基于Canny算子改进的图像边缘检测算法研究  

Study on the Improved Image Edge Detection Algorithm Based on Canny Algorithm

作　　者：郭忠峰[1] 唐晓晓[1] 任仲伟[1] 刘娅菲 匙皓 

机构地区：[1]沈阳工业大学机械工程学院,辽宁沈阳110870 [2]贵州红林机械有限公司,贵州贵阳550009

出　　处：《机械研究与应用》2017年第2期123-125,共3页Mechanical Research & Application

摘　　要：针对Canny算法的图像边缘检测技术,提出用中值滤波和均值滤波合成的混合滤波器代替传统高斯滤波器,加强了对复杂噪声的平滑降噪;在边缘连接细化中,采用基于Ostu提出的最大类间方差法的自适应阈值选择法,提高了效率和适应性。并进行MATLAB仿真,结果表明提出的改进算法明显提高了图像边缘检测的精度。In view of image edge detection technology of the Canny algorithm, the hybrid filter with the median filter and meanfilter instead of the traditional Gaussian filter is proposed in this paper. The smooth reduction of complex noise is enhanced.The adaptive threshold selection method of the maximum classes square error based on Ostu is applied during the refine edgeconnection, thus the efficiency and the adaptation are improved. The results through MATLAB experiments show that the pre-cision of image edge detection is greatly increased by the improved algorithm in this article.

关 键 词：CANNY算法 中值滤波 均值滤波 图像边缘检测 边缘连接 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]