Research And Application Of Part Size Detection Algorithm Based On Machine Vision

Workpiece has a wide range of applications in daily life and industrial manufacturing, workpiece sizes is an important criterion to judge the work eligibility. The traditional detection method is manual detection, this detection method prone to eye fatigue, resulting in false detection and missed, and the detection efficiency is relatively low. Under this background, the technology of measuring the dimension of parts based on machine vision, as a non-contact detection technology with high sensitivity, short time-consuming, high-precision advantages, has received extensive attention and development.In this paper, with a small metal workpiece as research object, to carry out the research on metal surface defect detection algorithm based on machine vision, the main work is as follows:1. A kind of ellipse detection algorithm based on maximum inscribed circle is proposed. The traditional ellipse detection algorithm is easy to be disturbed by the background noise of the image, and the detection efficiency is low. Using elliptic and related circular properties, first with a Gaussian filter and morphological processing to filtering and de-noising on elliptical hole group, using Canny edge detection operator for edge detection to detect the ellipse hole group and get the coordinate of the hole group center, long and short axial length and angle of inclination.2. Aiming at the dimension measurement of the abnormity sheet parts, a new method for the detection of thin sheet parts with contour feature points extraction based on image structure is proposed. Combining with the definition and properties of concave and convex functions, the feature points of the part contour were extracted and the contour line was cut apart by calculating the difference of distance between arc midpoint and chord midpoint and the difference of value between left and right derivative during the contour point and the points of neighborhood window respectively. The feature point detection technology can locate and segment the parts contour, form different curve segments, and then estimate the parameters of the curve segment. Due to the algorithm adopts the idea of subsection analysis part outline, there is no specific requirement for the shape of the parts to be measured, with a certain degree of adaptability.3. Combine with ellipse detection algorithm based on maximum inscribed circle and the detection method of thin sheet parts with contour feature points extraction based on image structure, using MATLAB language to develop a part size detection interface. The software is divided into image preprocessing module, ellipse hole group parts detection module and abnormity sheet detection module, provides a brief description for each module and demonstrates the operation for the experimental parts.