Method for determining number of product quality accidents based on convolutional neural network

The invention discloses a method for determining the number of product quality accidents based on a convolutional neural network. The method for determining the number of product quality accidents based on the convolutional neural network comprises the following steps: constructing an appearance quality data set; constructing a quality evaluation model; and product quality accident quantity analysis. According to the invention, the preset model and the preset optimization model are trained and verified through the constructed appearance quality data set to obtain the initial quality evaluation model, and model screening is carried out according to the model performance coefficient and the model comprehensive coefficient to obtain the quality evaluation model; then obtaining a corresponding appearance defect ratio and a significant appearance defect ratio according to product appearance defect data of a to-be-inspected product obtained through detection of the quality evaluation model, and finally determining