Deep learning-based detection method and control system for form and quantity of sugarcane seeds in sugarcane collection box

The invention discloses a method for detecting the form and quantity of sugarcane seeds in a sugarcane collecting box based on deep learning. The method comprises the following steps: step 1, acquiring image information of the sugarcane seeds in the sugarcane collecting box by using a camera device; step 2, calibrating the sugarcane seed image, making a data set, and dividing the data set into a training set, a verification set and a test set according to a proportion; 3, performing iterative training on the calibrated training set and verification set by using a convolutional neural network to obtain an optimal weight file, and obtaining a sugarcane seed form and seed quantity network model; and step 4, using the trained sugarcane seed form and seed quantity network model, performing prediction verification on the distributed test set, and evaluating the network model. The detection method provided by the invention has the advantages of high recognition rate and high recognition speed. A control system is fo