Hyperspectral Data Processing: Algorithm Design and Analysis

29.6 Conclusions

This chapter presents a new application of wavelet in hyperspectral signature characterization. In particular, a WSCA is developed for signature self-correction, self-tuning, self-denoising, self-discrimination, self-classification, and self-identification. The experimental results demonstrated that the proposed algorithm WSCA performs effectively in self-correction, self-tuning and self-denoising with high accuracy for a given reference. Most importantly, the results also show that WSCA can successfully self-classify and also self-identify mixed hyperspectral signatures and subpixel targets.