Search

Wildfires play a crucial role in the transformation of forest ecosystems and exert a significant influence on the global climate over geological timescales. Recent shifts in climate patterns and intensified human–forest interactions have led to an increase in the incidence of wildfires. These fires are characterized by their extensive coverage, higher frequency, and prolonged duration, rendering them increasingly destructive. To mitigate the impact of wildfires on climate change, ecosystems, and biodiversity, it is imperative to conduct systematic monitoring of wildfire progression and evaluate their environmental repercussions on a global scale. Satellite remote sensing is a powerful tool, offering precise and timely data on terrestrial changes, and has been extensively utilized for wildfire identification, tracking, and impact assessment at both local and regional levels. The Canada Centre for Mapping and Earth Observation, in collaboration with the Canadian Forest Service, has developed a comprehensive National Burned Area Composite (NBAC). This composite serves as a benchmark for curating a bi-temporal multi-source satellite image dataset for change detection, compiled from the archives of Sentinel-2, Sentinel-1, and ALOS-2 PALSAR-2. To our knowledge, this dataset is the inaugural large-scale, multi-source, and multi-frequency satellite image dataset with 20 m spatial resolution for wildfire mapping, monitoring, and evaluation. It harbors significant potential for enhancing wildfire management strategies, building upon the profound advancements in deep learning that have contributed to the field of remote sensing. Based on our curated dataset, which encompasses major wildfire events in Canada, we conducted a systematic evaluation of the capability of multi-source satellite earth observation data in identifying wildfire-burned areas using statistical analysis and deep learning. Our analysis compares the difference between burned and unburned areas using post-event observation solely or bi-temporal (pre- and post-event) observations across diverse land cover types. We demonstrate that optical satellite data yield higher separability than C-Band and L-Band Synthetic Aperture Radar (SAR), which exhibit considerable overlap in burned and unburned sample distribution, as evidenced by SAR-based boxplots. With U-Net, we further explore how different input channels influence the detection accuracy. Our findings reveal that deep neural networks enhance SAR's performance in mapping burned areas. Notably, C-Band SAR shows a higher dependency on pre-event data than L-Band SAR for effective detection. A comparative analysis of U-Net and its variants indicates that U-Net works best with single-sensor data, while the late fusion architecture marginally surpasses others in the fusion of optical and SAR data. Accuracy across sensors is highest in closed forests, with sequentially lower performance in open forests, shrubs, and grasslands. Future work will extend the data from both spatial and temporal dimensions to encompass varied vegetation types and climate zones, furthering our understanding of multi-source and multi-frequency satellite remote sensing capabilities in wildfire detection and monitoring. [ABSTRACT FROM AUTHOR]

Deep convolutional neural networks (CNNs) have achieved tremendous successes but tend to suffer from high computation costs mainly due to heavy over-parameterization, resulting in the difficulty of directly applying them to the ever-growing application demands based on low-end edge devices with strong power restriction and real-time inference requirement. Recently, there has much research attention devoted to compressing the network via pruning to address this issue. Most of the existing methods rely on some hand-designed pruning rules, which suffer from several limitations. Firstly, manually designed rules are only applicable to limited application scenarios, which can hardly generalize well in a broader scope. And these rules are typically designed based on human experience and via trial and error, and thus highly subjective. Then, channels of different layers in a network may have diverse distributions, which means the same pruning rule is not appropriate for each layer. To address these limitations, we propose a novel channel pruning scheme, in which the task-irrelevant channels are removed in a task-driven manner. Specifically, an adaptively differentiable search module is proposed to find the best pruning rule automatically for different layers in CNNs under sparsity constraints. Besides, we employed knowledge distillation to alleviate the excessive performance loss. Once the training process is finished, a compact network will be obtained by removing channels based on layer-wise pruning rules. We have evaluated the proposed method on some well-known benchmark datasets including CIFAR, MNIST, and ImageNet in comparison to several state-of-the-art pruning methods. Experimental results demonstrate the superiority of our method over the compared ones in terms of both parameters and FLOPs reduction. [ABSTRACT FROM AUTHOR]

The few-shot scene classification is dedicated to identifying unseen remote sensing classes when only a very small number of labeled samples are available for reference. Most of the existing few-shot scene classification methods are based on meta-learning and use the episodic learning for training, which lacks the consideration for the utilization of data efficiency. In this article, instead of designing sophisticated meta-learning-based algorithms, we are committed to training a feature extractor with good generalization performance and strong feature extraction capability. Specifically, we propose a novel two-path aggregation attention network with quad-patch data augmentation, called data architecture network (DANet), to solve the problem of few-shot scene classification from both data and architecture aspects. In terms of data, we design a new data augmentation strategy named quad-patch augmentation. We use the characteristics of remote sensing images to chunk and reassemble any existing data, thereby generating pseudo-new data to enrich the training set. In terms of architecture, we present a two-path aggregation attention module that makes it easier for the model to focus on the key clues in a targeted manner. The comparative experiments in natural image datasets and remote sensing image datasets demonstrate the effectiveness of our two innovations. In addition, DANet achieves competitive or state-of-the-art (SOTA) results on three benchmark scene classification datasets. [ABSTRACT FROM AUTHOR]

Unsupervised change detection based on supervised or semisupervised classifiers has achieved strong adaptability and robustness to obtain satisfactory change detection results. However, these methods suffer from an issue that it is hard to collect reliable training samples in an unsupervised manner. In this article, a group self-paced learning (GSPL) framework is proposed to mine the reliable training samples. In the proposed method, each sample is assigned a weight to indicate its reliability. The proposed scheme is able to learn the weighted samples and update the weights iteratively in a self-paced manner to identify the reliable training samples. In the phase of updating weights, a grouping strategy is designed to avoid selecting training samples from homogeneous regions. Furthermore, a novel time-varying self-paced regularizer is proposed to automatically determine the learning scheme of self-paced learning. Finally, three classifiers, including SoftMax, backpropagation neural network, and support vector machine, are investigated under this proposed framework. Experiments on five change detection data sets demonstrate that the proposed framework can significantly outperform those state-of-art methods for change detection in terms of accuracy and robustness. [ABSTRACT FROM AUTHOR]

Abstract This paper presents a deep associative neural network (DANN) based on unsupervised representation learning for associative memory. In brain, the knowledge is learnt by associating different types of sensory data, such as image and voice. The associative memory models which imitate such a learning process have been studied for decades but with simpler architectures they fail to deal with large scale complex data as compared with deep neural networks. Therefore, we define a deep architecture consisting of a perception layer and hierarchical propagation layers. To learn the network parameters, we define a probabilistic model for the whole network inspired from unsupervised representation learning models. The model is optimized by a modified contrastive divergence algorithm with a novel iterated sampling process. After training, given a new data or corrupted data, the correct label or corrupted part is associated by the network. The DANN is able to achieve many machine learning problems, including not only classification, but also depicting the data given a label and recovering corrupted images. Experiments on MNIST digits and CIFAR-10 datasets demonstrate the learning capability of the proposed DANN. Highlights • We propose a deep associative neural network for associative memory. • The network parameters are learnt by establishing a probability model. • The probability model aims to make the network best capture the data distributions. • The model is optimized by an improved contrastive divergence algorithm. • A novel sampling method is proposed to sample from the deep architecture based model. • The network can achieve several pattern recognition tasks with only one training. [ABSTRACT FROM AUTHOR]

• We analyze the limitations of directly applying the high-order multi-hop neighbors to graph neural networks, including computational inefficiency and limited representation ability of the multi-hop neighbor. • We propose a Markov Clustering Regularized Graph Neural Network (MCMGN) to overcome the limitations. By introducing iterative approximation and Regularized Markov Clustering strategies, the proposed model improves the performance of the multi-hop graph neural network for graph-level representation learning. • Extensive experimental results have verified the effectiveness of the proposed Markov Clustering Regularized Graph Neural Network for graph classification on eight graph benchmark datasets. • In this paper, we develop a Markov Clustering Regularized Multi-hop Graph Neural Network (MCMGN) to overcome the two limitations of higher-order multi-hop graph neural network, i.e., computational inefficiency and limited representation ability of multi-hop neighbor. We analyze the reasons of these two limitations in detail and give the corresponding solutions, i.e., iteration approximation and Markov Clustering regularized approaches, to achieve a excellent performance for graph-level representation learning. To evaluate the effectiveness of the proposed model, we carry out a series of experiments including the comparison with existing models, prune threshold comparison, impact of the number of iteration, impact of the normalization of node feature matrix, and computation cost comparison. Extensive experimental results strongly demonstrate the superiority of MCMGN in predictive performance for graph-level representation learning. [Display omitted] Graph Neural Networks (GNNs) have shown great potential for graph data analysis. In this paper, we focus on multi-hop graph neural networks and aim to extend existing models to a high-order multi-hop form for graph-level representation learning. However, such a directly extending method suffers from two limitations, i.e., computational inefficiency and limited representation ability of the multi-hop neighbor. For the former limitation, we utilize an iteration approach to approximate the power of a complex adjacency matrix to achieve linear computational complexity. For the latter limitation, we introduce the Regularized Markov Clustering (R-MCL) to regularize the flow matrix, i.e., the adjacency matrix, in each iteration step. With these two strategies, we construct Markov Clustering Regularized Multi-hop Graph Neural Network (MCMGN) for graph-level representation learning tasks. Specifically, MCMGN consists of a multi-hop message passing phase and a readout phase, where the multi-hop message passing phase aims to learn multi-hop node embedding, and then the readout phase aggregates multi-hop node representations to generate graph embedding for graph-level representation learning tasks. Extensive experiments on eight graph benchmark datasets strongly demonstrate the effectiveness of Markov Clustering Regularized Multi-hop Graph Neural Network, leading to superior performance on graph classification. [ABSTRACT FROM AUTHOR]

Ternary change detection aims to detect changes and group the changes into positive change and negative change. It is of great significance in the joint interpretation of spatial-temporal synthetic aperture radar images. In this study, sparse autoencoder, convolutional neural networks (CNN) and unsupervised clustering are combined to solve ternary change detection problem without any supervison. Firstly, sparse autoencoder is used to transform log-ratio difference image into a suitable feature space for extracting key changes and suppressing outliers and noise. And then the learned features are clustered into three classes, which are taken as the pseudo labels for training a CNN model as change feature classifier. The reliable training samples for CNN are selected from the feature maps learned by sparse autoencoder with certain selection rules. Having training samples and the corresponding pseudo labels, the CNN model can be trained by using back propagation with stochastic gradient descent. During its training procedure, CNN is driven to learn the concept of change, and more powerful model is established to distinguish different types of changes. Unlike the traditional methods, the proposed framework integrates the merits of sparse autoencoder and CNN to learn more robust difference representations and the concept of change for ternary change detection. Experimental results on real datasets validate the effectiveness and superiority of the proposed framework. [ABSTRACT FROM AUTHOR]

With the rapid technological development of various satellite sensors, high-resolution remotely sensed imagery has been an important source of data for change detection in land cover transition. However, it is still a challenging problem to effectively exploit the available spectral information to highlight changes. In this paper, we present a novel change detection framework for high-resolution remote sensing images, which incorporates superpixel-based change feature extraction and hierarchical difference representation learning by neural networks. First, highly homogenous and compact image superpixels are generated using superpixel segmentation, which makes these image blocks adhere well to image boundaries. Second, the change features are extracted to represent the difference information using spectrum, texture, and spatial features between the corresponding superpixels. Third, motivated by the fact that deep neural network has the ability to learn from data sets that have few labeled data, we use it to learn the semantic difference between the changed and unchanged pixels. The labeled data can be selected from the bitemporal multispectral images via a preclassification map generated in advance. And then, a neural network is built to learn the difference and classify the uncertain samples into changed or unchanged ones. Finally, a robust and high-contrast change detection result can be obtained from the network. The experimental results on the real data sets demonstrate its effectiveness, feasibility, and superiority of the proposed technique. [ABSTRACT FROM PUBLISHER]

In this paper, we establish a deep neural network using stacked Restricted Boltzmann Machines (RBMs) to analyze the difference images and detect changes between multitemporal synthetic aperture radar (SAR) images. Given the two multitemporal images, a difference image which shows difference degrees between corresponding pixels is generated. Then, RBMs are stacked to form a deep hierarchical neural network to learn to analyze the difference image and recognize the changed pixels and unchanged pixels. The learning process includes unsupervised layer-wise feature learning and supervised fine-tuning of network parameters. Unsupervised learning aims to learn the representation of the difference image. Supervised fine-tuning aims to learn to classify the changed and unchanged pixels. The network can learn from datasets that have few labeled data. The labeled data can be selected from the results obtained by other methods because there is no prior information in image change detection. The system learns to detect the changes instead of recognizing the changes by fixed equations as in traditional change detection algorithms. We test the network with real synthetic aperture radar datasets and the labeled samples are extracted from the results obtained, respectively, by several methods, including a thresholding method, a level set method and two clustering methods. The results achieved by the trained network outperform that of other methods. [ABSTRACT FROM AUTHOR]

Due to the fact that weak and fake laser-induced damages may occur in the surface of optical elements in high-energy laser facilities, it is still a challenging issue to effectively detect the real laser-induced damage changes of optical elements in optical images. Different from the traditional methods, in this paper, we put forward a similarity metric optimization driven supervised learning model to perform the laser-induced damage change detection task. In the proposed model, an end-to-end siamese convolutional neural network is designed and trained which can integrate the difference image generating and difference image analysis into a whole network. Thus, the damage changes can be highlighted by the pre-trained siamese network that classifies the central pixel between input multi-temporal image patches into changed and unchanged classes. To address the problem of unbalanced distribution between positive and negative samples, a modified average frequency balancing based weighted softmax loss is used to train the proposed network. Experiments conducted on two real datasets demonstrate the effectiveness and superiority of the proposed model. • This article presents a novel model for damage change detection of optical elements. • A similarity metric optimization driven deep learning model is proposed. • An end-to-end siamese CNN is designed for the damage change detection task. • A modified weighted softmax loss is utilized to train the network. [ABSTRACT FROM AUTHOR]

With the increasing resolution of optical remote sensing images, ship detection in optical remote sensing images has attracted a lot of research interests. The current ship detection methods usually adopt the coarse-to-fine detection strategy, which firstly extracts low-level and manual features, and then performs multi-step training. Inadequacies of this strategy are that it would produce complex calculation, false detection on land and difficulty in detecting the small size ship. Aiming at these problems, a sea-land separation algorithm that combines gradient information and gray information is applied to avoid false alarms on land, the feature pyramid network (FPN) is used to achieve small ship detection, and a multi-scale detection strategy is proposed to achieve ship detection with different degrees of refinement. Then the feature extraction structure is adopted to fuse different hierarchical features to improve the representation ability of features. Finally, we propose a new coarse-to-fine ship detection network (CF-SDN) that directly achieves an end-to-end mapping from image pixels to bounding boxes with confidences. A coarse-to-fine detection strategy is applied to improve the classification ability of the network. Experimental results on optical remote sensing image set indicate that the proposed method outperforms the other excellent detection algorithms and achieves good detection performance on images including some small-sized ships and dense ships near the port. [ABSTRACT FROM AUTHOR]

Multi-view graph embedding is aimed at learning low-dimensional representations of nodes that capture various relationships in a multi-view network, where each view represents a type of relationship among nodes. Multitudes of existing graph embedding approaches concentrate on single-view networks, that can only characterize one simple type of proximity relationships among objects. However, most of the real-world complex systems possess multiple types of relationships among entities. In this paper, a novel approach of graph embedding for multi-view networks is proposed, named Multi-view Graph Attention Networks (MGAT). We explore an attention-based architecture for learning node representations from each single view, the network parameters of which are constrained by a novel regularization term. In order to collaboratively integrate multiple types of relationships in different views, a view-focused attention method is explored to aggregate the view-wise node representations. We evaluate the proposed algorithm on several real-world datasets, and it demonstrates that the proposed approach outperforms existing state-of-the-art baselines. [ABSTRACT FROM AUTHOR]

Coupled networks are extremely fragile because a node failure of a network would trigger a cascade of failures on the entire system. Existing studies mainly focused on the cascading failures and the robustness of coupled networks when the networks suffer from attacks. In reality, it is necessary to recover the damaged networks, and there are cascading failures in recovery processes. In this study, firstly, we analyze the cascading failures of coupled networks during recoveries. Then, a recovery robustness index is presented for evaluating the resilience of coupled networks to cascading failures in the recovery processes. Finally, we propose a technique aiming at protecting several influential nodes for enhancing robustness of coupled networks under the recoveries, and adopt six strategies based on the potential knowledge of network centrality to find the influential nodes. Experiments on three coupling networks demonstrate that with a small number of influential nodes protected, the robustness of coupled networks under the recoveries can be greatly enhanced. [ABSTRACT FROM AUTHOR]