Research on house change detection algorithm based on satellite remote sensing and big data

In order to meet the monitoring needs of dynamic changes in housing under the background of rapid urbanization, a housing change detection algorithm that integrates deep learning and multi-scale features is proposed based on multi-source satellite remote sensing data. The study utilized an improved deep residual U-Net model, combined with Landsat 8 OLI, Sentinel-2 MSI, and Sentinel-1 SAR data, to achieve high-sensitivity detection of small-scale and subtle changes.