Research On Dimension Detection Algorithm Of Threaded Steel Pipe Based On Machine Vision

Most of the traditional thread detection methods are contact measurement,which is time-consuming and laborious and easily damages the thread.With the increase of the requirements for the speed and accuracy of thread inspection in the industry,the thread inspection technology based on machine vision has developed rapidly.Machine vision-based thread detection technology has the advantages of fast detection speed,high accuracy,and non-contact measurement method will not cause damage to the thread.This paper designs a visual inspection system for hollow threaded steel pipes,focuses on the related algorithms of image processing in the inspection system,and uses this method to achieve high-precision detection of the external thread size of steel pipe.The thesis mainly completes the following tasks:1.Starting from the actual project and testing needs,a hardware platform for detecting the size of a steel pipe composed of an industrial camera,a telecentric lens,and a light source is established;2.Through the graying processing,filtering,edge detection,sub-pixel corner detection and polygon approximation of the external thread image,the exact coordinate values of the thread edge feature points are obtained.According to the actual detection requirements,the bilateral filtering algorithm based on the local filtering template is researched,and the Canny edge detection algorithm based on the maximum variance between classes is used to improve the speed and accuracy of the image graphics processing system;3.The camera mapping coefficient of the size detection system and the center distance of the two cameras are calibrated.The mathematical calculation principles for the external thread pitch,tooth angle,large diameter and middle diameter in the size detection system are explained,and the visual inspection is calculated separately.The thread size value detected by the system and error analysis was performed.The test results show that the visual inspection algorithm can make the external thread size detection accuracy of the steel pipe reach 0.01 mm,and the false detection rate and missed detection rate can meet the actual industrial inspection requirements.