Robust Unsupervised Geo-Spatial Change Detection Algorithm for SAR Images

Geo-spatial change detection plays a crucial role in identifying alterations on the earth's surface over time intervals. This paper introduces a novel unsupervised grid graph generation algorithm specifically designed for change detection using Synthetic Aperture Radar (SAR) images. The proposed technique encompasses a multi-step process: starting with an improved log-ratio based difference image generation, followed by shortest path vector computation and thresholding from the generated grid graph utilizing Dijkstra's algorithm. Furthermore, a voting-major based tuning approach is employed to effectively eliminate any residual noisy corruptions. The algorithm is rigorously evaluated on qualitative and quantitative scales using standard image data and is benchmarked against state-of-the-art methods. Results demonstrate the resilience and superior performance of the proposed method in comparison to existing approaches in terms of robustness and speed.