A General Model for Urban Change Detection Based on Remote Sensing Images

With the rapid development of aerospace technology and satellite technology, remote sensing image change detection, as one of the important branches of remote sensing technology, has also been greatly developed and improved. This paper introduces a comprehensive framework for identifying alterations in roadways within satellite imagery, utilizing a deep learning approach based on the deep forest architecture. The approach outlined in this paper leverages the multi-scale scanning attributes of deep forests along with the data categorization prowess of cascade forests, thereby enhancing the efficacy of current road change detection techniques in a broad sense. In this general model, common remote sensing image change detection methods are first used for preliminary detection to obtain the detection results to be optimized, and then the classification effect of the deep forest is used to obtain optimized road change detection results. Experiments were conducted on a variety of common remote sensing road change detection models and change detection datasets including remote sensing images. The results show that the proposed general method achieves significant improvements in indicators such as precision, F1 score, and recall.