为了提高传统接触式测量方法的测量效率并避免人为误差的影响，同时克服机器视觉中常用的Hough 变换圆检测算法计算量大以及不适用于同心圆检测的情况，提出一种新的视觉测量方法用于实现环形目标直径测量。该方法通过阈值分割与边缘检测得到物体的轮廓图像，进行轮廓Radon 变换，得到包含四条光带的处理结果并反映了所有方向角上环形目标直径的测量数据，利用灰度重心法提取光带中心线，进而根据中心线之间的距离计算出环形目标直径。仿真与实验表明，所提方法测量结果科学可靠，拥有较高的测量精度，能够实现同心与偏心目标直径的自动非接触测量，将具有很好的工程应用价值。

To improve the measuring efficiency and avoid the influence from artificial error in the traditional contact method, and to overcome some limits presented in the Hough transform circle detection algorithm that often uses to machine vision, such as heavy calculation amount and inapplicable to the concentric circles detection, a new vision measurement method to measure the ring object diameter is proposed. The proposed method obtains the contour image of object through threshold segmentation and edge detection.Then the Radon transform to the contour image is executed, and the pretreatment result that includes in four light stripes reflects the measuring data of the ring diameter in the all direction angle. Finally, the central lines are extracted from the light strips with the gray gravity algorithm, furthermore the ring diameters are calculated according the distances between central lines. The simulation and experiment indicate that the proposed method holds scientific and reliable measurement result with higher measuring accuracy. Also, it can realize the automatic and non-contact measurement of concentric and eccentric object diameter and hence will hold good engineering application prospect.