Real-time high-precision detection method for appearance defects of power adapter

Abstract

The invention relates to a real-time high-precision detection method for appearance defects of a power adapter, and the method comprises the steps: fixing a plurality of high-resolution cameras at the left, right and right above an assembly line, and collecting the appearance images of the power adapter; extracting the defect parts in the defect pictures, intercepting a rectangular marking frame part from an original picture, and then enhancing the number and types of the defect pictures in an image rotation mode; and performing multi-stage cross-scale convolutional neural network model training by using the expanded and enhanced defect data set. The method has the beneficial effects that the calculated amount can be reduced, the detection speed can be improved, the detection precision is improved through multi-scale feature fusion and rule filtering, a practical and available technical scheme is provided for real-time detection of the surface defects of the mobile phone power adapter in an industrial assembly line, and the detection range of a defect target is effectively expanded.

Description

technical field [0001] The invention belongs to the field of defect detection of industrial products, in particular to a power adapter appearance defect detection using computer vision technology. Background technique [0002] In recent years, with the rapid update iteration and development of intelligent terminal equipment, the production scale of power adapters for equipment in industrial assembly lines has also expanded. However, in the production process, the surface of power adapters may be dirty, scratched, etc. due to improper operation. defect. Automatic defect detection of power adapters in the production line has increasingly become a hot topic of research, especially for irregular and small defect detection, which can ensure product quality. [0003] With the rapid development of machine vision technology, people's production and living environment have been completely changed. Machine vision detection technology combines machine vision and automation technology...

AI Technical Summary

Problems solved by technology

[0004] However, the existing deep learning methods have great limitations and are difficult to apply to the appearance inspection of power adapters for smart terminal equipment: First, deep learning methods are data-driven and require a large number of labeled defect samples. It is difficult to collect a large number of defective samples in the production line; secondly, the characteristics of the training samples need to cover the target samples used for detection in the application, but because the defects have large size differences and random appearance characteristics; finally, due to industrial The defect detection in the software pays great attention to real-time and no missed detection, and most of the deep learning will have different degrees of missed detection

Therefore, in the industrial assembly line, deep learning often has no way to replace manual work and meet the needs of practical applications.

Images