基于图像灰度梯度的圆度检测算法

Aroundness Evaluation Algorithm Based on the Gray Gradient of Image

在产品实时检测系统图像采集的过程中,难免采集到一些噪声。提取灰度图像目标边缘进行圆度检测时,就需要进行滤波处理,这样会造成被检测边缘的部分信息丢失,检测精度降低。为减小圆度检测误差,保证产品质量,对图像灰度梯度研究时,提出先将灰度梯度的幅值进行非极大值抑制,然后用理想阈值分割图像,再应用最小二乘圆对结果进行圆度检测的算法。最后,通过激光光斑圆度检测实验,对激光光斑的圆度误差进行评定,证实了算法的优越性。

In the process of detection roundness of detection system iage colletion, noise can＇ t be avoided. Be- cause of filtering processing, some information of the image edge will be loss and the detection precision will be re- duced. There is a algorithm is presented which proceed non-maximum suppression of image＇ s gray gradient magnitude and find the ideal threshold to split image. Then applicate the least square circle to detect the roundness. Finally we assess the roundness error of laser spot by laser roundness detection experiment, confirm the superiority of the proposed algorithm.