Imagery Change Detection Research Articles

Change detection on multi-sensor imagery using mixed interleaved group convolutional network

The difference of the spatio-spectral features of multi-sensor image causes big difficulty in change detection because of the difficulties of the feature extraction. Unlike the traditional approaches that mainly relying on manually feature design, the advances of deep learning-based methods in deep feature extraction provide new alternatives for multi-sensor imagery change detection. Specifically, the incorporation of multi-scale information from remote sensing images holds paramount importance in change detection, consistently applied in the design of various deep learning models. This study investigated a change detection approach utilizing a mixed interleaved group convolutional network (MIGCNet) on multi-sensor remote sensing imagery, with a specific focus on fine-grained kernel space and multi-scale feature analysis within convolution operations. The proposed MIGCNet, with parallel branches as the fundamental architecture, can distinguish the change information effectively by the proposed mixed interleaved group convolution (MIGC) module, which combined mixed convolution with interleaved group convolution. Meanwhile, multi-loss supervision is utilized to promote the performance of the proposed MIGCNet. Experimental results demonstrate the outperformance of the MIGCNet to handle change detection with multi-sensor images on urban area. Considering different datasets, the Overall Accuracy and Kappa Coefficient are reaching 0.97 and 80.67%, respectively, and the miss detection rate and the false alarm rate are as low as 0.17 and 0.18, respectively.

Deep collaborative learning with class-rebalancing for semi-supervised change detection in SAR images

Deep learning reveals excellent potential for accomplishing change detection in SAR imagery. Yet, it suffers from the problem of requiring large amounts of labeled samples, whilst labeling SAR imagery for change detection requires experts to label individual images at the pixel level, which is extremely tedious and time-consuming. Also, sample imbalance continues to present a serious challenge for the existing change detection techniques. To tackle these problems, in this study, a Deep Collaborative semi-supervised learning Framework with Class-Rebalancing (DCF-CRe) is proposed for SAR imagery change detection, by exploiting Convolutional Neural Network (CNN) and deep clustering. In particular, a Siamese Difference Fusion Network (SDFNet) is devised to implement change detection while effectively reducing the information loss due to the generation of difference images and highlighting features of the changed regions.In so doing, only a tiny batch of labeled samples is utilized to train SDFNet in order to obtain predicted change map and deep features. In addition, the Approximate Rank-Order Clustering (AROC) algorithm is employed to cluster the deep features, generating pseudo-labels for abundant unlabeled samples. DCF-CRe is then applied to select appropriate pseudo-labels and to add labeled samples to train SDFNet. Experimental results evaluated on six challenging datasets show that this proposed approach can achieve performance superior to state-of-the-art change detection methods for SAR imagery.

PCBA-Net: Pyramidal Convolutional Block Attention Network for Synthetic Aperture Radar Image Change Detection

Synthetic aperture radar (SAR) imagery change detection (CD) is still a crucial and challenging task. Recently, with the boom of deep learning technologies, many deep learning methods have been presented for SAR CD, and they achieve superior performance to traditional methods. However, most of the available convolutional neural networks (CNN) approaches use diminutive and single convolution kernel, which has a small receptive field and cannot make full use of the context information and some useful detail information of SAR images. In order to address the above drawback, pyramidal convolutional block attention network (PCBA-Net) is proposed for SAR image CD in this study. The proposed PCBA-Net consists of pyramidal convolution (PyConv) and convolutional block attention module (CBAM). PyConv can not only extend the receptive field of input to capture enough context information, but also handles input with incremental kernel sizes in parallel to obtain multi-scale detailed information. Additionally, CBAM is introduced in the PCBA-Net to emphasize crucial information. To verify the performance of our proposed method, six actual SAR datasets are used in the experiments. The results of six real SAR datasets reveal that the performance of our approach outperforms several state-of-the-art methods.

Spatial–Spectral Decoupling Interaction Network for Multispectral Imagery Change Detection

We present a spatial–spectral decoupling interaction network for multispectral imagery change detection, which can exploit the underlying information of the multispectral imagery adequately through simultaneously considering the discriminative attribute of each pixel and robust spatial structure of the corresponding patch. Specifically, a 1-D convolutional neural network (1D-CNN) is applied to the spectral vector of each pixel to extract its discriminative feature, while a 2D-CNN is applied to the patch centering on the corresponding pixel to explore the spatial structure information. In addition, an interaction mechanism is incorporated into the feature fusion module to enhance the spatial–spectral consistency.

Object-level change detection with a dual correlation attention-guided detector

Automatic change detection from remotely sensed imagery is extremely important for many applications, including land use mapping. In recent years, a growing number of researchers have applied capable deep-learning methods to the research on change detection. The majority of deep learning-based change detection methods currently perform pixel-by-pixel classification at the original image scale, but they can hardly avoid the false changes caused by strong parallax effects and projected shadows, without considering the totality of changed objects/regions. In this study, we propose an object-level change detection framework to detect changed geographic entities (such as newly built buildings or changed artificial structures) by paying more attention to the overall characteristics and context association of changed object instances. The detected changed objects are represented as bounding boxes, which are simple, regular, and convenient to use in object feature extraction. In terms of data handling, a special data augmentation method for change detection called Alternative-Mosaic is proposed to effectively accelerate model training and improve model performance. For the model, we propose a one-stage change detection network called dual correlation attention-guided detector (DCA-Det) to detect the changed objects. In particular, we feed the dual-temporal images into a weight-shared backbone network to extract the change features of different scales. The change features on the same scale are further refined, and then the features between different scales are fused by the correlation attention-guided feature fusion neck. Finally, the change detection heads output the prediction results of the changed objects/regions of different scales. Experiments were conducted on public LEVIR building change detection and aerial imagery change detection (AICD) datasets. The quantitative evaluation and visualization results proved the superiority and robustness of our framework. Our DCA-Det can obtain state-of-the-art performance on object-level metrics (99.50% APIoU=.50 and 79.72% APIoU=.50:.05:.95) on the AICD-2012 dataset.

Training Compact Change Detection Network for Remote Sensing Imagery

Change Detection (CD) is a hot remote sensing topic where the change zones are highlighted by analyzing bi-temporal or multi-temporal images. Recently, Deep learning (DL) paved the road to implement various reliable change detection approaches that overcome traditional CD methods limitation. However, high performance DL based approaches have explosion number of parameters that demanded extensive computation and memory usage in addition to large volumes of training data. To address this issue, we proposed a teacher-student setting for remote sensing imagery change detection. To distill the knowledge from the over-parameterized Siamese teacher network, we proposed tiny student network that was trained using the obtained categorical distribution of probability from the teacher paired Softmax output at high temperature. Practical Swarm Optimization (PSO) was applied in order to optimally configure student architecture. Finally, ample experiments were conducted on LEVIR-CD dataset. Also, we introduced EGSAR-CD dataset, which contains of a large set of bi-temporal SAR images with 460 image pairs ( 256 ×256). Experiment results indicate that we can reach up to 5.4× reduction rate in number of parameters with loss of accuracy between 5% and 6% on the LEVIR-CD and EGSAR-CD datasets utilizing self-knowledge distillation.

Change Detection Based on IR-MAD Model for GF-5 Remote Sensing Imagery

GF-5 is the domestic full spectrum satellite with the most spectral bands, and it can comprehensive observe earth and atmosphere, the data can effectively monitor the changes of ground objects. However, due to the high wave dimension and large data of the hyperspectral remote sensing image, which reduces the processing and operation speed, and brings great uncertainty to the accuracy. In order to improve the accuracy and processing speed of hyperspectral imagery change detection, a method of iterative weighted multivariate change detection based on IR-MAD is proposed. In this paper, the high-resolution remote sensing image of GF-5 is used as the data source. After geometric correction, removal of bad line and other pre-processing methods, the change detection results are obtained by the iterative weighted multivariate (IR-MAD) change detection method. The experiments show that: the algorithm in this paper is compared with change vector analysis (CVA) change detection, principal component change vector analysis (PCA-CVA) change detection method, and iterative weighted multivariate (IR-MAD) detection method without principal component extraction. The detection accuracy of this method is high, and the error rate and missed rate are also low.

Incorporating Metric Learning and Adversarial Network for Seasonal Invariant Change Detection

Change detection by comparing two bitemporal images is one of the most fundamental challenges for dynamic monitoring of the Earth surface. In this article, we propose a metric learning-based generative adversarial network (GAN) (MeGAN) to automatically explore seasonal invariant features for pseudochange suppressing and real change detection. To achieve this purpose, a seasonal invariant term is introduced to maximally suppress pseudochanges, whereas the MeGAN explores the transition patterns between adjacent images in a self-learning fashion. Different from the previous works on bitemporal imagery change detection, the proposed MeGAN have the following contributions: 1) it automatically explores change patterns from the complex bitemporal background without human intervention and 2) it aims to maximally exclude pseudochanges from the seasonal transition term and map out real changes efficiently. To our best knowledge, this is the first time we incorporate the seasonal transition term and GAN for change detection between bitemporal images. At last, to demonstrate the robustness of the proposed method, we included two data sets which are the Google Earth data and the Landsat data, for bitemporal change detection and evaluation. The experimental results indicated that the proposed method is able to perform change detection with precision can be as high as 81% and 88% for the Google Earth and Landsat data set, respectively.

High-resolution optical remote sensing imagery change detection through deep transfer learning

Change detection is a challenging task that has received much attention in the remote sensing field. Whereas numerous remote sensing change detection methods have been developed, the efficiency of these approaches is insufficient to meet the real-world applications’ requirements. Recently, deep learning methods have been largely used for remote sensing imagery change detection, most of these approaches are limited by their training dataset. However, adapting a pretrained convolutional neural network (CNN) on an image classification task to change detection is extremely challenging. An automatic land cover/use change detection approach based on fast and accurate frameworks for optical high-resolution remote sensing imagery is proposed. The fast framework is designed for applications that require immediate results with less complexity. The accurate framework is designed for applications that require high levels of precision, it decomposes large images into small processing blocks and forwards them into CNN. The proposed frameworks can learn transferable features from one task to another and escape the use of the expensive and inaccurate handcrafted features and the requirements of the big training dataset. A number of experiments were carried out to validate the proposed approach on three real bitemporal images. The experimental results illustrate the superiority of the proposed approach over other state-of-the-art methods.

Selection of Imagery Change Detection Methods Concerning Seasonal Difference

Updating land cover with imagery change detection methods is a hot topic. Many imagery change detection methods have been developed from different views; however, these methods often have their specific applicability, indicating that there is no single method can be universally applicable for land cover updating especially in large area. The main challenge is that not all remote sensing images used in change detection would be aquired from the same season. In this paper, two typical change detection methods SGD and PCC were analyzed and selected respectively for land cover updating concerning the seasonal difference of multi-temporal remote sensed images. Experiment in a case study of Shandong province was conducted by using the two methods. Results indicated that SGD is suitable for consistent seasonal phase images, while PCC can be used for those inconsistent seasonal phase images. It can achieve better accuracy in land cover updating using suitable change detection method concerning seasonal difference.

基于知识库的像斑光谱向量相似度土地覆盖变化检测方法

土地利用/覆盖变化检测是国内外全球化进程研究的重要内容,选择适当的变化检测方法对西北地区土地利用/覆盖变化进行研究在"生态十年项目"中具有重要的意义.选择西北地区具有典型代表性的TM轨道号134033区域作为变化检测方法验证的试验区,采用2005和2010年两期Landsat TM影像,在eCognition Developer 8.64软件支持下,采用基于像斑的光谱特征特征向量相似度方法进行变化检测,并利用2010年土地覆盖数据作为先验知识库对变化区域分类,提取土地利用/覆盖变化信息,并对变化结果进行定量分析.结果表明,采用基于像斑的光谱特征特征向量相似度方法对于试验区的土地利用/覆盖变化制图具有检测快速、检测精度高等优点,适合试验区以及整个西北地区的土地利用/覆盖变化的检测.最终采用该方法以及分类后比较法获得了西北地区2000-2010年近10年的土地利用/覆盖分类图.;Land use/cover change (LUCC) is an important field of global and local environmental change research. LUCC and its corresponding effects have a direct impact on the environment and ecological processes, and in turn natural resources management and related decisions. In China, LUCC research is at the core of all issues related to sustainable development, with tremendous practical significance. Remote sensing imagery change detection has great application value in many areas, such as the survey of land-use change, urban expansion, and vegetation resource monitoring. The choice of an effective change detection method for a given study area is a central issue of such detection. In research conducted in various countries, remote sensing imagery change detection methods have transitioned from pixel level to feature level and knowledge level. Currently used LUCC detection methods are in two categories, direct spectrum comparison and comparison of classification results. The spectrum feature vector of a segment in a Landsat Thematic Mapper (TM) image can be regarded as a vector of six-dimensional feature space. If the angle between two vectors is smaller and vector mode closer, the more similar are the two vectors. Thus, we used the cosine of the angle between two vectors and the ratio of vector mode to establish a vector similarity index for measuring vector similarity. Combined with the National Change of Ecological Environment Decade (2000-2010) Remote Sensing Survey and Assessment project, jointly organized and implemented by the Ministry of Environmental Protection and Chinese Academy of Sciences, this paper focuses on application of a change detection method. This method uses segment similarity of a spectrum vector based on a knowledge base in the northwest region. There is also an evaluation of method accuracy. The area covered by TM path 134 row 33, representative of the northwest region, was chosen as a change detection method validation test site. We used 2005 and 2010 two-phase Landsat TM imagery to detect land-cover change, using spectrum vector similarity based on a segment with support from eCognition Developer 8.64 software. We used 2010 land-cover data as a priori knowledge base to classify regions of change. The results showed the following. 1) Segment similarity of the spectrum vector method is appropriate for change detection in the northwest region, and accuracy of the 2005 land-cover database established using the 2010 land-cover database as a priori knowledge was relatively high. 2) The method of segment similarity of the spectrum vector has many advantages, such as less computation, fast operation, and suitability for large-scale rapid change detection. 3) The method is more effective for farmland, water, built-up land and vegetation cover change. 4) Accuracy of the land-cover database established by change detection depends on accuracy of the baseline land-cover database. Compared with the latter database, accuracy of the former database declined slightly. The main cause of this decline is change that was difficult to detect and change area misclassification, attributable to land-cover types with similar spectral features such as lakes and reservoirs, bare land, and sparse vegetation. 5) Because of the effects of imagery phase and cloud coverage, the land-cover database established by change detection requires additional manual modifications to improve its accuracy.

Temporal Sequences Quantify the Contributions of Individual Fixations in Complex Perceptual Matching Tasks

Perceptual tasks such as object matching, mammogram interpretation, mental rotation, and satellite imagery change detection often require the assignment of correspondences to fuse information across views. We apply techniques developed for machine translation to the gaze data recorded from a complex perceptual matching task modeled after fingerprint examinations. The gaze data provide temporal sequences that the machine translation algorithm uses to estimate the subjects' assumptions of corresponding regions. Our results show that experts and novices have similar surface behavior, such as the number of fixations made or the duration of fixations. However, the approach applied to data from experts is able to identify more corresponding areas between two prints. The fixations that are associated with clusters that map with high probability to corresponding locations on the other print are likely to have greater utility in a visual matching task. These techniques address a fundamental problem in eye tracking research with perceptual matching tasks: Given that the eyes always point somewhere, which fixations are the most informative and therefore are likely to be relevant for the comparison task?

Updating the 2001 National Land Cover Database land cover classification to 2006 by using Landsat imagery change detection methods

The recent release of the U.S. Geological Survey (USGS) National Land Cover Database (NLCD) 2001, which represents the nation's land cover status based on a nominal date of 2001, is widely used as a baseline for national land cover conditions. To enable the updating of this land cover information in a consistent and continuous manner, a prototype method was developed to update land cover by an individual Landsat path and row. This method updates NLCD 2001 to a nominal date of 2006 by using both Landsat imagery and data from NLCD 2001 as the baseline. Pairs of Landsat scenes in the same season in 2001 and 2006 were acquired according to satellite paths and rows and normalized to allow calculation of change vectors between the two dates. Conservative thresholds based on Anderson Level I land cover classes were used to segregate the change vectors and determine areas of change and no-change. Once change areas had been identified, land cover classifications at the full NLCD resolution for 2006 areas of change were completed by sampling from NLCD 2001 in unchanged areas. Methods were developed and tested across five Landsat path/row study sites that contain several metropolitan areas including Seattle, Washington; San Diego, California; Sioux Falls, South Dakota; Jackson, Mississippi; and Manchester, New Hampshire. Results from the five study areas show that the vast majority of land cover change was captured and updated with overall land cover classification accuracies of 78.32%, 87.5%, 88.57%, 78.36%, and 83.33% for these areas. The method optimizes mapping efficiency and has the potential to provide users a flexible method to generate updated land cover at national and regional scales by using NLCD 2001 as the baseline.

Theoretical foundation of the intersecting cortical model and its use for change detection of aircraft, cars, and nuclear explosion tests

The intersecting cortical model (ICM) is a model based on neural network techniques especially designed for image processing. It was derived from several visual cortex models and is basically the intersection of these models, i.e. the common elements amongst these models. The theoretical foundation of the ICM is given and it is shown how the ICM can be derived as a reduced set of equations of the pulse-coupled neural network based upon models proposed by Eckhorn and Reitboeck. Tests of the ICM are presented: one on a series of images of an aircraft moving in the sky; two on car detection; and one on preparations of underground nuclear explosions. The ICM is shown here, in a few examples, to be useful in imagery change detection: aircraft moving against a homogeneous background without precise geometric matching; car on a road; two cars moving in an urban setting without precise geometric matching; and for a linear structure in a complex background. The ICM can be used when the moving objects are not too small and the background is not too difficult. Changes involving larger linear structures can be detected even if the background is not homogeneous.