[Study on hyperspectral estimation of pigment contents in leaves of cotton under disease stress].

The spectrum reflectance and pigment contents of cotton leaves infected with Verticillium wilt were measured in cotton disease nursery and field in different growth phases, and severity level of Verticillium wilt was investigated. The correlation between pigment contents of cotton leaves with Verticillium wilt and spectra reflectance, the first derivative of reflectance and spectral characteristic parameters were analyzed respectively. The estimation models about leaves pigment contents of disease cotton were established and tested. The results indicated that the correlations were best significant between chlorophyll a, b and a+b contents of leaves and spectral reflectance in visible wave bands, the first derivative spectrum at the wavelength regions of blue edge, yellow edge and red edge, and all spectral characteristic parameters (excluding red edge swing Dr). The models of transformed chlorophyll absorption in reflectance index (TCAR) and the new model of normalized difference vegetation index (NDVI [702,758]) had best estimated precision, and the relative errors were in average within 1.3%. Given that the model of NDVI [702,758] is very simple and practical, it was commended as a best model to estimate chlorophyll a, b and a+b contents of disease cotton. This study shows that leaves hyperspectra data can be used to estimate the pigment contents of cotton leaves quantitatively. This conclusion has also great practice and application value for monitoring the growth state and disease influence evaluation on cotton by using hyperspectral remote sensing.