Change detection in synthetic aperture radar images based on image fusion and fuzzy clustering

Commonality Feature Representation Learning for Unsupervised Multimodal Change Detection

The main challenge of multimodal change detection (MCD) is that multimodal bitemporal images (MBIs) cannot be compared directly to identify changes. To overcome this problem, this paper proposes a novel commonality feature representation learning (CFRL) and constructs a CFRL-based unsupervised MCD framework. The CFRL is composed of a Siamese-based encoder and two decoders. First, the Siamese-based encoder can map original MBIs in the same feature space for extracting the representative features of each modality. Then, the two decoders are used to reconstruct the original MBIs by regressing themselves, respectively. Meanwhile, we swap the decoders to reconstruct the pseudo-MBIs to conduct modality alignment. Subsequently, all reconstructed images are input to the Siamese-based encoder again to map them in a same feature space, by which representative features are obtained. On this basis, latent commonality features between MBIs can be extracted by minimizing the distance between these representative features. These latent commonality features are comparable and can be used to identify changes. Notably, the proposed CFRL can be performed simultaneously in two modalities corresponding to MBIs. Therefore, two change magnitude images (CMIs) can be generated simultaneously by measuring the difference between the commonality features of MBIs. Finally, a simple threshold algorithm or a clustering algorithm can be employed to divide CMIs into binary change maps. Extensive experiments on six publicly available MCD datasets show that the proposed CFRL-based framework can achieve superior performance compared with other state-of-the-art approaches.

A survey on the utilization of Superpixel image for clustering based image segmentation

Superpixel become increasingly popular in image segmentation field as it greatly helps image segmentation techniques to segment the region of interest accurately in noisy environment and also reduces the computation effort to a great extent. However, selection of proper superpixel generation techniques and superpixel image segmentation techniques play a very crucial role in the domain of different kinds of image segmentation. Clustering is a well-accepted image segmentation technique and proved their effective performance over various image segmentation field. Therefore, this study presents an up-to-date survey on the employment of superpixel image in combined with clustering techniques for the various image segmentation. The contribution of the survey has four parts namely (i) overview of superpixel image generation techniques, (ii) clustering techniques especially efficient partitional clustering techniques, their issues and overcoming strategies, (iii) Review of superpixel combined with clustering strategies exist in literature for various image segmentation, (iv) lastly, the comparative study among superpixel combined with partitional clustering techniques has been performed over oral pathology and leaf images to find out the efficacy of the combination of superpixel and partitional clustering approaches. Our evaluations and observation provide in-depth understanding of several superpixel generation strategies and how they apply to the partitional clustering method.

Changed Detection Based on Patch Robust Principal Component Analysis

Change detection on retinal fundus image pairs mainly seeks to compare the important differences between a pair of images obtained at two different time points such as in anatomical structures or lesions. Illumination variation usually challenges the change detection methods in many cases. Robust principal component analysis (RPCA) takes intensity normalization and linear interpolation to greatly reduce the illumination variation between the continuous frames and then decomposes the image matrix to obtain the robust background model. The matrix-RPCA can obtain clear change regions, but when there are local bright spots on the image, the background model is vulnerable to illumination, and the change detection results are inaccurate. In this paper, a patch-based RPCA (P-RPCA) is proposed to detect the change of fundus image pairs, where a pair of fundus images is normalized and linearly interpolated to expand a low-rank image sequence; then, images are divided into many patches to obtain an image-patch matrix, and finally, the change regions are obtained by the low-rank decomposition. The proposed method is validated on a set of large lesion image pairs in clinical data. The area under curve (AUC) and mean average precision (mAP) of the method proposed in this paper are 0.9832 and 0.8641, respectively. For a group of small lesion image pairs with obvious local illumination changes in clinical data, the AUC and mAP obtained by the P-RPCA method are 0.9893 and 0.9401, respectively. The results show that the P-RPCA method is more robust to local illumination changes than the RPCA method, and has stronger performance in change detection than the RPCA method.

Novel Higher-Order Clique Conditional Random Field to Unsupervised Change Detection for Remote Sensing Images

Change detection (CD) is one of the most important topics in remote sensing. In this paper, we propose a novel higher-order clique conditional random field model to unsupervised CD for remote sensing images (termed HOC2RF), by defining a higher-order clique potential. The clique potential, constructed based on a well-designed higher-order clique of image objects, takes the interaction between the neighboring objects in both feature and location spaces into account. HOC2RF consists of five principle steps: (1) Two difference images with complementary change information are produced by change vector analysis and using the spectral correlation mapper, which describe changes from the perspective of the vector magnitude and angle, respectively. (2) The fuzzy partition matrix of each difference image is calculated by fuzzy clustering, and the fused partition matrix is obtained by fusing the calculated partition matrices with evidence theory. (3) An object-level map is created by segmenting the difference images with an adaptive morphological reconstruction based watershed algorithm. (4) The energy function of the proposed HOC2RF, composed of unary, pairwise, and higher-order clique potentials, is computed based on the difference images, the fusion partition matrix, and the object-level map. (5) The energy function is minimized by the graph cut algorithm to achieve the binary CD map. The proposed HOC2RF CD approach combines the complementary change information extracted from the perspectives of vector magnitude and angle, and synthetically exploits the pixel-level and object-level spatial correlation of images. The main contributions of this article include: (1) proposing the idea of using the interaction between neighboring objects in both feature and location spaces to enhance the CD performance; and (2) presenting a method to construct a higher-order clique of objects, developing a higher-order clique potential function, and proposing a novel CD method HOC2RF. In the experiments on three real remote sensing images, the Kappa coefficient/overall accuracy values of the proposed HOC2RF are 0.9655/0.9967, 0.9518/0.9910, and 0.7845/0.9651, respectively, which are superior to some state-of-the-art CD methods. The experimental results confirm the effectiveness of the proposed method.

Change Detection Based on Fusion Difference Image and Multi-Scale Morphological Reconstruction for SAR Images

Synthetic aperture radar (SAR) image-change detection is widely used in various fields, such as environmental monitoring and ecological monitoring. There is too much noise and insufficient information utilization, which make the results of change detection inaccurate. Thus, we propose an SAR image-change-detection method based on multiplicative fusion difference image (DI), saliency detection (SD), multi-scale morphological reconstruction (MSMR), and fuzzy c-means (FCM) clustering. Firstly, a new fusion DI method is proposed by multiplying the ratio (R) method based on the ratio of the image before and after the change and the mean ratio (MR) method based on the ratio of the image neighborhood mean value. The new DI operator ratio–mean ratio (RMR) enlarges the characteristics of unchanged areas and changed areas. Secondly, saliency detection is used in DI, which is conducive to the subsequent sub-area processing. Thirdly, we propose an improved FCM clustering-change-detection method based on MSMR. The proposed method has high computational efficiency, and the neighborhood information obtained by morphological reconstruction is fully used. Six real SAR data sets are used in different experiments to demonstrate the effectiveness of the proposed saliency ratio–mean ratio with multi-scale morphological reconstruction fuzzy c-means (SRMR-MSMRFCM). Finally, four classical noise-sensitive methods are used to detect our DI method and demonstrate the strong denoising and detail-preserving ability.

Change Detection Techniques with Synthetic Aperture Radar Images: Experiments with Random Forests and Sentinel-1 Observations

This work aims to clarify the potential of incoherent and coherent change detection (CD) approaches for detecting and monitoring ground surface changes using sequences of synthetic aperture radar (SAR) images. Nowadays, the growing availability of remotely sensed data collected by the twin Sentinel-1A/B sensors of the European (EU) Copernicus constellation allows fast mapping of damage after a disastrous event using radar data. In this research, we address the role of SAR (amplitude) backscattered signal variations for CD analyses when a natural (e.g., a fire, a flash flood, etc.) or a human-induced (disastrous) event occurs. Then, we consider the additional pieces of information that can be recovered by comparing interferometric coherence maps related to couples of SAR images collected between a principal disastrous event date. This work is mainly concerned with investigating the capability of different coherent/incoherent change detection indices (CDIs) and their mutual interactions for the rapid mapping of “changed” areas. In this context, artificial intelligence (AI) algorithms have been demonstrated to be beneficial for handling the different information coming from coherent/incoherent CDIs in a unique corpus. Specifically, we used CDIs that synthetically describe ground surface changes associated with a disaster event (i.e., the pre-, cross-, and post-disaster phases), based on the generation of sigma nought and InSAR coherence maps. Then, we trained a random forest (RF) to produce CD maps and study the impact on the final binary decision (changed/unchanged) of the different layers representing the available synthetic CDIs. The proposed strategy was effective for quickly assessing damage using SAR data and can be applied in several contexts. Experiments were conducted to monitor wildfire’s effects in the 2021 summer season in Italy, considering two case studies in Sardinia and Sicily. Another experiment was also carried out on the coastal city of Houston, Texas, the US, which was affected by a large flood in 2017; thus, demonstrating the validity of the proposed integrated method for fast mapping of flooded zones using SAR data.

Change detection based on unsupervised sparse representation for fundus image pair

Detecting changes is an important issue for ophthalmology to compare longitudinal fundus images at different stages and obtain change regions. Illumination variations bring distractions on the change regions by the pixel-by-pixel comparison. In this paper, a new unsupervised change detection method based on sparse representation classification (SRC) is proposed for the fundus image pair. First, the local neighborhood patches are extracted from the reference image to build a dictionary of the local background. Then the current image patch is represented sparsely and its background is reconstructed by the obtained dictionary. Finally, change regions are given through background subtracting. The SRC method can correct automatically illumination variations through the representation coefficients and filter local contrast and global intensity effectively. In experiments of this paper, the AUC and mAP values of SRC method are 0.9858 and 0.8647 respectively for the image pair with small lesions; the AUC and mAP values of the fusion method of IRHSF and SRC are 0.9892 and 0.9692 separately for the image pair with the big change region. Experiments show that the proposed method in this paper is more robust than RPCA for the illumination variations and can detect change regions more effectively than pixel-wised image differencing.

MSGATN: A Superpixel-Based Multi-Scale Siamese Graph Attention Network for Change Detection in Remote Sensing Images

With the rapid development of Earth observation technology, how to effectively and efficiently detect changes in multi-temporal images has become an important but challenging problem. Relying on the advantages of high performance and robustness, object-based change detection (CD) has become increasingly popular. By analyzing the similarity of local pixels, object-based CD aggregates similar pixels into one object and takes it as the basic processing unit. However, object-based approaches often have difficulty capturing discriminative features, as irregular objects make processing difficult. To address this problem, in this paper, we propose a novel superpixel-based multi-scale Siamese graph attention network (MSGATN) which can process unstructured data natively and extract valuable features. First, a difference image (DI) is generated by Euclidean distance between bitemporal images. Second, superpixel segmentation is employed based on DI to divide each image into many homogeneous regions. Then, these superpixels are used to model the problem by graph theory to construct a series of nodes with the adjacency between them. Subsequently, the multi-scale neighborhood features of the nodes are extracted through applying a graph convolutional network and concatenated by an attention mechanism. Finally, the binary change map can be obtained by classifying each node by some fully connected layers. The novel features of MSGATN can be summarized as follows: (1) Training in multi-scale constructed graphs improves the recognition over changed land cover of varied sizes and shapes. (2) Spectral and spatial self-attention mechanisms are exploited for a better change detection performance. The experimental results on several real datasets show the effectiveness and superiority of the proposed method. In addition, compared to other recent methods, the proposed can demonstrate very high processing efficiency and greatly reduce the dependence on labeled training samples in a semisupervised training fashion.

Few-Shot Learning with Collateral Location Coding and Single-Key Global Spatial Attention for Medical Image Classification

Humans are born with the ability to learn quickly by discerning objects from a few samples, to acquire new skills in a short period of time, and to make decisions based on limited prior experience and knowledge. The existing deep learning models for medical image classification often rely on a large number of labeled training samples, whereas the fast learning ability of deep neural networks has failed to develop. In addition, it requires a large amount of time and computing resource to retrain the model when the deep model encounters classes it has never seen before. However, for healthcare applications, enabling a model to generalize new clinical scenarios is of great importance. The existing image classification methods cannot explicitly use the location information of the pixel, making them insensitive to cues related only to the location. Besides, they also rely on local convolution and cannot properly utilize global information, which is essential for image classification. To alleviate these problems, we propose a collateral location coding to help the network explicitly exploit the location information of each pixel to make it easier for the network to recognize cues related to location only, and a single-key global spatial attention is designed to make the pixels at each location perceive the global spatial information in a low-cost way. Experimental results on three medical image benchmark datasets demonstrate that our proposed algorithm outperforms the state-of-the-art approaches in both effectiveness and generalization ability.

Adaptive Contourlet Fusion Clustering for SAR Image Change Detection

In this paper, a novel unsupervised change detection method called adaptive Contourlet fusion clustering based on adaptive Contourlet fusion and fast non-local clustering is proposed for multi-temporal synthetic aperture radar (SAR) images. A binary image indicating changed regions is generated by a novel fuzzy clustering algorithm from a Contourlet fused difference image. Contourlet fusion uses complementary information from different types of difference images. For unchanged regions, the details should be restrained while highlighted for changed regions. Different fusion rules are designed for low frequency band and high frequency directional bands of Contourlet coefficients. Then a fast non-local clustering algorithm (FNLC) is proposed to classify the fused image to generate changed and unchanged regions. In order to reduce the impact of noise while preserve details of changed regions, not only local but also non-local information are incorporated into the FNLC in a fuzzy way. Experiments on both small and large scale datasets demonstrate the state-of-the-art performance of the proposed method in real applications.

TCD-Net: A Novel Deep Learning Framework for Fully Polarimetric Change Detection Using Transfer Learning

Due to anthropogenic and natural activities, the land surface continuously changes over time. The accurate and timely detection of changes is greatly important for environmental monitoring, resource management and planning activities. In this study, a novel deep learning-based change detection algorithm is proposed for bi-temporal polarimetric synthetic aperture radar (PolSAR) imagery using a transfer learning (TL) method. In particular, this method has been designed to automatically extract changes by applying three main steps as follows: (1) pre-processing, (2) parallel pseudo-label training sample generation based on a pre-trained model and fuzzy c-means (FCM) clustering algorithm, and (3) classification. Moreover, a new end-to-end three-channel deep neural network, called TCD-Net, has been introduced in this study. TCD-Net can learn more strong and abstract representations for the spatial information of a certain pixel. In addition, by adding an adaptive multi-scale shallow block and an adaptive multi-scale residual block to the TCD-Net architecture, this model with much lower parameters is sensitive to objects of various sizes. Experimental results on two Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) bi-temporal datasets demonstrated the effectiveness of the proposed algorithm compared to other well-known methods with an overall accuracy of 96.71% and a kappa coefficient of 0.82.

Change Target Extraction Based on Scale-Adaptive Difference Image and Morphology Filter for KOMPSAT-5

Multitemporal synthetic aperture radar (SAR) images have been widely used for change detection and monitoring of the environment owing to their competency under all weather conditions. However, owing to speckle backgrounds and strong reflections, change detection in urban areas is challenging. In this study, to automatically extract changed objects, we developed a model that integrated change detection and object extraction in multiple Korean Multi-Purpose Satellite-5 (KOMPSAT-5) images. Initially, two arbitrary L1A-level SAR images were input into the proposed model, and after pre-processing, such as radio calibration and coordinate system processing, change detection was performed. Subsequently, the desired targets were automatically extracted from the change detection results. Finally, the model obtained images of the extraction targets and metadata, such as date and location. Noise was removed by applying scale-adaptive modification to the generated difference image during the change detection process, and the detection accuracy was improved by emphasizing the occurrence of the change. After polygonizing the pixel groups of the change detection map in the target extraction process, the morphology-based object filtering technique was applied to minimize the false detection rate. As a result of the proposed approach, the changed objects in the KOMPSAT-5 images were automatically extracted with 90% accuracy.

A novel self-learning weighted fuzzy local information clustering algorithm integrating local and non-local spatial information for noise image segmentation

Fuzzy clustering algorithm (FCM) can be directly used to segment images, it takes no account of the neighborhood information of the current pixel and does not have a robust segmentation noise suppression. Fuzzy Local Information C-means Clustering (FLICM) is a widely used robust segmentation algorithm, which combines spatial information with the membership degree of adjacent pixels. In order to further improve the robustness of FLICM algorithm, non-local information is embedded into FLICM algorithm and a fuzzy C-means clustering algorithm has local and non-local information (FLICMLNLI) is obtained. When calculating distance from pixel to cluster center, FLICMLNLI algorithm considers two distances from current pixel and its neighborhood pixels to cluster center. However, the algorithm gives the same weight to two different distances, which incorrectly magnifies the importance of neighborhood information in calculating the distance, resulting in unsatisfactory image segmentation effects and loss of image details. In order to solve this problem, we raise an improved self-learning weighted fuzzy algorithm, which directly obtains different weights in distance calculation through continuous iterative self-learning, then the distance metric with the weights obtained from self-learning is embedded in the objective function of the fuzzy clustering algorithm in order to improve the segmentation performance and robustness of the algorithm. A large number of experiments on different types of images show that the algorithm can not only suppress the noise but also retain the details in the image, the effect of segmenting complex noise images is better, and it provides better image segmentation results than the existing latest fuzzy clustering algorithms.

Deep Siamese Networks Based Change Detection with Remote Sensing Images

Although considerable success has been achieved in change detection on optical remote sensing images, accurate detection of specific changes is still challenging. Due to the diversity and complexity of the ground surface changes and the increasing demand for detecting changes that require high-level semantics, we have to resort to deep learning techniques to extract the intrinsic representations of changed areas. However, one key problem for developing deep learning metho for detecting specific change areas is the limitation of annotated data. In this paper, we collect a change detection dataset with 862 labeled image pairs, where the urban construction-related changes are labeled. Further, we propose a supervised change detection method based on a deep siamese semantic segmentation network to handle the proposed data effectively. The novelty of the method is that the proposed siamese network treats the change detection problem as a binary semantic segmentation task and learns to extract features from the image pairs directly. The siamese architecture as well as the elaborately designed semantic segmentation networks significantly improve the performance on change detection tasks. Experimental results demonstrate the promising performance of the proposed network compared to existing approaches.

Spherical Linear Diophantine Fuzzy Soft Rough Sets with Multi-Criteria Decision Making

Modeling uncertainties with spherical linear Diophantine fuzzy sets (SLDFSs) is a robust approach towards engineering, information management, medicine, multi-criteria decision-making (MCDM) applications. The existing concepts of neutrosophic sets (NSs), picture fuzzy sets (PFSs), and spherical fuzzy sets (SFSs) are strong models for MCDM. Nevertheless, these models have certain limitations for three indexes, satisfaction (membership), dissatisfaction (non-membership), refusal/abstain (indeterminacy) grades. A SLDFS with the use of reference parameters becomes an advanced approach to deal with uncertainties in MCDM and to remove strict limitations of above grades. In this approach the decision makers (DMs) have the freedom for the selection of above three indexes in [0,1]. The addition of reference parameters with three index/grades is a more effective approach to analyze DMs opinion. We discuss the concept of spherical linear Diophantine fuzzy numbers (SLDFNs) and certain properties of SLDFSs and SLDFNs. These concepts are illustrated by examples and graphical representation. Some score functions for comparison of LDFNs are developed. We introduce the novel concepts of spherical linear Diophantine fuzzy soft rough set (SLDFSRS) and spherical linear Diophantine fuzzy soft approximation space. The proposed model of SLDFSRS is a robust hybrid model of SLDFS, soft set, and rough set. We develop new algorithms for MCDM of suitable clean energy technology. We use the concepts of score functions, reduct, and core for the optimal decision. A brief comparative analysis of the proposed approach with some existing techniques is established to indicate the validity, flexibility, and superiority of the suggested MCDM approach.

Unsupervised Change Detection Using Fuzzy Topology-Based Majority Voting

Remote sensing change detection (CD) plays an important role in Earth observation. In this paper, we propose a novel fusion approach for unsupervised CD of multispectral remote sensing images, by introducing majority voting (MV) into fuzzy topological space (FTMV). The proposed FTMV approach consists of three principal stages: (1) the CD results of different difference images produced by the fuzzy C-means algorithm are combined using a modified MV, and an initial fusion CD map is obtained; (2) by using fuzzy topology theory, the initial fusion CD map is automatically partitioned into two parts: a weakly conflicting part and strongly conflicting part; (3) the weakly conflicting pixels that possess little or no conflict are assigned to the current class, while the pixel patterns with strong conflicts often misclassified are relabeled using the supported connectivity of fuzzy topology. FTMV can integrate the merits of different CD results and largely solve the conflicting problem during fusion. Experimental results on three real remote sensing images confirm the effectiveness and efficiency of the proposed method.

Performance Evaluation of Change Detection in SAR Images Based on Hybrid Antlion DWT Fuzzy c-Means Clustering

In this paper, the main objective is to detect changes in the geographical area of Ottawa city in Canada due to floods. Two multi-temporal Synthetic Aperture Radar (SAR) images have been taken to evaluate the un-supervised change detection process. In this process, two ratio operators named as Log-Ratio and Mean-Ratio are used to generate a difference image. Performing image fusion based on DWT by selecting optimum filter coefficients by satisfying the wavelet filter coefficient properties through a novel image fusion technique is named as ADWT. GA, PSO, AntLion Optimization algorithms (ALO) and Hybridized AntLion Algorithm (HALO) have been adapted to perform the ADWT based image fusion. Segmentation has been performed based on fuzzy c-Means clustering to detect changed and unchanged pixels. Finally, the performance of the proposed method will be analysed by comparing the segmented image with the ground truth image in terms of sensitivity, accuracy, specificity, precision, F1-score.

Patch-Based Change Detection Method for SAR Images with Label Updating Strategy

Convolutional neural networks (CNNs) have been widely used in change detection of synthetic aperture radar (SAR) images and have been proven to have better precision than traditional methods. A two-stage patch-based deep learning method with a label updating strategy is proposed in this paper. The initial label and mask are generated at the pre-classification stage. Then a two-stage updating strategy is applied to gradually recover changed areas. At the first stage, diversity of training data is gradually restored. The output of the designed CNN network is further processed to generate a new label and a new mask for the following learning iteration. As the diversity of data is ensured after the first stage, pixels within uncertain areas can be easily classified at the second stage. Experiment results on several representative datasets show the effectiveness of our proposed method compared with several existing competitive methods.

A Novel Squeeze-and-Excitation W-Net for 2D and 3D Building Change Detection with Multi-Source and Multi-Feature Remote Sensing Data

Building Change Detection (BCD) is one of the core issues in earth observation and has received extensive attention in recent years. With the rapid development of earth observation technology, the data source of remote sensing change detection is continuously enriched, which provides the possibility to describe the spatial details of the ground objects more finely and to characterize the ground objects with multiple perspectives and levels. However, due to the different physical mechanisms of multi-source remote sensing data, BCD based on heterogeneous data is a challenge. Previous studies mostly focused on the BCD of homogeneous remote sensing data, while the use of multi-source remote sensing data and considering multiple features to conduct 2D and 3D BCD research is sporadic. In this article, we propose a novel and general squeeze-and-excitation W-Net, which is developed from U-Net and SE-Net. Its unique advantage is that it can not only be used for BCD of homogeneous and heterogeneous remote sensing data respectively but also can input both homogeneous and heterogeneous remote sensing data for 2D or 3D BCD by relying on its bidirectional symmetric end-to-end network architecture. Moreover, from a unique perspective, we use image features that are stable in performance and less affected by radiation differences and temporal changes. We innovatively introduced the squeeze-and-excitation module to explicitly model the interdependence between feature channels so that the response between the feature channels is adaptively recalibrated to improve the information mining ability and detection accuracy of the model. As far as we know, this is the first proposed network architecture that can simultaneously use multi-source and multi-feature remote sensing data for 2D and 3D BCD. The experimental results in two 2D data sets and two challenging 3D data sets demonstrate that the promising performances of the squeeze-and-excitation W-Net outperform several traditional and state-of-the-art approaches. Moreover, both visual and quantitative analyses of the experimental results demonstrate competitive performance in the proposed network. This demonstrates that the proposed network and method are practical, physically justified, and have great potential application value in large-scale 2D and 3D BCD and qualitative and quantitative research.

Adaptive Distance-Weighted Voronoi Tessellation for Remote Sensing Image Segmentation

The spatial fragmentation of high-resolution remote sensing images makes the segmentation algorithm put forward a strong demand for noise immunity. However, the stronger the noise immunity, the more serious the loss of detailed information, which easily leads to the neglect of effective characteristics. In view of the difficulty of balancing the noise immunity and effective characteristic retention, an adaptive distance-weighted Voronoi tessellation technology is proposed for remote sensing image segmentation. The distance between pixels and seed points in Voronoi tessellation is established by the adaptive weighting of spatial distance and spectral distance. The weight coefficient used to control the influence intensity of spatial distance is defined by a monotone decreasing function. Following the fuzzy clustering framework, a fuzzy segmentation model with Kullback–Leibler (KL) entropy regularization is established by using multivariate Gaussian distribution to describe the spectral characteristics and Markov Random Field (MRF) to consider the neighborhood effect of sub-regions. Finally, a series of parameter optimization schemes are designed according to parameter characteristics to obtain the optimal segmentation results. The proposed algorithm is validated on many multispectral remote sensing images with five comparing algorithms by qualitative and quantitative analysis. A large number of experiments show that the proposed algorithm can overcome the complex noise as well as better ensure effective characteristics.

Noise robust hybrid algorithm for segmenting image with unequal cluster sizes based on chaotic crow search and improved fuzzy c-means

Adding spatial penalty to fuzzy C-means (FCM) model is an important way to reduce the influence of noise in image segmentation. However, these improved algorithms easily cause segmentation failures when the image has the characteristics of unequal cluster sizes. Besides, they often fall into local optimal solutions if the initial cluster centers are improper. This paper presents a noise robust hybrid algorithm for segmenting image with unequal cluster sizes based on chaotic crow search algorithm and improved fuzzy c-means to overcome the above defects. Firstly, each size of clusters is integrated into the objective function of noise detecting fuzzy c-means algorithm (NDFCM), which can reduces the contribution of larger clusters to objective function and then the new membership degree and cluster centers are deduced. Secondly, a new expression called compactness, representing the pixel distribution of each cluster, is introduced into the iteration process of clustering. Thirdly, we use two- paths to seek the optimal solutions in each step of iteration: one path is produced by the chaotic crow search algorithm and the other is originated by gradient method. Furthermore, the better solutions of the two-paths go to next generation until the end of the iteration. Finally, the experiments on the synthetic and non–destructive testing (NDT) images show that the proposed algorithm behaves well in noise robustness and segmentation performance.

Graph-Based Data Fusion Applied to: Change Detection and Biomass Estimation in Rice Crops

The complementary nature of different modalities and multiple bands used in remote sensing data is helpful for tasks such as change detection and the prediction of agricultural variables. Nonetheless, correctly processing a multi-modal dataset is not a simple task, owing to the presence of different data resolutions and formats. In the past few years, graph-based methods have proven to be a useful tool in capturing inherent data similarity, in spite of different data formats, and preserving relevant topological and geometric information. In this paper, we propose a graph-based data fusion algorithm for remotely sensed images applied to (i) data-driven semi-unsupervised change detection and (ii) biomass estimation in rice crops. In order to detect the change, we evaluated the performance of four competing algorithms on fourteen datasets. To estimate biomass in rice crops, we compared our proposal in terms of root mean squared error (RMSE) concerning a recent approach based on vegetation indices as features. The results confirm that the proposed graph-based data fusion algorithm outperforms state-of-the-art methods for change detection and biomass estimation in rice crops.

Suspicious Lesion Segmentation on Brain, Mammograms and Breast MR Images Using New Optimized Spatial Feature Based Super-Pixel Fuzzy C-Means Clustering

Suspicious lesion or organ segmentation is a challenging task to be solved in most of the medical image analyses, medical diagnoses and computer diagnosis systems. Nevertheless, various image segmentation methods were proposed in the previous studies with varying success levels. But, the image segmentation problems such as lack of versatility, low robustness, high complexity and low accuracy in up-to-date image segmentation practices still remain unsolved. Fuzzy c-means clustering (FCM) methods are very well suited for segmenting the regions. The noise-free images are effectively segmented using the traditional FCM method. However, the segmentation result generated is highly sensitive to noise due to the negligence of spatial information. To solve this issue, super-pixel-based FCM (SPOFCM) is implemented in this paper, in which the influence of spatially neighbouring and similar super-pixels is incorporated. Also, a crow search algorithm is adopted for optimizing the influential degree; thereby, the segmentation performance is improved. In clinical applications, the SPOFCM feasibility is verified using the multi-spectral MRIs, mammograms and actual single spectrum on performing tumour segmentation tests for SPOFCM. Ultimately, the competitive, renowned segmentation techniques such as k-means, entropy thresholding (ET), FCM, FCM with spatial constraints (FCM_S) and kernel FCM (KFCM) are used to compare the results of proposed SPOFCM. Experimental results on multi-spectral MRIs and actual single-spectrum mammograms indicate that the proposed algorithm can provide a better performance for suspicious lesion or organ segmentation in computer-assisted clinical applications.

Self-Supervised Representation Learning for Remote Sensing Image Change Detection Based on Temporal Prediction

Traditional change detection (CD) methods operate in the simple image domain or hand-crafted features, which has less robustness to the inconsistencies (e.g., brightness and noise distribution, etc.) between bitemporal satellite images. Recently, deep learning techniques have reported compelling performance on robust feature learning. However, generating accurate semantic supervision that reveals real change information in satellite images still remains challenging, especially for manual annotation. To solve this problem, we propose a novel self-supervised representation learning method based on temporal prediction for remote sensing image CD. The main idea of our algorithm is to transform two satellite images into more consistent feature representations through a self-supervised mechanism without semantic supervision and any additional computations. Based on the transformed feature representations, a better difference image (DI) can be obtained, which reduces the propagated error of DI on the final detection result. In the self-supervised mechanism, the network is asked to identify different sample patches between two temporal images, namely, temporal prediction. By designing the network for the temporal prediction task to imitate the discriminator of generative adversarial networks, the distribution-aware feature representations are automatically captured and the result with powerful robustness can be acquired. Experimental results on real remote sensing data sets show the effectiveness and superiority of our method, improving the detection precision up to 0.94–35.49%.

A Novel Change Detection Method Based on Statistical Distribution Characteristics Using Multi-Temporal PolSAR Data

Unsupervised change detection approaches, which are relatively straightforward and easy to implement and interpret, and which require no human intervention, are widely used in change detection. Polarimetric synthetic aperture radar (PolSAR), which has an all-weather response capability with increased polarimetric information, is a key tool for change detection. However, for PolSAR data, inadequate evaluation of the difference image (DI) map makes the threshold-based algorithms incompatible with the true distribution model, which causes the change detection results to be ineffective and inaccurate. In this paper, to solve these problems, we focus on the generation of the DI map and the selection of the optimal threshold. An omnibus test statistic is used to generate the DI map from multi-temporal PolSAR images, and an improved Kittler and Illingworth algorithm based on either Weibull or gamma distribution is used to obtain the optimal threshold for generating the change detection map. Multi-temporal PolSAR data obtained by the Radarsat-2 sensor over Wuhan in China are used to verify the efficiency of the proposed method. The experimental results using our approach obtained the best performance in East Lake and Yanxi Lake regions with false alarm rates of 1.59% and 1.80%, total errors of 2.73% and 4.33%, overall accuracy of 97.27% and 95.67%, and Kappa coefficients of 0.6486 and 0.6275, respectively. Our results demonstrated that the proposed method is more suitable than the other compared methods for multi-temporal PolSAR data, and it can obtain both effective and accurate results.

Bipartite Differential Neural Network for Unsupervised Image Change Detection

Image change detection detects the regions of change in multiple images of the same scene taken at different times, which plays a crucial role in many applications. The two most popular image change detection techniques are as follows: pixel-based methods heavily rely on accurate image coregistration while object-based approaches can tolerate coregistration errors to some extent but are sensitive to image segmentation or classification errors. To address these issues, we propose an unsupervised image change detection approach based on a novel bipartite differential neural network (BDNN). The BDNN is a deep neural network with two input ends, which can extract the holistic features from the unchanged regions in the two input images, where two learnable change disguise maps (CDMs) are used to disguise the changed regions in the two input images, respectively, and thus demarcate the unchanged regions therein. The network parameters and CDMs will be learned by optimizing an objective function, which combines a loss function defined as the likelihood of the given input image pair over all possible input image pairs and two constraints imposed on CDMs. Compared with the pixel-based and object-based techniques, the BDNN is less sensitive to inaccurate image coregistration and does not involve image segmentation or classification. In fact, it can even skip over coregistration if the degree of transformation (due to the different view angles and/or positions of the camera) between the two input images is not that large. We compare the proposed approach with several state-of-the-art image change detection methods on various homogeneous and heterogeneous image pairs with and without coregistration. The results demonstrate the superiority of the proposed approach.

Two-Phase Object-Based Deep Learning for Multi-Temporal SAR Image Change Detection

Change detection is one of the fundamental applications of synthetic aperture radar (SAR) images. However, speckle noise presented in SAR images has a negative effect on change detection, leading to frequent false alarms in the mapping products. In this research, a novel two-phase object-based deep learning approach is proposed for multi-temporal SAR image change detection. Compared with traditional methods, the proposed approach brings two main innovations. One is to classify all pixels into three categories rather than two categories: unchanged pixels, changed pixels caused by strong speckle (false changes), and changed pixels formed by real terrain variation (real changes). The other is to group neighbouring pixels into superpixel objects such as to exploit local spatial context. Two phases are designed in the methodology: (1) Generate objects based on the simple linear iterative clustering (SLIC) algorithm, and discriminate these objects into changed and unchanged classes using fuzzy c-means (FCM) clustering and a deep PCANet. The prediction of this Phase is the set of changed and unchanged superpixels. (2) Deep learning on the pixel sets over the changed superpixels only, obtained in the first phase, to discriminate real changes from false changes. SLIC is employed again to achieve new superpixels in the second phase. Low rank and sparse decomposition are applied to these new superpixels to suppress speckle noise significantly. A further clustering step is applied to these new superpixels via FCM. A new PCANet is then trained to classify two kinds of changed superpixels to achieve the final change maps. Numerical experiments demonstrate that, compared with benchmark methods, the proposed approach can distinguish real changes from false changes effectively with significantly reduced false alarm rates, and achieve up to 99.71% change detection accuracy using multi-temporal SAR imagery.

An enhanced adaptive non-local means algorithm for Rician noise reduction in magnetic resonance brain images

Background

The Rician noise formed in magnetic resonance (MR) imaging greatly reduced the accuracy and reliability of subsequent analysis, and most of the existing denoising methods are suitable for Gaussian noise rather than Rician noise. Aiming to solve this problem, we proposed fuzzy c-means and adaptive non-local means (FANLM), which combined the adaptive non-local means (NLM) with fuzzy c-means (FCM), as a novel method to reduce noise in the study.

Method

The algorithm chose the optimal size of search window automatically based on the noise variance which was estimated by the improved estimator of the median absolute deviation (MAD) for Rician noise. Meanwhile, it solved the problem that the traditional NLM algorithm had to use a fixed size of search window. Considering the distribution characteristics for each pixel, we designed three types of search window sizes as large, medium and small instead of using a fixed size. In addition, the combination with the FCM algorithm helped to achieve better denoising effect since the improved the FCM algorithm divided the membership degrees of images and introduced the morphological reconstruction to preserve the image details.

Results

The experimental results showed that the proposed algorithm (FANLM) can effectively remove the noise. Moreover, it had the highest peak signal-noise ratio (PSNR) and structural similarity (SSIM), compared with other three methods: non-local means (NLM), linear minimum mean square error (LMMSE) and undecimated wavelet transform (UWT). Using the FANLM method, the image details can be well preserved with the noise being mostly removed.

Conclusion

Compared with the traditional denoising methods, the experimental results showed that the proposed approach effectively suppressed the noise and the edge details were well retained. However, the FANLM method took an average of 13 s throughout the experiment, and its computational cost was not the shortest. Addressing these can be part of our future research.

Mapping and assessing spatial extent of floods from multitemporal synthetic aperture radar images: a case study on Brahmaputra River in Assam State, India

Brahmaputra is one of the perennial rivers in India which causes floods every year in the north-east state of Assam causing hindrance to normal life and damage to crops. The availability of temporal Remote Sensing (RS) data helps to study the periodical changes caused by flood event and its eventual effect on natural environment. Integrating RS and GIS methods paved a way for effective flood mapping over a large spatial extent which helps to assess the damage accurately for mitigation. In the present study, multitemporal Sentinel-1A data is exploited to assess the 2017 flood situation of Brahmaputra River in Assam state. Five data sets that are taken during flood season and one reference data taken during the non-monsoon season are used to estimate the area inundated under floods for the quantification of damage assessment. A visual interpretation map is produced using colour segmentation method by estimating the thresholds from histogram analysis. A new method is developed to identify the optimum value for threshold from statistical distribution of Synthetic Aperture Radar (SAR) data that separates flooded water and non-flooded water. From this method, the range of backscatter values for normal water are identified as - 18 to - 30 dB and the range is identified as - 19 to - 24 dB for flooded water. The results showed that the method is able to separate the flooded and non-flooded region on the microwave data set, and the derived flood extent using this method shows the inundated area of 3873.14 Km² on peak flood date for the chosen study area.

Enhanced Fuzzy-based Local Information Algorithm for Sonar Image Segmentation

The recent boost in undersea operations has led to the development of high-resolution sonar systems mounted on autonomous vehicles. These vehicles are used to scan the seafloor in search of different objects such as sunken ships, archaeological sites, and submerged mines. An important part of the detection operation is the segmentation of sonar images, where the object's highlight and shadow are distinguished from the seabed background. In this work, we focus on the automatic segmentation of sonar images. We present our enhanced fuzzybased with Kernel metric (EnFK) algorithm for the segmentation of sonar images which, in an attempt to improve segmentation accuracy, introduces two new fuzzy terms of local spatial and statistical information. Our algorithm includes a preliminary de-noising algorithm which, together with the original image, feeds into the segmentation procedure to avoid trapping to local minima and to improve convergence. The result is a segmentation procedure that specifically suits the intensity inhomogeneity and the complex seabed texture of sonar images. We tested our approach using simulated images, real sonar images, and sonar images that we created in two different sea experiments, using multibeam sonar and synthetic aperture sonar. The results show accurate segmentation performance that is far beyond the stateof-the-art results.

Fuzzy Clustering Algorithm with Non-Neighborhood Spatial Information for Surface Roughness Measurement Based on the Reflected Aliasing Images

Due to the limitation of the fixed structures of neighborhood windows, the quality of spatial information obtained from the neighborhood pixels may be affected by noise. In order to compensate this drawback, a robust fuzzy c-means clustering with non-neighborhood spatial information (FCM_NNS) is presented. Through incorporating non-neighborhood spatial information, the robustness performance of the proposed FCM_NNS with respect to the noise can be significantly improved. The results indicate that FCM_NNS is very effective and robust to noisy aliasing images. Moreover, the comparison of other seven roughness indexes indicates that the proposed FCM_NNS-based F index can characterize the aliasing degree in the surface images and is highly correlated with surface roughness (R² = 0.9327 for thirty grinding samples).

Automated Quantification of Hyperreflective Foci in SD-OCT With Diabetic Retinopathy

The presence of hyperreflective foci (HFs) is related to retinal disease progression, and the quantity has proven to be a prognostic factor of visual and anatomical outcome in various retinal diseases. However, lack of efficient quantitative tools for evaluating the HFs has deprived ophthalmologist of assessing the volume of HFs. For this reason, we propose an automated quantification algorithm to segment and quantify HFs in spectral domain optical coherence tomography (SD-OCT). The proposed algorithm consists of two parallel processes namely: region of interest (ROI) generation and HFs estimation. To generate the ROI, we use morphological reconstruction to obtain the reconstructed image and histogram constructed for data distributions and clustering. In parallel, we estimate the HFs by extracting the extremal regions from the connected regions obtained from a component tree. Finally, both the ROI and the HFs estimation process are merged to obtain the segmented HFs. The proposed algorithm was tested on 40 3D SD-OCT volumes from 40 patients diagnosed with non-proliferative diabetic retinopathy (NPDR), proliferative diabetic retinopathy (PDR), and diabetic macular edema (DME). The average dice similarity coefficient (DSC) and correlation coefficient (r) are 69.70%, 0.99 for NPDR, 70.31%, 0.99 for PDR, and 71.30%, 0.99 for DME, respectively. The proposed algorithm can provide ophthalmologist with good HFs quantitative information, such as volume, size, and location of the HFs.

Similarity Measure-Based Possibilistic FCM With Label Information for Brain MRI Segmentation

Magnetic resonance imaging (MRI) is extensively applied in clinical practice. Segmentation of the MRI brain image is significant to the detection of brain abnormalities. However, owing to the coexistence of intensity inhomogeneity and noise, dividing the MRI brain image into different clusters precisely has become an arduous task. In this paper, an improved possibilistic fuzzy c -means (FCM) method based on a similarity measure is proposed to improve the segmentation performance for MRI brain images. By introducing the new similarity measure, the proposed method is more effective for clustering the data with nonspherical distribution. Besides that, the new similarity measure could alleviate the "cluster-size sensitivity" problem that most FCM-based methods suffer from. Simultaneously, the proposed method could preserve image details as well as suppress image noises via the use of local label information. Experiments conducted on both synthetic and clinical images show that the proposed method is very effective, providing mitigation to the cluster-size sensitivity problem, resistance to noisy images, and applicability to data with more complex distribution.

A novel unsupervised change detection approach based on reconstruction independent component analysis and ABC-Kmeans clustering for environmental monitoring

In this paper, I propose a new unsupervised change detection method for optical satellite imagery. The proposed technique consists of three phases. In the first stage, difference images are calculated using four different functions. Two of the functions were first used in this study. In the second stage, using Reconstruction Independent Component Analysis, this four-difference matrix is projected to one feature. In the last stage, clustering is performed. Kmeans tuned by Artificial Bee Colony (ABC-Kmeans) clustering technique has been developed and proposed by following a different strategy in the clustering phase. The effectiveness of the proposed approach was examined using two different datasets, Sardinia and Mexico. Quantitative evaluation was performed in two stages. In the first stage, proposed method was compared with different unsupervised change detection algorithms using False Alarm, Missed Alarm, Total Error, and Total Error Rate metrics which are calculated using ground truth image in dataset. In the second experimental study, the proposed approach is compared in detail with PCA-Kmeans approach, which is quite often preferred for similar studies, using the Mean Squared Error, Peak Signal to Noise Ratio, Structural Similarity Index, and Universal Image Quality Index metrics. According to quantitative and qualitative analysis, proposed approach can produce quite successful results using optical remote sensing data.

Newly Built Construction Detection in SAR Images Using Deep Learning

Remote sensing data can be utilized to help developing countries monitor the use of land. However, the problem of constant cloud coverage prevents us from taking full advantage of satellite optical images. Therefore, we instead opt to use data from synthetic-aperture radar (SAR), which can capture images of the Earth’s surface regardless of the weather conditions. In this study, we use SAR data to identify newly built constructions. Most studies on change detection tend to detect all of the changes that have a similar temporal change characteristic occurring on two occasions, while we want to identify only the constructions and avoid detecting other changes such as the seasonal change of vegetation. To do so, we study various deep learning network techniques and have decided to propose the fully convolutional network with a skip connection. We train this network with pairs of SAR data acquired on two different occasions from Bangkok and the ground truth, which we manually create from optical images available from Google Earth for all of the SAR pairs. Experiments to assign the most suitable patch size, loss weighting, and epoch number to the network are discussed in this paper. The trained model can be used to generate a binary map that indicates the position of these newly built constructions precisely with the Bangkok dataset, as well as with the Hanoi and Xiamen datasets with acceptable results. The proposed model can even be used with SAR images of the same specific satellite from another orbit direction and still give promising results.

Kernel-Based Robust Bias-Correction Fuzzy Weighted C-Ordered-Means Clustering Algorithm

The spatial constrained Fuzzy C-means clustering (FCM) is an effective algorithm for image segmentation. Its background information improves the insensitivity to noise to some extent. In addition, the membership degree of Euclidean distance is not suitable for revealing the non-Euclidean structure of input data, since it still lacks enough robustness to noise and outliers. In order to overcome the problem above, this paper proposes a new kernel-based algorithm based on the Kernel-induced Distance Measure, which we call it Kernel-based Robust Bias-correction Fuzzy Weighted C-ordered-means Clustering Algorithm (KBFWCM). In the construction of the objective function, KBFWCM algorithm comprehensively takes into account that the spatial constrained FCM clustering algorithm is insensitive to image noise and involves a highly intensive computation. Aiming at the insensitivity of spatial constrained FCM clustering algorithm to noise and its image detail processing, the KBFWCM algorithm proposes a comprehensive algorithm combining fuzzy local similarity measures (space and grayscale) and the typicality of data attributes. Aiming at the poor robustness of the original algorithm to noise and outliers and its highly intensive computation, a Kernel-based clustering method that includes a class of robust non-Euclidean distance measures is proposed in this paper. The experimental results show that the KBFWCM algorithm has a stronger denoising and robust effect on noise image.

An Unsupervised SAR Change Detection Method Based on Stochastic Subspace Ensemble Learning

As synthetic aperture radar (SAR) is playing an increasingly important role in Earth observations, many new methods and technologies have been proposed for change detection using multi-temporal SAR images. Especially with the development of deep learning, numerous methods have been proposed in recent years. However, the requirement to have a certain number of high-quality samples has become one of the main reasons for the limited development of these methods. Thus, in this paper, we propose an unsupervised SAR change detection method that is based on stochastic subspace ensemble learning. The proposed method consists of two stages: The first stage involves the automatic determination of high-confidence samples, which includes a fusion strategy and a refinement process; and the second stage entails using the stochastic subspace ensemble learning module, which contains three steps: obtaining the subsample sets, establishing and training a two-channel network, and applying the prediction results and an ensemble strategy. The subsample sets are used to solve the problem of imbalanced samples. The two-channel networks are used to extract high-dimensional features and learn the relationship between the neighborhood of the pixels in the original images and the labels. Finally, by using an ensemble strategy, the results predicted by all patches reclassified in each network are integrated as the detection result. The experimental results of different SAR datasets prove the effectiveness and the feasibility of the proposed method.

Change Analysis in Urban Areas Based on Statistical Features and Temporal Clustering Using TerraSAR-X Time-Series Images

The existing unsupervised multitemporal change detection approaches for synthetic aperture radar (SAR) images based on the pixel level usually suffer from the serious influence of speckle noise, and the classification accuracy of temporal change patterns is liable to be affected by the generation method of similarity matrices and the pre-specified cluster number. To address these issues, a novel time-series change detection method with high efficiency is proposed in this paper. Firstly, spatial feature extraction using local statistical information on patches is conducted to reduce the noise and for subsequent temporal grouping. Secondly, a density-based clustering method is adopted to categorize the pixel series in the temporal dimension, in view of its efficiency and robustness. Change detection and classification results are then obtained by a fast differential strategy in the final step. The experimental results and analysis of synthetic and realistic time-series SAR images acquired by TerraSAR-X in urban areas demonstrate the effectiveness of the proposed method, which outperforms other approaches in terms of both qualitative results and quantitative indices of macro F1-scores and micro F1-scores. Furthermore, we make the case that more temporal change information for buildings can be obtained, which includes when the first and last detected change occurred and the frequency of changes.

Fusion Network for Change Detection of High-Resolution Panchromatic Imagery

This paper proposes a fusion network for detecting changes between two high-resolution panchromatic images. The proposed fusion network consists of front- and back-end neural network architectures to generate dual outputs for change detection. Two networks for change detection were applied to handle image- and high-level changes of information, respectively. The fusion network employs single-path and dual-path networks to accomplish low-level and high-level differential detection, respectively. Based on two dual outputs, a two-stage decision algorithm was proposed to efficiently yield the final change detection results. The dual outputs were incorporated into the two-stage decision by operating logical operations. The proposed algorithm was designed to incorporate not only dual network outputs but also neighboring information. In this paper, a new fused loss function was presented to estimate the errors and optimize the proposed network during the learning stage. Based on our experimental evaluation, the proposed method yields a better detection performance than conventional neural network algorithms, with an average area under the curve of 0.9709, percentage correct classification of 99%, and Kappa of 75 for many test datasets.

Small-Target Detection between SAR Images Based on Statistical Modeling of Log-Ratio Operator

The log-ratio (LR) operator is well suited for change detection in synthetic aperture radar (SAR) amplitude or intensity images. In applying the LR operator to change detection in multi-temporal SAR images, a crucial problem is how to develop precise models for the LR statistics. In this study, we first derive analytically the probability density function (PDF) of the LR operator. Subsequently, the PDF of the LR statistics is parameterized by three parameters, i.e., the number of looks, the coherence magnitude, and the true intensity ratio. Then, the maximum-likelihood (ML) estimates of parameters in the LR PDF are also derived. As an example, the proposed statistical model and corresponding ML estimation are used in an operational application, i.e., determining the constant false alarm rate (CFAR) detection thresholds for small target detection between SAR images. The effectiveness of the proposed model and corresponding ML estimation are verified by applying them to measured multi-temporal SAR images, and comparing the results to the well-known generalized Gaussian (GG) distribution; the usefulness of the proposed LR PDF for small target detection is also shown.

Change Detection in Remote Sensing Images Based on Image Mapping and a Deep Capsule Network

Homogeneous image change detection research has been well developed, and many methods have been proposed. However, change detection between heterogeneous images is challenging since heterogeneous images are in different domains. Therefore, direct heterogeneous image comparison in the way that we do it is difficult. In this paper, a method for heterogeneous synthetic aperture radar (SAR) image and optical image change detection is proposed, which is based on a pixel-level mapping method and a capsule network with a deep structure. The mapping method proposed transforms an image from one feature space to another feature space. Then, the images can be compared directly in a similarly transformed space. In the mapping process, some image blocks in unchanged areas are selected, and these blocks are only a small part of the image. Then, the weighted parameters are acquired by calculating the Euclidean distances between the pixel to be transformed and the pixels in these blocks. The Euclidean distance calculated according to the weighted coordinates is taken as the pixel gray value in another feature space. The other image is transformed in a similar manner. In the transformed feature space, these images are compared, and the fusion of the two different images is achieved. The two experimental images are input to a capsule network, which has a deep structure. The image fusion result is taken as the training labels. The training samples are selected according to the ratio of the center pixel label and its neighboring pixels’ labels. The capsule network can improve the detection result and suppress noise. Experiments on remote sensing datasets show the final detection results, and the proposed method obtains a satisfactory performance.