An improved generative adversarial network based image defect segmentation method for eddy current inspection of aluminum plates

Abstract

The invention relates to an image defect segmentation method for eddy current detection of aluminum plates based on an improved generative confrontation network, and belongs to the technical field of image defect segmentation for eddy current detection of aluminum plates. First, the surface defects of the aluminum plate are detected by the C-scan detection method through the eddy current inspection platform; secondly, the collected signals are processed to obtain the eddy current inspection image of the surface defect of the aluminum plate; then the generative adversarial network model for the defect segmentation of the eddy current image of the aluminum plate is constructed. It consists of a generator and a discriminator. The generator adopts the idea of ​​the U-net model, and connects the low-level features with the corresponding high-level features. Finally, in order to enhance the target features and suppress the background features, the attention module is used before the fusion of high-level and low-level features. , to adjust the weights when splicing low-level features and high-level features. Compared with the traditional image segmentation method, the invention improves the utilization of image feature information, and the segmentation image is more accurate, and meanwhile, it still has higher recognition ability under noise interference.

Description

technical field [0001] The invention relates to an image defect segmentation method for eddy current detection of aluminum plates based on an improved generative confrontation network, and belongs to the technical field of image defect segmentation for eddy current detection of aluminum plates. Background technique [0002] In production and life, aluminum plate in metal sheet is widely used as an important industrial raw material. During its application, it is easily affected by various environmental factors, which inevitably causes crack defect damage, which seriously restricts the normal and safe operation of aluminum plate. The existence of defects not only affects the appearance of the aluminum sheet material, but also seriously reduces the corrosion resistance, wear resistance and other characteristics of the material. If the existence of defects is not detected in time, it may lead to serious accidents. Nondestructive testing methods currently available for metal mate...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is to provide a method for image defect segmentation of aluminum plate eddy current detection based on improved generative confrontation network, to enhance the robustness of the image segmentation algorithm, and to eliminate the excessive dependence of the model on the number and distribution of training samples. There is background noise interference in the defect area of ​​the eddy current image, which solves the problem that the edge area of ​​the aluminum plate defect image is not easy to identify and the segmentation effect needs to be improved

Images