[Study on hyperspectra estimation of pigment contents in canopy leaves of winter wheat under disease stress].

The canopy reflectance of winter wheat infected with stripe rust was measured in the field through artificial inoculation, and the pigment contents of the wheat leaves were determined indoor. The correlation between pigment contents and canopy hyperspectra data and the first derivative data of the disease wheat were analyzed respectively. Using linear and non-linear regression methods, and choosing a part of samples, the estimation models about pigment contents of disease wheat were built. Through the test of the other part samples, the result shows that the model containing the normalized value of the sum of first derivative within green edge (SD(g)) and the sum of first derivative within red edge (SD(r)) is the best one. The model was used to estimate the contents of chlorophyll a and chlorophyll b and carotenoid of the disease wheat, and the relative errors were 17.0%, 16.3% and 12.4%, respectively. This study shows that canopy hyperspectra data can be used to estimate the pigment contents of crops leaves and the estimation precision is high. This conclusion has great practice and application value to monitor the grow-ing way of and disease influence on crops by using hyperspectral remote sensing.