Multi-disease detection method in crop intercropping mode

The invention discloses a multi-disease detection method in a crop intercropping mode, and relates to the technical field of crop plant disease detection.Collected corn and soybean hyperspectral sample data are preprocessed through a first-order derivative method, and a multi-disease detection result is obtained through a competitive self-adaptive reweighting algorithm; the method comprises the following steps: selecting wavelength points with large absolute values of regression coefficients in a PLS (Partial Least Square) model by using an adaptive reweighted sampling technology, removing wavelength points with small weights, selecting a subset with the lowest RMSECV finger by using interactive verification, finding out an optimal variable combination, and finally detecting and identifying intercropping corn and soybean diseases by using a two-way long-short-term memory algorithm optimized by a dung beetle algorithm. Compared with the prior art, the method has the advantages that nondestructive detection of