Multi-mode morchella esculenta growth state rapid discrimination method

The invention discloses a multi-modal morchella growth state rapid discrimination method, which comprises the steps of carrying out feature extraction on an image I based on a Mask R-CNN backbone network and obtaining a feature map, carrying out boundary feature extraction on the image I based on an NSCT algorithm and a Canny operator and obtaining a sub-band set, unifying the spatial dimensions of the feature map and the boundary feature map, and carrying out rapid discrimination on the growth state of morchella. Calculating entropy information of each piece of morchella esculenta environmental data based on discrete entropy and a clustering binning method, generating a high-resolution entropy information graph of the environmental data based on an inverse distance interpolation method and the entropy information of the environmental data, performing feature fusion on the feature graph, the boundary feature graph and the high-resolution entropy information graph, and representing the fused features; and obta