Application Of Groung-based Hyperspectral Remote Sensing To Detect Spider Mite

Hyperspectral remote sensing inject new energy into the application of remote sensing on agriculture through it's more bands and abundant data. It is very important that monitoring the crop growth status and the diseases, insect pests' damage to them. In 2005 leaf and canopy reflectance of cotton infected with different level cotton spider mites were measure using ASD Fieldspec FR2500 instrument in Shihezi Xinjiang. At the same time the index of spider mites was investigated and the photosynthesis parameters were measured in the range of spectrum measuring. The spectral feature and difference of cotton spider mites at visible light ,near infrared and middle infrared region and the changes of red edge parameters were studied qualitatively and quantitatively. Meanwhile, the correlations between the index of spider mites and the canopy spectrum and the photosynthesis parameters were analyzed. The results are as follows: The spectral reflectance decreased in near infrared region with the increase of infestation and this change is most significant between leaf spectral reflectance and canopy spectral reflectance. The changes of leaf and canopy red edge were different. At leaf level the red edge was remarkably moved to short wavelength with more heavy infestation, but at canopy level the red edge was unremarkably moved to short wavelength. The correlations were significant between the index of spider mites and red edge swing(dλred) , red edge minimum swing (dλmin), the ratio of red edge swing to minimum swing (dλred /dλmin) , red edge area (∑dλ680～750), so the red edge parameters could be detected cotton spider mites. The pigment had good correlation with the index of spider mites and canopy spectrum. Finally, we defined the DVI, RVI, NDVI by 634 nm,694 nm and 767 nm bands that were significant with the index of spider mites that have good correlation with the the index of spider mites. So those models can be use to support the monitoring the index of spider mites. Meanwhile, we constructed multi-band model to diagnose the index of spider mites by 694 nm and 767 nm bands. This research provided theoretic foundation to further monitor cotton spider mites at large scale using airborne and airspace remote sensing data.