Research Of Automatic Detection Algorithm Based On Machine Vision

In recent years,with the development of the industrial mechanization,more and more manufacturing industry began to improve itself from manual operation to automatic machinery production.Currently in the manufacturing industry,the production task of the sample page is manually sorting approach.However,the manual sorting method is inefficient and cannot endure a high-intensity continuous work.In order to improve the efficiency of the production,the sample page production and processing enterprises try to achieve the mechanical production instead of manual production,but it also need to overcome the problems encountered in automatic detection.Under the limited cost of investment,equipment production is not good enough to ensure the accuracy.So the sample page in the transmission process prone to offset,Situation,affecting the accuracy of detection.After a large research,an automatic detection algorithm is designed to adapt to the enterprise industrial production,which is on the base of the machine vision technology and image detection technology.Taking the characteristics of the actual paging production of the sample page into account,this paper designs the surf-based feature detection algorithm and the OCR-based page number detection algorithm to detect different sample page paging conditions.The feature detection algorithm extracts the surf feature point and the feature descriptor of the template image to match the feature of the target image.Then the matching result is screened by KNN and the mapping relation is established by using the single matrix.The page detection algorithm needs totrain the page characters first,and then use the fast positioning algorithm to calculate the offset angle of the image and intercept the detection area before the detection.What's more,the correlation coefficient template matching algorithm is used to locate the page detection image accurately.At the same time,the page detection image need to be rotated again according to the target image Offset angle.While prepare work's done,the training data is adopt to identify the page image,and the difference matching algorithm is designed to further confirm whether the test results are correct.In order to realize the industrial production of the sample pages,this paper designs and implements a set of automatic detection system,based on the automatic detection algorithm,which adapts to the real industrial detection environment.And successfully applied to the enterprise sample page mechanized paging production.The automatic detection algorithm can solve the problem of automatic detection and meet the real-time detection requirements of industrial production.Through the installation and operation of the actual industrial testing environment,it has good running efficiency,accuracy and robustness,and can adapt to different detection conditions.