基于改进滤波的Canny医学图像边缘检测算法

Canny Medical Image Edge Detection Algorithm Based on Improved Filter

医学图像的边缘细节是分析病情与临床诊断的关键,针对传统Canny算法存在易丢失图像边缘细节的问题,提出了一种改进的Canny边缘检测算法,该算法采用基于形态学算子的各向异性扩散滤波代替高斯滤波平滑图像,同时引入Otsu法自动获得Canny算子高低门限;实验结果表明,与传统Canny算法以及其他改进Canny算法相比,新算法较好地保留了医学图像边缘细节信息,并且很好地抑制了虚假边缘,另外高低门限的自动获得,提高了边缘检测的自适应能力,为后续医学分析诊断提供关键支持。

The edge of the details of the medical image is key for the analysis of disease and clinical diagnosis, for the traditional Canny al- gorithm is easy to loss the image edge details , the improved Canny edge detection algorithm is put forward, the algorithm based on morpho- logical operators anisotropic diffusion filtering instead of the Gaussian filter smoothed image, while the introduction of the Otsu method to au- tomatically get the threshold of Canny operator. Experimental results show that, compared with the traditional Canny algorithm and other improvements, the new algorithm is better retain the details of the medical image edge, and inhibition of the false edge. In addition the high and low threshold automatically obtained , improving the edge detection adaptive ability to provide critical support for the follow--up medical analysis and diagnosis.