Syringe Defect Detection Method Based on Semantic Segmentation

Abstract

The present invention provides a syringe defect detection method based on semantic segmentation, including the following steps, step S1: injector image acquisition, step S2: convert the json file into a png mask image with code, step S3: image processing, that is, the syringe data set is completed production, step S4: build a fully convolutional neural network model, input the training data set into the semantic segmentation network, and obtain a convergent segmentation model after iterating the model parameters, step S5: test the model, and input the test data set into the semantic segmentation network In the segmentation network, obtain the syringe segmentation map, compare the defects on the original image, and judge the segmentation accuracy of the model. Step S6: Export the model file, perform defect detection through the semantic segmentation network, and automatically determine whether the syringe is defective. Compared with traditional manual detection , faster and more accurate, improving the degree of automation of the manufacturing process, can greatly reduce the cost of enterprises, while increasing productivity.

Description

Technical field [0001] The present invention relates to the field of vision technology thereof, in particular to a syringe defect detection method based on semantic segmentation. Background [0002] In the manufacturing process of syringes, there will inevitably be defective products, so the processing plant needs to select these defective products, and syringe manufacturers have used to use the human eye to judge their manufacturing defects, because the number of syringe manufacturing is very large, completely rely on people to detect manufacturing defects, not only low detection efficiency, but also high cost, testing work requires a lot of manpower and material resources, the final test results will also affect the detection accuracy and efficiency because of human uncertainty. [0003] Therefore, it is necessary to provide a semantic segmentation-based syringe defect detection method to overcome this challenge. Contents of the Invention [0004] The present invention provid...

AI Technical Summary

Problems solved by technology

[0004] The invention provides a syringe defect detection method based on semantic segmentation, which overcomes the traditional artificial syringe defect detection method, uses the semantic segmentation model for real-time syringe defect detection, and can be used for non-destructive and non-contact detection

Images