基于极坐标计盒维数的圆形通孔缺陷检测

The round hole defect detection based on polar coordinates and box-counting dimension

针对实际生产过程对金属零件圆形通孔的高精度检测要求,在实现硬件系统优化成像质量的基础上,提出了一种基于极坐标变换的圆边缘信息判断方法,并通过计盒维数分类零件通孔圆边缘信息来识别通孔缺陷.该方法具体过程为:首先对工业CCD相机采集到的通孔图像进行预处理得到其轮廓,再通过最小二乘法确定其圆心,并以此为中心将图像转换到极坐标下,然后采用层次聚类和计盒维数的方法分析提取圆边缘特征信息,进而判断通孔缺陷类型.实验证明,该圆边缘信息检测的方法能够有效地检测通孔缺陷,并提高通孔缺陷检测的准确性和效率.

For high accuracy requirement measuring the round hole of metal parts in practical application, this paper proposed a method for round edge information judgement and defect classification based on polar coordinate transformation and box-counting dimension after op-timizing the hardware system, to eventually recognize hole; s defects by examining the edge information of metal parts7 round holes. It got the image contour of the round hole captured by CCD camera,and then confirmed the center of the round hole through least square meth-od. Depending on that,it transformed the image into the condition of polar coordinate to ac-curately judge hole;s defects according to the edge information of round hole by using cluster and box-counting dimension. Experiments have shown that the efficiency and accuracy of the examination are improved through this method.