The Triangular Thread3D Detection And Research Based On Machine Vision

Modern industrial has posed an increasing requirement on thread detection, traditionalapproach of external thread detection is inefficient, which cannot fulfill the detection needs ofmodern production, however, machine vision inspection technology has the advantages of fastdetection speed, high precision, non-contact, advanced automated degree. Currently, theapplication of machine vision technology to detect the thread is based on the detection ofplanar contours or regions, which does not reflect the three-dimensional information of thethread. For meeting the demands of high precision thread detection, this subject raises amethod of the triangular external thread three-dimensional detection based on machine vision.This subject investigates a3D model reconstruction method of the triangular externalthread. According to the triangular profile characteristics of the external thread, this paperdesigns a reasonable experimental method and3D reconstruction algorithm, which employsvisual C++programming language to achieve the results. Firstly, gather the images of theexternal thread with the CCD camera. Secondly, preprocess the external thread images,including image enhancement, thread edge detection and extraction of the threaded featurepoints, Lastly, use the extracted thread feature points to reconstruct the external threadthree-dimensional model. The external thread3D models generated by the experiment achievepreferable results, which can truly reflect the threaded parts of the surface contour shape. Thismethod is a feasible approach for the external thread3D reconstruction.A3D detection algorithm for triangular external thread is studied in this paper, whichconduct the experiment with MATLAB programming language. The first step is proposing amatching algorithm between the reconstructed3D thread model and the standard threadmodel. After that, use a function to quantitatively assess the similarity between thereconstructed3D thread model and the standard thread model. Lastly, charge certainparameters of the standard thread to analyze the rationality of the matching similarityalgorithm.3D detection experiments are conducted with the car wheel bolts and hex headbolts, results indicate that the quantitative description of the similarity has high rationality.