Industrial Parts Measurement Based On Machine Vision Technology

Owing to the fact that machine vision technology has a huge advantage in the field of industrial detection, the research and application of vision technology in this area has appeared more and more paper in recent years. This paper starts from a common industrial part called rivets and analyzes the application of image processing and machine vision technology in this field. Traditional rivets detection based on artificial way has extremely low efficiency. This way would spend more efforts for the detection of large quantities of rivets and would not meet the requirements of modern industrial online measurement. In order to overcome defects of the traditional artificial rivets detection, this paper proposes a non-contact automatic detection algorithm of multi-rivets based on machine vision and reveals the basic structure of visual technology in the field of industrial parts measuring.The framework of this paper is divided into four parts:(1) Camera calibration. This part starts from four coordinate systems and introduces in detail the camera imaging model. And then a camera calibration method is given in this system via the analysis of an algorithm based on radial alignment constraint.(2) Image acquisition and pre-processing. This part, firstly, describes the works of CCD and CMOS image sensors and the importance of light source. Secondly, this part analyzes in detail the source of the noise and handles noisy image using two classic filtering algorithms. In the end of this part we propose an improved median filter algorithm. This algorithm adds the judgment for signal points and noise points making, it can effectively improves the effect.(3) Rivets contour extraction. Image segmentation technology status and its research hierarchy described in this section. This part analyses the problem of edge contour and introduces several segmentation algorithms based on differential and regional. Finally, we propose an improved region-based OTSU segmentation method. The experimental results show that this method can effectively reduce the false division.(4) Rivets parameters strike. Judging the rivet is qualified to detection need to get the full-length, short diameter, length diameter and riveting length parameter. We get parameters such as full-length and short diameter in last chapters. In this section, we identify the outline feature points by calculating curvature of each rivet contour point in the both sides of the support region. When we know the position of feature points, length diameter and riveting length parameter can be easily obtained.