Scratch detection method based on deep convolutional neural network and image segmentation

Abstract

The invention discloses a scratch detection method based on a deep convolutional neural network and image segmentation, and the method comprises the steps: designing a multi-feature fusion module according to the morphological characteristics of a scratch, adding the multi-feature fusion module into a deep convolutional neural network structure, carrying out the recognition and positioning of the scratch, and extracting a scratch prediction frame image; and calculating principal component points of the scratch prediction frame image by using a principal component analysis method, and performing 8-neighborhood region growth on the scratch prediction frame image subjected to adaptive mean segmentation by taking the principal component points as growing points to realize accurate segmentation of scratch pixels. According to the invention, high-precision identification, positioning and segmentation can be carried out on the scratch, and convenience is provided for subsequent precision measurement of the size of the scratch.

Description

technical field [0001] The invention belongs to the field of defect automatic detection and precision measurement, and in particular relates to a scratch detection method based on a deep convolutional neural network and image segmentation. Background technique [0002] The improvement of modern industrial intelligence puts forward higher and higher requirements for product surface quality. Usually, product surface quality problems will appear in areas with uneven physical or chemical properties on the surface of the product, such as scratches on key parts of automobiles and trains, weld seams of laser welding products, concrete cracks, chip packaging scratches, etc., according to the morphological characteristics of these defects It can be attributed to scratches, and effective detection of surface scratches is of great significance to ensure product quality and performance, improve the new rate, and ensure the safety of use. As a non-contact and non-damaging automatic dete...

AI Technical Summary

Problems solved by technology

[0004] Traditional scratch detection methods based on machine vision usually rely on manually designed features, and need to optimize the design of the algorithm according to the actual detection requirements and application scenarios. These algorithms focus on the segmentation and extraction of scratches, while ignoring their identification , resulting in susceptibility to noise or other types of imperfections

In recent years, scratch detection methods based on deep learning have gradually emerged to make up for the shortcomings of traditional methods in scratch recognition. This method requires training and learning of a large number of labeled data, focusing on the recognition and location of scratches. It is difficult to meet the precision detection requirements for the accuracy of scratch pixel segmentation

Images