Development Of Image-based Thread Detection Devices

China is a big manufacturing country,and the market has a high demand on the quality of bolt production.In the production and life,the use of defective bolts can lead to a variety of accidents.In order to reduce the occurrence of such accidents,this paper designs an image-based thread detection device,and uses a reasonable detection algorithm to improve the detection efficiency of defective thread.In this paper,we set up a screw thread detection device,can detect the thread surface defects and helix abnormal situation.The two-dimensional thread image and the thread image with structured light are obtained by the detection device.First,the thread-based detection method is used to detect the thread image,but the method can only detect the surface defect of the thread.Then,the bolt image with structured light is detected by using the thread detection method of monocular stereoscopic vision based on structured light,because the bolt is a high-reflective device,the reconstruction is too difficult,so the method uses the local traverse detection method to detect the defect thread.The 3D reconstruction of the point cloud of the defect thread is carried out by MATLAB,and the reconstructed model of the thread is compared with the standard thread database.Finally,the spiral thread of abnormal is selected.The detection process of the image-based thread detection device designed in this paper takes about 4.5 seconds,through the measurement of the standard block,the accuracy of the three-dimensional detection system is 0.06 mm,the system to meet the industrial automation testing requirements and accuracy requirements,and to overcome the traditional manual detection method is slow,low precision.