Detection of anthracnose in tea plants based on hyperspectral imaging

Abstract

Anthracnose (Gloeosporium theae-sinesis Miyake) is an important and common foliar disease in tea plants and is a severe threat to tea quality and production. Hyperspectral imaging technology enables non-invasive, objective detection of the damages ca by foliar disease and offers significant potential for plant disease prevention and phenotyping. This study proposes a novel method for detecting anthracnose in tea plants based on hyperspectral imaging. By analyzing the spectral sensitivity, we identified disease-sensitive bands at 542, 686, and 754 nm and used these bands to create two new disease indices: the Tea Anthracnose Ratio Index (TARI) and the Tea Anthracnose Normalized Index (TANI). Based on an optimized set of spectral features, a strategy combining unsupervised classification and adaptive two-dimensional thresholding was developed to detect disease scabs. Compared with traditional pixel-based classification methods, the proposed method was not affected by leaf background differences and thereby provides an effective means for disease identification and damage analysis. The validation results gave an overall accuracy of 98% for identifying the disease at the leaf level and 94% at the pixel level. These results suggest that automated and accurate detection of anthracnose-infected tea leaves is possible by using hyperspectral imaging for practical tea-plant protection.