提出一种改进的医学图像边缘检测方法。利用图像四叉树分解技术和模糊形态学的思想,根据图像像素点方向的不同性,构造不同方向不同尺度的结构元素对边缘像素进行模糊形态学运算,进而检测到图像边缘。实验结果表明,算法提取的边缘能够精确地显示物体的内部细节信息,较好地解决了抑制噪声和保留精细边缘之间的矛盾,与其它边缘检测算子相比是一种较实用的边缘检测算法。
An improved method of medical image edge detection is introduced in the paper. It uses image quadtree decomposition technology and the idea of fuzzy morphology, constructs structural elements with different directions and scales according to the difference of image pixels' directions to perform fuzzy morphology operation against edge pixels and then to detect image edge. The experiment demonstrates that the extracted edge can accurately display image's internal detail information. It can solve the contradiction preferably between anti-noise and thin edge preserving and is a practical algorithm for edge detection comparing with other edge detection operators.