Hyperspectral remote sensing of yellow mosaic severity and associated pigment losses in Vigna mungo using multinomial logistic regression models

摘要: Abstract Yellow mosaic disease (YMD) has been a serious threat to blackgram cultivation especially during post-monsoon season. Visual assessment of severity is qualitative and time consuming. Rapid non-destructive estimation YMD by hyperspectral remote sensing not attempted so far on any its hosts. Field studies were conducted for two seasons with eight genotypes having differential response YMD. Comparison mean reflectance spectra the healthy infested leaves showed changes in all broad band regions. However, sensitivity analysis narrow-band data revealed sharp increase from diseased compare at 669 (red), 505 510 nm (blue). ANOVA significant decrease leaf chlorophyll ( p R 2 ) between percent one nm wavelength interval, individual bands (R 571 ; 705 ratios /R 721 593 highest values selected. These linear relationship SPAD readings range 0.781–0.814) spectrometric estimates total content 0.477–0.565). Further, was stronger compared single bands. With optimal spectral as inputs, prediction models built using multinomial logistic regression (MLR) technique. Based model fit statistics, found better than . Validation MLR an independent test set that overall percentage correct classification plant into categories essentially same both (68.75%). dependent variable greater accuracy it gave lower standard errors slopes β G 9.79–36.73) highly intercept slope