Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

A Novel Hyperspectral Waveband Selection Algorithm for Insect Attack Detection

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  Transactions of the ASABE. 55(1): 281-291. (doi: 10.13031/2013.41238) @2012
Authors:   Y. Zhao, X. Xu, F. Liu, Y. He
Keywords:   Brown planthoppers, Hyperspectral waveband selection algorithm (HWSA), Least squares support vector machine (LS-SVM), Spectral signature

A novel hyperspectral waveband selection algorithm (HWSA) is proposed and applied to detect the injury severity of rice plants caused by brown planthoppers. Rice plants that were controlled or injured by brown planthoppers were sampled by the hyperspectral system. After preprocessing, the instability index (ISI) was calculated in order to measure the sensitivity of wavelengths to spectral variability, and the tradeoff index (TI) was set to remove insensitive wavelengths. The optimal wavelengths were then selected and used as inputs of the least squares support vector machine (LS-SVM) model, and the percentage of injured pixels was calculated. Different combinations of optimal wavelengths were obtained to satisfy different accuracy requirements. The wavelengths of 543.11, 568.33, and 602.35 nm were the most optimal combination, resulting in classification accuracy of 90%. The combination of 16 wavelengths, in which the wavelengths of 543.11, 568.33, and 602.35 nm were included, led to ideal classification accuracy of 98% with a suitable number of wavebands.

(Download PDF)    (Export to EndNotes)