Abstract
To further improve target detection performance of hyperspectral image, this paper presents a novel method named multi-scale analysis-based target detection (MATD). The proposed method first applies multi-scale wavelet analysis technology to extract multi-scale features of hyperspectral data. Then, these features are converted into a tensor form, and is processed and analyzed by using tensor analysis method. Through solving the tensor subspace, the reduced-dimension feature coefficients can be extracted. Finally, based on these feature coefficients, a better target detection result can be obtained by using the detection algorithm. Experimental results of real world hyperspectral data show that the proposed MATD method can effectively improve detection performance.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call