Nondestructive Identification of Litchi Downy Blight at Different Stages Based on Spectroscopy Analysis

Litchi downy blight caused by Peronophythora litchii is the most serious disease in litchi production, storage and transportation. Existing identification technology has difficulty identifying sufficiently early, resulting economic losses. Thus, use of diffuse reflectance spectroscopy to identify at different stages disease, particularly achieve early blight, very important. The spectral data fruits inoculated with P. were collected wavelength range 350–1350 nm. According duration inoculation expert evaluation, they divided into four categories: healthy, latent, mild severe. First, SG smoothing method derivation used denoise curves. Then, screening methods competitive adaptive reweighted sampling (CARS) successive projections algorithm (SPA) compared verify that SPA was more effective. Eleven characteristic wavelengths selected, accounting for only 1.1% original data. Finally, tested six classification models, their accuracy calculated. Among them, ANN model performed best, an 90.7%. results showed spectroscopic potential stages, providing technical support subsequent development related automatic detection devices.