基于改进Canny算子的坯料挤压变形边缘提取  

Edge detection of extruded billet based on improved Canny operator

作　　者：潘晟旻[1,2] 钟毅[1] 王建华[2] 

机构地区：[1]昆明理工大学材料科学与工程学院,云南昆明650093 [2]昆明理工大学计算中心,云南昆明650500

出　　处：《山东大学学报（工学版）》2013年第5期19-23,共5页Journal of Shandong University（Engineering Science）

基　　金：国家科技支撑计划资助项目(2011BAE13B06);国家自然科学基金资助项目(50870455)

摘　　要：为了研究塑性加工过程中金属坯料的流动行为,提出了一种边缘检测算法。先运用灰度增强与形态学膨胀、腐蚀相结合进行图像预处理,再通过双阀值约束的Canny算子进行坯料变形轮廓提取,避免了孤立灰度阶跃点对边缘判断的干扰。在等量椒盐噪声下,本算法的峰值信噪比高于其他典型算法。实验结果表明,该算法所提取的变形区边缘能够满足量化分析金属流动特征的需求。To deal with flow behavior of metal billet in plastic deforming process, an algorithm of edge detection was proposed. Firstly, the sample image was preprocessed by used the method of gray-scale enhance combine with morpho logical corrosion and expansion, and then the billet deformation contour was extracted by Canny operator that is limited by double thresholds. It avoided the isolating step grayscale point on the edge. In the same amount of salt and pepper noise, the peak signal to noise ratio of this algorithm was higher than other typical algorithms. The experimental results showed that the edge of deformation zone that it be extracted by this algorithm, can meet the quantitative analysis of metal flowing characteristics.

关 键 词：图像处理 CANNY算子 边缘检测 图像增强 信噪比 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]