Improved Spectral Detection of Nitrogen Deficiency and Yellow Mosaic Disease Stresses in Wheat Using a Soil Effect Removal Algorithm and Machine Learning

Wheat yellow mosaic disease is a low-temperature and soil-borne disease. Crop infection by the virus can lead to severe yield economic losses. It easily confused with nitrogen deficiency based on plant’s morphological characteristics. Timely detection crop management in field require precise identification of stress types. However, often underappreciated. were investigated wheat physiological biochemical experiments conducted collect agronomic indicators, four years reflectance spectral data at green-up jointing collected, then studies for stresses carried out. The continuous removal (CR), first-order derivative (FD), standard normal variate (SNV), separation soil vegetation (3SV) preprocessing methods 96 indices evaluated. threshold method variance inflation factor (TVIF) used as feature selection combined machine learning develop method. results show that most sensitive wavelengths are found 725–1000 nm region, while sensitivity spectrum 400–725 region lower. PRI670,570, B, RARSa detect leaf stress, OA Kappa values 93.87% 0.873, respectively, which best index. A 3SV-TVIF-SVM was proposed, using 96.97% 0.931, validation. study provide technical support theoretical basis accurate control fertilizer field.