Luo Xiaogang, Liu Ting, Peng Chenglin, Wen Li
College of Bioengineering, Chongqing University, Chongqing 400044, China.
Sheng Wu Yi Xue Gong Cheng Xue Za Zhi. 2009 Feb;26(1):177-80, 183.
An improved edge detection algorithm is proposed in this paper for the medical images with strong noises and fuzzy edges. The algorithm modified the combination of morphological operations, so that the unclear edges of the images are avoided. In this paper is also introduced the algorithm of multi-structure elements which can reserve integrated edges from different directions of the images. Furthermore, the contrast enhancement and morphological filter processing are implemented. This method can detect the edges efficiently, keep the detected edges smooth and obtain coherent image edges. Experiments demonstrate that this edge detector has a better performance of noise reduction and keeps the edges more accurate than do the traditional edge detectors; thus its practicality is enhanced.
本文针对具有强噪声和模糊边缘的医学图像提出了一种改进的边缘检测算法。该算法对形态学运算的组合进行了改进，从而避免了图像边缘不清晰的问题。本文还介绍了多结构元素算法，该算法可以保留图像不同方向的完整边缘。此外，还进行了对比度增强和形态学滤波处理。该方法能够高效地检测边缘，使检测到的边缘保持平滑，并获得连贯的图像边缘。实验表明，该边缘检测器具有比传统边缘检测器更好的降噪性能，并且能够更准确地保留边缘；因此增强了其实用性。
