“In response to the limitations of ordinary skip connections in remote sensing image change detection and the insufficient feature extraction ability of encoders, the research team proposes a new CBAM UNet++network structure. This network combines the change detection requirements of high-resolution remote sensing images and effectively enhances the network's ability to capture full-scale change information and the encoder's ability to learn salient features by introducing a coupled attention mechanism CBAM. In the experiment, the research team validated using two high-resolution remote sensing image change detection datasets with different types of changes. The results showed that CBAM UNet++achieved the highest accuracy on the LEBEDEV multi terrain change dataset, with F1 and OA values reaching 88.9% and 97.3%, respectively. On the LEVIR-CD building change dataset, the second highest accuracy was also achieved, with F1 and OA values of 86.7% and 96.8%, respectively. These achievements demonstrate the superior performance of CBAM UNet+++in remote sensing image change detection. In addition, the study also found that CBAM UNet+++can selectively obtain deep semantic information, and its qualitative results are superior to other benchmark networks. This discovery provides new ideas and methods for the field of remote sensing image change detection, and lays the foundation for subsequent research.”