[IEEE 2019 10th Workshop on Hyperspectral Imaging and Signal Processing: Evolution in Remote Sensing (WHISPERS) - Amsterdam, Netherlands (2019.9.24-2019.9.26)] 2019 10th Workshop on Hyperspectral Imaging and Signal Processing: Evolution in Remote Sensing (WHISPERS) - Closed-Loop Moving Windows Wavelength Selection Method With Application To Near-Infrared Spectroscopic Analysis

摘要： Wavelength selection plays a very important role in near-infrared (NIR) spectroscopy, which has always been an important research direction. Based on the reverse-deletion waveband idea, the closed-loop moving window-partial least squares (closed-loop MW-PLS) method was proposed, which can eliminate interference wavelengths and enable flexible multi-band wavelength selection. NIR analysis of soil organic matter was taken as an example to evaluate the performance of closed-loop MW-PLS. And the MW-PLS was also used for comparison. The proposed algorithm traversed all the sub-bands of the original range to perform forward and backward optimization. The forward optimization was just the original MW-PLS. Therefore, the closed-loop MW-PLS completely covered the MW-PLS. The results of soil organic matter indicated that the close-loop MW-PLS was strictly superior to original MW-PLS, and the method extension was meaningful and had no increase in operational complexity. We believe that the closed-loop MW-PLS method will have a wider application.