Joint Variation Learning of Fusion and Difference Features for Change Detection in Remote Sensing Images

Jiang, Kaixuan; Zhang, Wenhua; Liu, Jia; Liu, Fang; Xiao, Liang

doi:10.1109/tgrs.2022.3226778

Cited by 16 publications

(7 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These proposed new architectures are in line with recent developments for 2D image change detection concerning the importance of data fusion [25], [29], [30] for change detection task. Indeed, by convolving change features in the encoder, we expect that the network will put more attention on changes and also better combine multi-scale change features.…”

Section: B New Models For 3d Point Clouds Change Detectionsupporting

confidence: 68%

“…For example, [28] propose a network based on the three results of addition, subtraction, and concatenation of features at multiple scales. The study in [29] takes a step aside from the traditional Siamese network with one input for each branch, by proposing to take in one branch the concatenation of the images, and in the other the difference, forming two sub-networks with different properties. At the output of each layer, the features of the two branches are summed and then concatenated at the corresponding scale in the decoder thanks to skip connections.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Change Detection in Urban Point Clouds: An Experimental Comparison with Simulated 3D Datasets

2021

View full text Add to dashboard Cite

In the context of rapid urbanization, monitoring the evolution of cities is crucial. To do so, 3D change detection and characterization is of capital importance since, unlike 2D images, 3D data contain vertical information of utmost importance to monitoring city evolution (that occurs along both horizontal and vertical axes). Urban 3D change detection has thus received growing attention, and various methods have been published on the topic. Nevertheless, no quantitative comparison on a public dataset has been reported yet. This study presents an experimental comparison of six methods: three traditional (difference of DSMs, C2C and M3C2), one machine learning with hand-crafted features (a random forest model with a stability feature) and two deep learning (feed-forward and Siamese architectures). In order to compare these methods, we prepared five sub-datasets containing simulated pairs of 3D annotated point clouds with different characteristics: from high to low resolution, with various levels of noise. The methods have been tested on each sub-dataset for binary and multi-class segmentation. For supervised methods, we also assessed the transfer learning capacity and the influence of the training set size. The methods we used provide various kinds of results (2D pixels, 2D patches or 3D points), and each of them is impacted by the resolution of the PCs. However, while the performances of deep learning methods highly depend on the size of the training set, they seem to be less impacted by training on datasets with different characteristics. Oppositely, conventional machine learning methods exhibit stable results, even with smaller training sets, but embed low transfer learning capacities. While the main changes in our datasets were usually identified, there were still numerous instances of false detection, especially in dense urban areas, thereby calling for further development in this field. To assist such developments, we provide a public dataset composed of pairs of point clouds with different qualities together with their change-related annotations. This dataset was built with an original simulation tool which allows one to generate bi-temporal urban point clouds under various conditions.

show abstract

Section: B New Models For 3d Point Clouds Change Detectionsupporting

confidence: 68%

mentioning

confidence: 99%

Change Detection in Urban Point Clouds: An Experimental Comparison with Simulated 3D Datasets

2021

View full text Add to dashboard Cite

show abstract

“…Generally, the changed objects in bi-temporal images often have significant size variations [24], which leads to incomplete detection targets and unsmooth boundaries in the results. Consequently, it is imperative to explore multi-scale feature fusion strategies to smooth the boundaries and improve the accuracy of results.…”

Section: Multi-scale Gated Fusion Modulementioning

confidence: 99%

“…Although the methods mentioned above have achieved performance results, they do not consider the characteristics of bi-temporal multi-scale features, thereby resulting in incomplete detection targets and limited accuracy of results. Inspired by the widely used multi-scale pyramid architecture for extracting multi-scale feature information in medical image segmentation [21], several methods have been proposed to address these problems by using multi-scale features [22][23][24]. For instance, Li et al [23] proposed a multi-scale convolutional channel attention mechanism to generate detailed local features and integral global features.…”

Section: Introductionmentioning

confidence: 99%

MSGFNet: Multi-Scale Gated Fusion Network for Remote Sensing Image Change Detection

Wang,

Hao

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

Change detection (CD) stands out as a pivotal yet challenging task in the interpretation of remote sensing images. Significant developments have been witnessed, particularly with the rapid advancements in deep learning techniques. Nevertheless, challenges such as incomplete detection targets and unsmooth boundaries remain as most CD methods suffer from ineffective feature fusion. Therefore, this paper presents a multi-scale gated fusion network (MSGFNet) to improve the accuracy of CD results. To effectively extract bi-temporal features, the EfficientNetB4 model based on a Siamese network is employed. Subsequently, we propose a multi-scale gated fusion module (MSGFM) that comprises a multi-scale progressive fusion (MSPF) unit and a gated weight adaptive fusion (GWAF) unit, aimed at fusing bi-temporal multi-scale features to maintain boundary details and detect completely changed targets. Finally, we use the simple yet efficient UNet structure to recover the feature maps and predict results. To demonstrate the effectiveness of the MSGFNet, the LEVIR-CD, WHU-CD, and SYSU-CD datasets were utilized, and the MSGFNet achieved F1 scores of 90.86%, 92.46%, and 80.39% on the three datasets, respectively. Furthermore, the low computational costs and small model size have validated the superior performance of the MSGFNet.

show abstract

“…The direct classification approach aims to identify both the region and category of changes using an end-to-end neural network architecture. Two principal types of such end-to-end change detection networks exist: early fusion and late fusion [24]. Early fusion architectures integrate bi-temporal images as multi-channel inputs, tailored to fit semantic segmentation networks.…”

Section: A Deep Learning-based Change Detectionmentioning

confidence: 99%

Land Cover Change Detection Based on Vector Polygons and Deep Learning With High-Resolution Remote Sensing Images

Zhang,

Liu,

Niu

et al. 2024

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Vector Polygons are valuable survey data, serving as crucial outputs of national geographical censuses and a fundamental data source for detecting changes in geographical conditions. Current remote-sensing image change detection methods rely on comparing images but overlook abundant historical vector results, struggle with model generalization, and lack adequate samples. Consequently, change detection remains a manual process primarily, unable to meet the requirements for automated and efficient monitoring of standardized geographical conditions. Hence, this paper proposes a change detection method for land cover vector polygons based on high-resolution remote sensing images and deep learning. Initially, the enhanced simple linear iterative clustering (SLIC) algorithm is applied to segment dual-temporal images from identical regions. Subsequently, an annotated dataset is generated using a multi-scale extraction, cropping-with-inpainting approach. Next, datasets derived from pre-and post-temporal images are used for training and testing, respectively, and the training set is purified by using twoclassifier cross-validation. Finally, an improved object-oriented convolutional neural network (CNN) model performs finegrained scene classification. The change rules and postprocessing method are then integrated to identify changed vector polygons. To validate the effectiveness and superiority of the proposed method, we conducted experiments on land cover change detection using datasets from two study areas. The results indicate that the proposed method achieves precision and recall rates of 91.89% and 94.44% on dataset-1, respectively. Similarly, in dataset-2, the precision and recall rates reach 87.59% and 91.41%, respectively. These findings demonstrate the method's efficacy in detecting changed vector polygons, reducing manual intervention, and enhancing detection efficiency.

show abstract

Joint Variation Learning of Fusion and Difference Features for Change Detection in Remote Sensing Images

Cited by 16 publications

References 48 publications

Change Detection in Urban Point Clouds: An Experimental Comparison with Simulated 3D Datasets

Change Detection in Urban Point Clouds: An Experimental Comparison with Simulated 3D Datasets

MSGFNet: Multi-Scale Gated Fusion Network for Remote Sensing Image Change Detection

Land Cover Change Detection Based on Vector Polygons and Deep Learning With High-Resolution Remote Sensing Images

Contact Info

Product

Resources

About