Estimating the severity of apple mosaic disease with hyperspectral images

Abstract

Soil Plant Analysis Development (SPAD) Chlorophyll Meter reading was used to effectively characterize chlorophyll content, which is an important indicator of the health status of plant leaves. In this study, the hyperspectral images of apple leaves infected by apple mosaic virus (ApMV) were captured, and their SPAD values were measured. The spectral reflectance of leaves with varying degree infection of disease is significantly different. In particular, the reflectance in visible wavebands of leaves with a more serious infection was higher than that of leaves with a less severe infection. Several hyperspectral vegetation indices were highly correlated with the SPAD values of apple leaves (correlation coefficient > 0.9). Models were established to estimate apple foliar SPAD values based on these vegetation indices. Among the models, the multivariate regression model with partial least square regression (PLSR) method achieved the highest accuracy. The SPAD value of a whole apple leaf was calculated from its SPAD distribution image and used as a quantitative index to represent the health status of an apple leaf. Furthermore, the SPAD value of a whole apple leaf could also be estimated rapidly and accurately by extracting the spectral average value of the whole leaf using a simple model. It can be used as a rapid detection method of SPAD values of apple leaves to monitor and describe the health conditions of apple leaves quantitatively. Keywords: hyperspectral image, apple leaf, mosaic disease, SPAD, plant health detection DOI: 10.25165/j.ijabe.20191204.4524 Citation: Ban S T, Tian M L, Chang Q R. Estimating the severity of apple mosaic disease with hyperspectral images. Int J Agric & Biol Eng, 2019; 12(4): 148–153.