DL-CD: A Deep Learning Generative Models for Accurate Change Detection in Remote Sensing Data

In recent years, deep learning technologies have emerged to be the major approach for the processing of remote-sensing images. Because of its numerous useful uses, deep learning has been implemented for autonomous change detection as well as accomplished significant success. Deep learning (DL) approaches have been demonstrated effective in detecting changes due to their usage in real life. A few investigators have even suggested that deep learning algorithms prevail over traditional methodologies and improves change detection accuracy. The study focuses on deep learning techniques, including supervised, unsupervised, as well as semi-supervised methods for various change detection models. Data collected from various satellites are readily available nowadays towards the rapid advancement of remote sensing (RS) technology. The images above considerably improve the data resources for change detection. CD is a procedure for detecting differences between photographs captured at different time intervals. The system combines a convolutional neural network (CNN) building construction, attention mechanisms, as well as multi-scale recognition of features to improve sensitivity to tiny changes simultaneously lowering false positives.