Workpiece defect detection method and device

Abstract

The invention provides a workpiece defect detection method and device, and the method comprises the following steps: obtaining defect marking data of a to-be-detected workpiece; enhancing the defect labeling data by adopting an image pixel decomposition and reconstruction algorithm; constructing a defect detection model according to the enhanced defect labeling data; and performing defect detection on the to-be-detected workpiece according to the defect detection model. According to the invention, richness of data used for detection can be ensured, so the performance of a detection model can be ensured, and detection precision and stability are improved.

Description

Technical field [0001] The present invention relates to the field of defect detection, and more particularly to a workpiece defect detection method and a workpiece detecting device. Background technique [0002] Convolutional neural network model is currently a very important technique in the field of workpiece defect detection. In convolivation neural network, excellent annotation of workpiece defect data sets have a great impact on the performance of the final quality inspection model. However, the label of the current workpiece defect data is generally completed, but due to the promotion of the cost of employment and the rareity of expert experience in the quality inspection, the output of the workpiece defective data of excellent labeling is about 100, to achieve sufficient convolutional neural network training. The amount of data required is required for a month, and if it starts training in the amount of data, the final workpiece defect detection model is difficult to obtai...

AI Technical Summary

Problems solved by technology

However, the "new data set" generated by this strategy is still not substantially different from the original data set feature mode, and the inter-class feature mode remains unchanged, which is difficult to be used for workpiece defect detection tasks

Images