Tensor Regression and Image Fusion-Based Change Detection Using Hyperspectral and Multispectral Images

Change detection is a popular topic in remote sensing that generally constrained to two images captured at different times. However, the optimal type of image for change tasks has not yet been determined. The use only hyperspectral (HSIs) with low spatial resolution or multispectral (MSIs) spectral cannot obtain satisfactory results. In this paper, we propose fusion simultaneously HSIs and MSIs tensor regression-based method detect regions from fused time points. method, non-local couple CP decomposition (NCTCP) initially applied fuse MSIs. A difference then obtained by subtracting Thereafter, tensors are extracted used classify final Experimental results three real datasets suggest proposed substantially outperforms existing state-of-the-art methods as well any using single-source images.