基于蚁群算法的Zernike矩亚像素工件边缘检测

Edge Detection of Zernike Moment SubpixelWorkpiece Based on ACO

边缘检测是影响工件尺寸检测精度的主要因素之一。为了进一步提高工件尺寸测量的准确性,本文提出了一个改进Zernike矩的亚像素边缘检测方法。首先,采用滤波和阈值分解等技术对图像进行预处理以得到二值图,再次,采用蚁群算法得到图像的边界,然后,采用Zernike矩的边界检测方法来确定图像亚像素边缘。结果表明,改进的边缘检测方法相较于传统方法,其提取的亚像素边缘坐标与原始坐标之间的绝对误差不超过0.5,且运行速度提高了约10倍,较好地满足了工件尺寸检测精度的要求。

Edge detection is one of the main factors affecting the accuracy of workpieces’size detection.In order to fur-ther improve the accuracy of the measurement,a sub-pixel edge detection method with improved Zernike moment is proposed in this paper.Firstly,the image is preprocessed by filtering,threshold decomposition and other techniques to obtain the binary image,and then the ant colony algorithm is used to get the image’s boundary,and then the boundary detection method of Zernike moment is used to determine the sub-pixel edge of the image.The results illustrate that,compared with the traditional method,the absolute error between the extracted sub-pixel edge coordinates and the original coordinates is less than 0.5,the operating speed is about 10 times higher,and the improved detection method better meets the requirements of workpiece size detection accuracy.