Spray image classification and quality detection method based on VGG-16 model

The invention provides a spray image classification and quality detection method based on a VGG-16 model, and the method comprises the following steps: S1, image preprocessing: carrying out the gray enhancement of all original images, and inputting the original images into a convolutional neural network for training; s2, model building: using a VGG-16 model in a convolutional neural network, loading a pre-trained weight, freezing weight parameters of a convolutional layer, removing an original full connection layer of VGG-16, manually adding a self-defined full connection layer, and connecting the manually added full connection layer and the original VGG-16 model; s3, model training: performing data enhancement on the images of the training set, and inputting the images into a model for training; s4, loading the test set; S5, measuring the spraying angle; and S6, measuring the SMD (Surface Mount Deposition) of the Soxt average diameter of the spraying liquid drop. The method has the beneficial effects that th