Abstract
Deep neural network (DNN) has been widely used in remote sensing image change detection (CD) in recent years. Due to the scarcity of training data, a large number of labeled data onto other fields become the source of DNN concept learning in remote sensing image CD. However, the distribution of features of the CD data and other data varies greatly, which prevents DNN from being better applied for one task to another. To solve this problem, a domain adaptive CD method based on segmentation map difference is proposed to this article, which includes the pretraining stage and the CD stage. In the pretraining stage, the domain adaptive UNet (Ada-UNet) is applied as the basic network of remote sensing image segmentation for network training with the purpose of learning the concepts of different features. In the CD stage, strict threshold segmentation results are used to train the channel attention network, which makes it more efficient to utilize the high-dimensional feature map. The probabilistic map generated by the three-channel attention networks is evaluated, and then it is used to accurately classify the changing pixels. In this article, experiments are carried out on datasets with different feature distributions. The results show that this method has strong domain adaptability and can greatly reduce the influence of the difference in feature distributions of the CD results.
Highlights
Remote sensing image is the main form of remote sensing data at present, which directly reflects the information of land utilization and land cover
We propose a new change detection method based on semantic segmentation feature maps to generate change maps from high-resolution remote sensing images
Adaptive semantic segmentation The change detection method proposed in this paper shows that the difference in feature distribution among different remote sensing image data sets can be solved by improving the structure of the feature extraction stage
Summary
Remote sensing image is the main form of remote sensing data at present, which directly reflects the information of land utilization and land cover. Change detection is one of the most important means in remote sensing image analysis and interpretation[4]. It aims to identify differences in remote sensing images of the same geographic location at different times. The other is based on the classification or segmentation method In this kind of method, images are classified by using neural network[11], and all kinds of generated feature images are fused and processed. The method based on classification requires DNN to extract image features. In the process of feature extraction, the ability of DNN to extract unbalanced features will be greatly reduced because the features of ground objects contained in a large range of remote sensing images are usually not evenly distributed [12]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
