B超图像边缘检测算法的研究与比较

Research and Comparison of Edge Detection Algorithms for B-Scan Images

边缘检测在B超图像处理中发挥着重要作用。本文介绍了五种边缘检测算法,选取了两幅不同特征的B超图像,胆囊B超图像和子宫B超图像,使用MATLAB编程工具,分别对这两副图像进行边缘检测,比较并分析各算法的检测效果。结果表明,Roberts算子、Prewitt算子、Sobel算子对于边缘清晰且与周围组织对比度明显的胆囊B超图像处理效果较好。使用复杂的Canny算法反而出现伪边缘,处理效果不好。而Canny算法对于对比度不明显的子宫B超图像处理效果要明显优于其他算子。

Edge detection is an important processing method in B-scan image processing.In this paper,five algorithms of edge detection are introduced,and chose two B-scan images with different characteristics,B-scan image of gallbladder and B-scan image of uterus,detect the edge of these two images by the five algorithms in MTALAB,compare and study the detection results and running time.The result shows,Roberts operator,Prewitt operator,Sobel operator for edge clear and obvious contrast with the surrounding tissue of B-scan image processing,the effect is better than Canny operator.But for less contrast B-scan image processing,Canny operator is better than other operators.