基于机器视觉的羽毛球圆度测量系统研究

Badminton Roundness Measurement System Based on Machine Vision

针对羽毛球圆度测量困难,效率低等问题,搭建了一套基于机器视觉的羽毛球圆度测量系统。首先,利用滤波降噪,阈值分割和孔洞填充方法对图像进行预处理;其次,通过Freeman链码轮廓跟踪和改进FAST算法完成羽毛球候选特征点的筛选,针对候选特征点存在的3个问题,分析羽毛球几何特征,计算候选特征点的轮廓尖锐度和支撑区域凹凸性,并利用局部非极大值抑制进行优化,得到最终的特征点;最后,以最小二乘法对特征点进行拟合,实现羽毛球的圆度测量。实验结果表明,系统测量半径平均误差为0.21 mm,半径最大误差为0.34 mm,能准确分类合格羽毛球和次品,测量平均速度为457 ms。因此,该系统能较好地完成羽毛球自动化生产中的测量任务。

Addressing the challenges and inefficiencies associated with measuring badminton shuttlecock roundness,a machine vision-based measurement system was developed.Initially,the system preprocesses the image using filtering,threshold segmentation,and hole filling techniques.Subsequently,candidate feature points of the shuttlecock are filtered using the Freeman chain code for contour tracking,combined with an enhanced FAST algorithm.Given the inherent issues with these candidate points,the system calculates their contour sharpness and the convex-concave nature of the supporting area.This is further optimized using local non-maximum suppression to determine the final feature points.These points are then fitted using the least square method to measure the shuttlecock′s roundness.Experimental results show an average radius measurement error of 0.21 mm,with a maximum error of 0.34 mm.The system can effectively differentiate between acceptable shuttlecocks and defects,boasting an average measurement speed is 457 ms.Consequently,this system proves highly effective for automated badminton production measurement task.