Defect detection method for camera glass

Abstract

The invention relates to a defect detection method of camera glass, comprising the following steps: 1) collecting images: imaging products through a camera; 2) positioning detection areas: extracting product areas to be detected from images; 3) extracting suspicious defects: Extract suspicious defects with high contrast and low contrast in the product area; 4) Defect classification: After extracting suspicious defect areas, divide suspicious defect areas into different defect types according to the characteristics of defect areas; 5) Defect screening: Screen the divided defect areas, filter minor defects and leave serious defects according to actual needs; 6) Detect product defects. The invention is superior to the scratch detection effect of the usual methods such as Gaussian filtering and wavelet change, and greatly improves the algorithm efficiency, achieves a good scratch detection effect, and greatly reduces the processing time of the algorithm.

Description

technical field [0001] The invention relates to the technical field of visual inspection, in particular to a method for detecting defects on the glass surface of a camera. Background technique [0002] The mobile phone camera glass inspection system is a set of machine vision system added to the automatic production line. Machine vision uses a computer to achieve a visual function similar to the human eye. It obtains images through cameras, etc., and then converts them into digital image signals. processing, the software then synthesizes the information to make correct calculations and judgments, and finally controls the on-site equipment actions based on the recognition results. Although machine vision technology has been widely used in various fields at present, there are still certain limitations in its wide application due to its own imperfections or supporting technologies. The efficiency of the image processing algorithm and the detection effect of the algorithm are t...

AI Technical Summary

Problems solved by technology

[0003] At present, the application of machine vision inspection on lens glass is still very difficult. The problems are mainly reflected in the low detection efficiency, high missed detection rate and overkill rate.

The main reasons for these problems are: 1. The imaging methods of different types and different positions of defects are not the same, and it is difficult for the optical imaging system to ensure that all defect types can be well imaged

Too complex optical system design will increase the cost and affect the efficiency of detection

2. The most important factor affecting the visual inspection system is the image processing algorithm. The main problem of the existing technology is that the algorithm processing efficiency is insufficient, and the detection output per unit time cannot meet the requirements; the detection ability is not enough, and the detection effect of minor defects is not good. , it is easy to cause missed detection or misjudgment, and it is difficult to control the missed detection rate and false detection rate at a low level at the same time

However, the traditional background grayscale estimation method will be interfered by the vehicle at the edge of the product. In addition, the traditional method takes too long, which affects the detection efficiency of the algorithm.

As for the slight scratches on the glass, algorithms such as Gaussian filtering and wavelet transform are traditionally used for detection, but these algorithms consume too long time and are far from meeting the time requirements in actual use.

Images