Industrial part defect detection method based on deep learning

Abstract

The invention discloses an industrial part defect detection method based on deep learning, and relates to the field of industrial quality inspection, and the method mainly comprises the steps: obtaining a preset number of industrial part original images and defect marking graphs after defect marking; obtaining a feature map after convolution pooling processing according to the defect labeling map, fusing the feature map with the output of each pooling layer in the pooling stage, and obtaining a segmentation network by using the initial convolution kernel; adjusting the size of the convolution kernel to train the segmentation network in sequence; performing classification training on output results of the corresponding segmentation networks according to the original images and the defect labeling graphs to obtain classification networks; and judging the defect degree, defect position and defect type of the original image of the industrial part according to the segmentation network and the classification network. According to the method, the problem of defect segmentation is converted into the problem of classification through sequential pooling-up-sampling-fusion processing, and the advantage that the convolutional neural network is good at classification is utilized, so that the high efficiency of defect marking and classification of industrial parts is realized.

Description

technical field [0001] The invention relates to the field of industrial quality inspection, in particular to a method for detecting defects in industrial parts based on deep learning. Background technique [0002] In the industrial production process, the defect detection on the surface of the workpiece is the most important. The most commonly used method is manual visual inspection, but it requires training for employees, and a small number of factories also use ResNet50 deep convolutional network and FPN structure to build feature pyramids to find defect features, and use RetinaNet to build defect detection head modules , using DeepLab V3 or U-NET to build a semantic feature extraction module to find defective labeled images, but they all ignore the most important OK / NG detection and cannot customize the severity of the detection, or use a separate OK / NG The NG classification model is detected by, for example, ImageNet for classification. [0003] The artificial visual m...

AI Technical Summary

Problems solved by technology

One solution is to use Deeplab v3 to build a semantic feature extraction module, but how to find a suitable rate and avoid model degradation while increasing the receptive field to solve the grid effect is a problem that requires multiple experiments to obtain results, which cannot satisfy The speed requirements of Industry 4.0, and the Deeplab v3 has a huge amount of parameters, the training time and the early data required for training are large, and the later detection time is also very long, which cannot well meet the real-time requirements of the industrial assembly line; another solution is to use ResNet50. However, the parameters of the model are still huge, and professional GPUs need to be connected in parallel and trained for a few days to achieve better results. Althou

Images