Method And Application Of A Linear Defect Detection On Quality Of Passport

The feature of straight line or line segment in the image is an important thread to the visual perception and recognition of the naked eye.It corresponds to the outline or boundary of the pattern in the image and is the key to image analysis processing and further understanding of the image.It has more than great influence on the specific image feature extraction.Traditional line segment detection algorithms begin with image edge detection,which is a fundamental problem in digital image processing and inspection technology of computer vision.Line segment detection as a research area for edge detection has great help in detecting and extracting specific image features.Therefore,the relevant detection algorithm for line segment feature extraction has been a hot research topic.The problem of line segment feature detection of different shapes often appears in many practical applications,but the line segment detection algorithm that can fully meet the actual needs is almost none,and the existing algorithm must be improved according to the actual situation,so as to meet the requirements of engineering applications.At present,the mainstream line segment detection technologies are mainly divided into the following categories: detection methods based on Hough transform spatial mapping;detection methods based on image boundary tracking and geometric information;and rapid detection based on pixel-based gradient information to implement line segment extraction detection methods.This article proposes a method to detect passport quality defects.The method will use the line segment detection algorithm to detect the defect image,and use some pre-experienced experience to pre-process the color three channels of the image to be detected.As long as the relevant defect line segment is detected in one of the channels,the defect image can be quickly located and the defective product can be singled out.Therefore,the main work is as follows:1)A typical line segment detection algorithm such as Hough transforms detection algorithm,Burns edge detection algorithm,and LSD line segment detection algorithm is analyzed and compared.Finally,the LSD detection algorithm is selected to analyze and deal with the defect detection mentioned in the article.2)Image pre-processing analysis: For the extraction of defect features in the image,the results of the pre-processing will directly determine the effect of the detection algorithm.In this paper,the information of each channel of the color image will be fully used to perform the necessary preprocessing of the image.3)We apply the down-sampling to the detected images,and use multiscale analysis methods to reduce the effect of the background pattern on the detection results.At the same time,these linear defects are almost all vertical existing knowledge to improve the detection accuracy,and an improved line segment detection algorithm is proposed.Using the technique of affine transformation,the defects of the dirty plate are detected after calibrating the detected pattern,thereby finding the relevant defect patterns.Finally,a series of images were pre-processed with three colorful channels through simulation experiments,and the algorithm's effect was verified by using improved LSD algorithm detection and dirty plate defect detection.The experiment result shows that using this method to quickly locate and detect defects on such images is very helpful.