisticated objects, and boundaries of objects, and produce high-quality segmentation maps. The presented networks were tested on the Massachusetts building and road datasets. The MCG-UNet improved the average F1 accuracy by 1.85%, and 1.19% and 6.67% and 5.11% compared with UNet and BCL-UNet for road and building extraction, respectively. Additionally, the presented MCG-UNet and BCL-UNet networks were compared with other state-of-the-art deep learning-based networks, and the results proved the superiority of the networks in multi-object segmentation tasks.

Road networks are one of the main urban features. Therefore, road parts extraction from high-resolution remotely sensed imagery and updated road database are beneficial for many GIS applications. However, owing to the presence of various types of obstacles in the images, such as shadows, cars, and trees, with similar transparency and spectral values as road class, achieving accurate road extraction using different classification and segmentation methods is still difficult. This paper proposes an integrated method combining segmentation and classification methods with connected components analysis to extract road class from orthophoto images. The proposed technique is threefold. First, multiresolution segmentation method was applied to segment images. Then, the main classification methods, namely, decision trees (DT), k-nearest neighbors (KNN), and support vector machines (SVM), were implemented based on spectral, geometric, and textural information to classify the obtained results into two classes: road and non-road. Three main accuracy evaluation measures, such as recall, precision, and F1-score, were evaluated to determine the performance of the proposed method, with respective average values of 87.62%, 89.71%, and 88.61%, respectively, for DT; 86.61%, 88.17%, and 87.30%, respectively, for KNN; and 89.83%, 89.52%, and 89.67%, respectively, for SVM. Finally, connected components labelling was used to extract road component parts, and morphological operation was employed to delete non-road parts and noises and improve the performance. These results were also compared with other prior works, which confirmed that the integrated method is an effective road extraction technique.

Urban vegetation mapping is critical in many applications, i.e., preserving biodiversity, maintaining ecological balance, and minimizing the urban heat island effect. It is still challenging to extract accurate vegetation covers from aerial imagery using traditional classification approaches, because urban vegetation categories have complex spatial structures and similar spectral properties. Deep neural networks (DNNs) have shown a significant improvement in remote sensing image classification outcomes during the last few years. These methods are promising in this domain, yet unreliable for various reasons, such as the use of irrelevant descriptor features in the building of the models and lack of quality in the labeled image. Explainable AI (XAI) can help us gain insight into these limits and, as a result, adjust the training dataset and model as needed. Thus, in this work, we explain how an explanation model called Shapley additive explanations (SHAP) can be utilized for interpreting the output of the DNN model that is designed for classifying vegetation covers. We want to not only produce high-quality vegetation maps, but also rank the input parameters and select appropriate features for classification. Therefore, we test our method on vegetation mapping from aerial imagery based on spectral and textural features. Texture features can help overcome the limitations of poor spectral resolution in aerial imagery for vegetation mapping. The model was capable of obtaining an overall accuracy (OA) of 94.44% for vegetation cover mapping. The conclusions derived from SHAP plots demonstrate the high contribution of features, such as Hue, Brightness, GLCM_Dissimilarity, GLCM_Homogeneity, and GLCM_Mean to the output of the proposed model for vegetation mapping. Therefore, the study indicates that existing vegetation mapping strategies based only on spectral characteristics are insufficient to appropriately classify vegetation covers.

Road network extraction from remotely sensed imagery has become a powerful tool for updating geospatial databases, owing to the success of convolutional neural network (CNN) based deep learning semantic segmentation techniques combined with the high-resolution imagery that modern remote sensing provides. However, most CNN approaches cannot obtain high precision segmentation maps with rich details when processing high-resolution remote sensing imagery. In this study, we propose a generative adversarial network (GAN)-based deep learning approach for road segmentation from high-resolution aerial imagery. In the generative part of the presented GAN approach, we use a modified UNet model (MUNet) to obtain a high-resolution segmentation map of the road network. In combination with simple pre-processing comprising edge-preserving filtering, the proposed approach offers a significant improvement in road network segmentation compared with prior approaches. In experiments conducted on the Massachusetts road image dataset, the proposed approach achieves 91.54% precision and 92.92% recall, which correspond to a Mathews correlation coefficient (MCC) of 91.13%, a Mean intersection over union (MIOU) of 87.43% and a F1-score of 92.20%. Comparisons demonstrate that the proposed GAN framework outperforms prior CNN-based approaches and is particularly effective in preserving edge information.

Two of the major applications in geospatial information system (GIS) and remote sensing fields are object detection and man-made feature extraction (e.g., road sections) from high-resolution remote sensing imagery. Extracting roads from high-resolution remotely sensed imagery plays a crucial role in multiple applications, such as navigation, emergency tasks, land cover change detection, and updating GIS maps. This study presents a deep learning technique based on a convolutional neural network (CNN) to classify and extract roads from orthophoto images. We applied the model on five orthophoto images to specify the superiority of the method for road extraction. First, we used principal component analysis and object-based image analysis for pre-processing to not only obtain spectral information but also add spatial and textural information for enhancing the classification accuracy. Then, the obtained results from the previous step were used as input for the CNN model to classify the images into road and non-road parts and trivial opening and closing operation are applied to extract connected road components from the images and remove holes inside the road parts. For the accuracy assessment of the proposed method, we used measurement factors such as precision, recall, F1 score, overall accuracy, and IOU. Achieved results showed that the average percentages of these factors were 91.09%, 95.32%, 93.15%, 94.44%, and 87.21%. The results were also compared with those of other existing methods. The comparison ascertained the reliability and superior performance of the suggested model architecture for extracting road regions from orthophoto images.Graphic Abstract

In this study, we present a new automatic deep learning-based network named Road Vectorization Network (RoadVecNet), which comprises interlinked UNet networks to simultaneously perform road segmentation and road vectorization. Particularly, RoadVecNet contains two UNet networks. The first network with powerful representation capability can obtain more coherent and satisfactory road segmentation maps even under a complex urban set-up. The second network is linked to the first network to vectorize road networks by utilizing all of the previously generated feature maps. We utilize a loss function called focal loss weighted by median frequency balancing (MFB_FL) to focus on the hard samples, fix the training data imbalance problem, and improve the road extraction and vectorization performance. A new module named dense dilated spatial pyramid pooling, which combines the benefit of cascaded modules with atrous convolution and atrous spatial pyramid pooling, is designed to produce more scale features over a broader range. Two types of high-resolution remote sensing datasets, namely, aerial and Google Earth imagery, were used for road segmentation and road vectorization tasks. Classification results indicate that the RoadVecNet outperforms the state-of-the-art deep learning-based networks with 92.51% and 93.40% F1 score for road surface segmentation and 89.24% and 92.41% F1 score for road vectorization from the aerial and Google Earth road datasets, respectively. In addition, the proposed method outperforms the other comparative methods in terms of qualitative results and produces high-resolution road segmentation and vectorization maps. As a conclusion, the presented method demonstrates that considering topological quality may result in improvement of the final road network, which is essential in various applications, such as GIS database updating.

One of the most challenging research subjects in remote sensing is feature extraction, such as road features, from remote sensing images. Such an extraction influences multiple scenes, including map updating, traffic management, emergency tasks, road monitoring, and others. Therefore, a systematic review of deep learning techniques applied to common remote sensing benchmarks for road extraction is conducted in this study. The research is conducted based on four main types of deep learning methods, namely, the GANs model, deconvolutional networks, FCNs, and patch-based CNNs models. We also compare these various deep learning models applied to remote sensing datasets to show which method performs well in extracting road parts from high-resolution remote sensing images. Moreover, we describe future research directions and research gaps. Results indicate that the largest reported performance record is related to the deconvolutional nets applied to remote sensing images, and the F1 score metric of the generative adversarial network model, DenseNet method, and FCN-32 applied to UAV and Google Earth images are high: 96.08%, 95.72%, and 94.59%, respectively.

This paper proposes a model to identify the changing of bare grounds into built-up or developed areas. The model is based on the fuzzy system and the Ordered Weighted Averaging (OWA) methods. The proposed model consists of four main sections, which include physical suitability, accessibility, the neighborhood effect, and a calculation of the overall suitability. In the first two parts, physical suitability and accessibility were obtained by defining fuzzy inference systems and applying the required map data associated with each section. However, in order to calculate the neighborhood effect, we used an enrichment factor method and a hybrid method consisting of the enrichment factor with the Few, Half, Most, and Majority quantifiers of the ordered weighted averaging (OWA) method. Finally, the three maps of physical suitability, accessibility, and the neighborhood effect were integrated by the fuzzy system method and the quantifiers of OWA to obtain the overall suitability maps. Then, the areas with high suitability were selected from the overall suitability map to be changed from bare ground into built-up areas. For this purpose, the proposed model was implemented and calibrated in the first period (2004–2010) and was evaluated by being applied to the second period (2010–2016). By comparing the estimated map of changes to the reference data and after the formation of the error matrix, it was determined that the OWA-Majority method has the best estimation compared to those of the other methods. Finally, the total accuracy and the Kappa coefficient for the OWA-Majority method in the second period were 98.98% and 98.98%, respectively, indicating this method’s high accuracy in predicting changes. In addition, the results were compared with those of other studies, which showed the effectiveness of the suggested method for urban development modeling.

Building extraction with high accuracy using semantic segmentation from high-resolution remotely sensed imagery has a wide range of applications like urban planning, updating of geospatial database, and disaster management. However, automatic building extraction with non-noisy segmentation map and obtaining accurate boundary information is a big challenge for most of the popular deep learning methods due to the existence of some barriers like cars, vegetation cover and shadow of trees in the high-resolution remote sensing imagery. Thus, we introduce an end-to-end convolutional neural network called Generative Adversarial Network (GAN) in this study to tackle these issues. In the generative model, we utilized SegNet model with Bi-directional Convolutional LSTM (BConvLSTM) to generate the segmentation map from Massachusetts building dataset containing high-resolution aerial imagery. BConvLSTM combines encoded features (containing of more local information) and decoded features (containing of more semantic information) to improve the performance of the model even with the presence of complex backgrounds and barriers. The adversarial training method enforces long-range spatial label vicinity to tackle with the issue of covering building objects with the existing occlusions such as trees, cars and shadows and achieve high-quality building segmentation outcomes under the complex areas. The quantitative results obtained by the proposed technique with an average F1-score of 96.81% show that the suggested approach could achieve better results through detecting and adjusting the difference between the segmentation model output and the reference map compared to other state-of-the-art approaches such as autoencoder method with 91.36%, SegNet+BConvLSTM with 95.96%, FCN-CRFs with 95.36%% SegNet with 94.77%, and GAN-SCA model with 96.36% accuracy.

One of the most important tasks in the advanced transportation systems is road extraction. Extracting road region from high-resolution remote sensing imagery is challenging due to complicated background such as buildings, trees shadows, pedestrians and vehicles and rural road networks that have heterogeneous forms with low interclass and high intraclass differences. Recently, deep learning-based techniques have presented a notable enhancement in the image segmentation results, however, most of them still cannot preserve boundary information and obtain high-resolution road segmentation map when processing the remote sensing imagery. In the present study, we introduce a new deep learning-based convolutional network called VNet model to produce a high-resolution road segmentation map. Moreover, a new dual loss function called cross-entropy-dice-loss (CEDL) is defined that synthesize cross-entropy (CE) and dice loss (DL) and consider both local information (CE) and global information (DL) to decrease the class imbalance influence and improve the road extraction results. The proposed VNet+CEDL model is implemented on two various road datasets called Massachusetts and Ottawa datasets. The suggested VNet+CEDL approach achieved an average F1 accuracy of 90.64% for Massachusetts dataset and 92.41% for Ottawa dataset. When compared to other state-of-the-art deep learning-based frameworks like FCN, Segnet and Unet, the proposed approach could improve the results to 1.09%, 2.45% and 0.39%, for Massachusetts dataset and 7.21%, 1.86% and 2.68%, for Ottawa dataset. Also, we compared the proposed method with the state-of-the-art road extraction techniques, and the results proved that the proposed technique outperformed other deep learning-based techniques in road extraction.

Desertification is a natural phenomenon that threatens the biomass of the world in various forms, and the adverse effects of this phenomenon can be observed in different parts of the planet. Some events and complications of the earth's surface, such as vegetation coverage, have changed over time due to natural or human factors, thereby affecting the ecosystem's condition and performance. Vegetation coverage is a critical factor in the assessment of desertification, and continuous production of accurate vegetation maps is an important tool for monitoring natural resources and the environment. Therefore, this paper used MODIS images to investigate the vulnerability of the Mashhad regions in Iran to desertification according to the enhanced vegetation index (EVI) for various periods. Experimental results showed that the Mashhad regions had the highest vulnerability to desertification during 2001–2005, given the highest variation in the EVI in this period.