为满足基于机器视觉的刀具尺寸测量系统快速及高精度的要求，提出一种基于直线截距直方图的 Arimoto 熵和 Zernike 矩的图像亚像素边缘检测方法．首先，通过高斯 滑动窗口获取图像的邻域平均灰度，构造图像的灰度－邻域平均灰度二维直方图，并利用直线截距法将其降为一维直方图; 然后，针对得到的直线截距直方图，依据 Arimoto 熵准则进行阈值分割，并将所得阈值映射回原二维直方图实现目标区域及像素级边缘的提取; 最后，由基于Zernike 矩的边缘模型对获取的像素级边缘进行重定位，以完成刀具图像亚 像素级边缘的提取．通过对刀具图像进行的大量实验，将文中方法与基于Canny 的、基于空间矩的、基于灰度矩的以及基于 Zernike 矩的边缘提取方法进行对比，发现文中方法运行速度更快且提取精度更高．

In order to meet the high-speed and high-accuracy demands of the machine vision-based measurement system of cutting tool sizes,an image sub-pixel edge detection method based on Arimoto entropy of linear intercept histograms and Zernike moment is proposed.In the method,first,the neighborhood-average grayscale of images is obtained through the Gaussian sliding window to construct a two-dimensional histogram,and the linear intercept method is adopted to reduce the two-dimensional histogram to a one-dimensional histogram.Then,aiming at the a- chieved linear intercept histogram,the thresholding is performed according to the Arimoto entropy,and the ob- tained threshold is mapped back to the two-dimensional histogram to extract a target region and pixel-level edges.Finally,the edge points are re-located by using the Zernike moment-based edge model,thus achieving the sub-pix- el-level edges of cutting tool images.By a large number of experiments on the cutting tool images,the proposed method is compared with the Canny-based,the space moment-based,the gray moment-based and the Zernike mo- ment-based edge extraction methods.The results show that the proposed method is superior in both speed and accu- racy.