基于阈值分割和轮廓提取的图像边缘检测算法  

Medical Cell Edge Detection Algorithm Based on Threshold Segmentation and Contour Extraction

作　　者：徐国雄[1] 张骁[1] 胡进贤 倪旭祥 

机构地区：[1]安徽工业大学,安徽马鞍山243002 [2]合肥赛为智能有限公司,安徽合肥230022 [3]现代光学仪器国家重点实验室,浙江杭州310027

出　　处：《计算机技术与发展》2015年第12期64-67,71,共5页Computer Technology and Development

基　　金：安徽省自然基金重点项目(kj2009a132)

摘　　要：文中提出了一种全新的基于阈值分割和轮廓提取的医学图像边缘检测算法,并在MATLAB平台上进行了模拟仿真。与传统的算法相比,该算法具有三方面的优点:数学形态学重建去除了图像中的阴影和反射光噪声;通过迭代算法计算出最佳阈值并对图像进行分割,有效地降低了噪声对最终检测结果的影响;利用形态学腐蚀算法提取细胞边缘轮廓,使细胞边缘的宽度仅有一个像素,提高了边缘定位的精度,避免了细胞边缘的重叠现象。同时,检测后得到的细胞边缘具有良好的连续性。A novel medical cell edge detection algorithm is presented in this paper based on threshold segmentation and contour extraction and tested on the platform MATLAB. Compared with the traditional classical edge detection algorithms, the result shows the novel algo- rithm has following advantages：Morphological reconstruction removes image noise shadows and reflected light. The image is segmented clearly through an iterative algorithm to calculate the optimal threshold value. The width of the cell edge is only one pixel to improve the positioning accuracy of the edge, to avoid overlapping cell edge, as well as the cell edge has good continuity.

关 键 词：医学图像 边缘检测 数学形态学 阈值分割 轮廓提取 仿真 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]