摘要
针对尿沉渣图像的背景和目标区分度低、有形成分复杂,导致提取困难的问题,设计了一种融合形态学和Snake模型的尿沉渣提取算法。采用基于区域生长的阈值分割方法对尿沉渣进行粗定位,通过形态学的方法确定Snake模型的初始蛇,并将C anny的边缘定位法融入到Snake模型的贪婪算法中,最后完成对尿沉渣的准确提取。实验表明融合算法性能稳定,所提取的尿沉渣区域定位准确;特别是解决了虚假边缘和边缘断裂现象较为严重的尿沉渣轮廓提取问题。
In order to solve the problem that urine sediment can＇t be extracted effectively because of poor discrimination between object and background and complex components in image, a method based on morphology and active contour model was designed for extracting urine sediment. First,the approximate location of urine sediment was gained by a method based on threshold segmentation and region growing. Second, a primary snake nearby the desired sediment contour was established with an algorithm based on morphology. Last, the primary snake was pushed toward the appropriate position step by step by fusing the Canny algorithm of edge-location into the greedy algorithm of the Snake-model. Experiment results show that this method has super contour localization accuracy and stable performance. It especially solves the contour extraction of urine sediment with many fake edges and break edges.
出处
《桂林电子科技大学学报》
2007年第6期482-486,共5页
Journal of Guilin University of Electronic Technology
基金
广西教育厅科研项目(200508107)
广西师范大学青年基金(师政科技[2006]5号)
