摘要
针对目前基于机器视觉技术的纸杯杯口圆度质量检测时,因光照不均而致杯口边缘轮廓定位准确率低和速度慢等问题,采用对图像径向灰度及梯度的特征分析,提出一种自适应阈值的双向径向扫描方法。应用双边滤波优化图像径向灰度分布规律,以及对灰度直方图均衡算法HE算法进行改进,使杯口边缘灰度对比更加明显,减弱了光照对获取边缘点的影响,提高了边缘点定位的准确率和效率。另外针对杯口圆度质量评估参数单一而导致漏检测问题,在获取杯口内外边缘点坐标之后,应用最小二乘法分别进行圆拟合,并以内外两个轮廓圆度来综合评定杯口圆度质量,提高了检测准确率。实验结果表明,所提算法对不同圆度缺陷的纸杯具有98%的检测准确率和0.123秒/个的平均检测速率,相较于其他算法更加快速有效。
Aiming at the problems of low positioning accuracy and slow speed of edge positioning due to uneven illumination in the roundness inspection for cups mouths by machine vision technology,based on analyzing the radial grayscale and gradient features,a bidirectional radial scanning method with adaptive threshold was proposed.Bilateral filtering was used to smooth the image and optimizing the radial grayscale distribution of the image.The HE algorithm was further improved to make the cup edge’s grayscale contrast more pronounced,reducing the impact of lighting on edge point detection,which improved the accuracy and efficiency of edge point positioning.To address the issue of missed detections due to the reliance on a single parameter for assessing cup mouth roundness,after obtaining the coordinates of the inner and outer edge points of the cup mouth,the least-squares method was used to fit the coordinates of the inner and outer edge points obtained by the bidirectional scanning method respectively,and the roundness quality was comprehensively evaluated by both the inner and outer contours’roundness,which improves the detection accuracy.The experimental results show that the method has a 98%detection accuracy and a 0.123 s/piece average detection rate for cups with different roundness defects,and more effective compared to other methods.
作者
黄晏哲
罗亚波
HUANG Yanzhe;LUO Yabo(School of Mechanical and Electronic Engineering,Wuhan University of Technology,Wuhan 430070,China)
关键词
圆度测量
机器视觉
径向扫描
双边滤波
图像增强
roundness measurement
machine vision
radial scan
bilateral filtering
image enhancement
作者简介
黄晏哲(1996-),男,湖北潜江人,武汉理工大学机电工程学院硕士研究生.
