机身蒙皮铆接质量检测算法的研究与分析

Research and analysis of riveting quality detection algorithm for fuselage skin

对机身蒙皮的铆接质量进行研究与分析。在自动化生产中,铆接质量的内涵非常丰富,包括几何量与非几何量。通过对基于机器视觉检测部分的铆钉几何量尺寸进行测量,分析铆钉形变端的半径和圆心,首先用标准的Hough变换圆检测算法提取蒙皮图像中的铆接特征检测出的圆定位不准确,像素精度低;然后使用改进的Hough变换精确检测圆的算法对圆形铆钉进行检测,结合最小二乘拟合圆的精确性提高圆心及半径的定位精度;最后实验分别对同一曲面下的相同直径的铆钉图像进行检测分析,从仿真分析和实验结果可知该算法在一定程度上提高了铆接几何量的测量精度。

The article studies and analyzes the riveting quality of the fuselage skin.In the automated production,the quality of riveting is particularly rich,including geometric and non-geometric quantities.By measuring the geometrical dimensions of the rivet based on the machine vision detection part,the radius and center of the deformed end of the rivet are analyzed.Firstly,the riveting feature of the skin image is extracted by the standard Hough transform circle detection algorithm,and the detected circular positioning is inaccurate and the pixel is detected.The precision is low.Then,the improved Hough transform is applied to detect the circle algorithm for the circular rivet accurately where the accuracy of the least squares fit circle is deployed to improve the positioning accuracy of the center and radius.Finally,the experiment is performed on the same diameter under the same surface to analyze the rivet image.From the simulation analysis and experimental results,the algorithm can improve the measurement accuracy of the riveting geometry to some extent.