Estimation of Severity Level of Wheat Powdery Mildew Based on Canopy Spectral Reflectance

Abstract

Understanding spectrum characteristics and sensitive bands of wheat infected by Blumeria graminis f.sp.tritici(Bgt) and estimating the disease severity will provide a basis for monitoring and subsequently accurately controlling powdery mildew in wheat planted in large scales using aerial remote sensing.Canopy reflectance of winter wheat infected by artificial inoculation of Bgt with different severity levels was measured in disease nursery,field,and pot experiments,and the severity levels in different growth phases were also investigated at the same time.The results indicated that spectrum reflectance increased significantly in visible light region(350–710 nm) with the increase of disease severity level,and the light region between 580 nm to 710 nm was the sensitive bands to wheat powdery mildew,which varied greatly in near-infrared region(710–1100 nm) across treatments with different disease severity levels.However,the correlation between spectrum reflectance and conventional disease index(DI) was low.When the conventional DI was replaced by modified DI,the correlation was improved significantly.An integrated linear regression equation of disease severity level to red edge width(Lwidth) described the dynamic pattern of disease severity level in wheat,with R2of 0.811 and relative error(RE) of 17.7%.Besides,correlation coefficients between spectral parameters(mND705,SIPI,CTR2,and TSAVI) and modified DI were higher than 0.6.No common integrated regression equation could be established due to poor compatibility among experiments.The relative spectral indices(ΔMSAVI and ΔmND705) exhibited high correlations(R2 0.76) with the disease sensitive level,with RE values of 18.4% and 19.4%,respectively.The results suggested that the models with sensitive spectral indices could retrieve and forecast the disease severity level accurately in a large area.