Detecting location changes accurately and robustly is a fundamental requirement in many computer vision applications, including surveillance, autonomous navigation, and assistive technologies. Traditional approaches based on single video images often suffer from depth ambiguity and sensitivity to illumination or viewpoint changes. The proposed system builds upon a baseline monocular framework by integrating stereo matching techniques and depth map analysis. Our improved framework combines deep learning–based place recognition with stereo vision–based depth estimation to enable robust, depth-aware change detection. By leveraging disparity information from stereo image pairs, our method provides improved spatial understanding and depth-aware scene comparison. The approach was validated on the KITTI Stereo Vision Benchmark and a custom indoor dataset, demonstrating improved reliability in detecting added obstacles while maintaining low false positive rates. Future work will investigate alternative methods for comparing depth maps, including histogram-based and region-based differencing, to further enhance accuracy in complex indoor environments.