摘要
为了解决传统边缘提取算法边缘定位不精确、抗噪能力差的问题,提出了一种基于数学形态学的边缘提取算法。该算法首先利用了双尺度双结构的数学形态学对目标图像进行滤波降噪处理,以提高目标图像的信噪比,然后利用多尺度多结构的数学形态学对目标图像进行边缘提取。利用该算法在配置了OpenCV的Visual Studio对Lena图像进行仿真处理,并将其处理结果与Canny算法处理结果进行对比。实验结果表明,该算法抗噪性能优异,对含有噪声的图像边缘的提取清晰且流畅、细节丰富。
In order to solve the traditional edge detection algorithms with poor noise immunity characteristics and inaccurate edge location,a new edge detection method based on mathematical morphology was put forward.Firstly,the mathematical morphology with double scale double kernel element was applied to image noise reduction to improve the signal to noise ratio of the image,then the mathematical morphology of multi-scale multi-structural elements was used for the image edge extraction.The algorithm was used in Visual Studio with Open CV for Lena image processing simulation.Experimental results show that the new algorithm has good anti-noise performance compared with the results by Canny algorithm.
出处
《机械工程与自动化》
2015年第1期46-48,共3页
Mechanical Engineering & Automation
基金
上海市研究生创新基金资助项目(JWCXSL1202)
关键词
数学形态学
边缘提取
滤波降噪
图像处理
抗噪处理
mathematical morphology
edge detection
filter noise
image processing
anti-noise processing
