种子区域生长法的改进算法及其在钉螺图像提取中的应用  

Research on the improved algorithm and application of seed region growth method and its sail image extration

作　　者：严深海[1] 黄贤通[1] 刘洋[1] 

机构地区：[1]赣南师范学院数学与计算机科学学院,江西赣州341000

出　　处：《韶关学院学报》2011年第10期21-24,共4页Journal of Shaoguan University

摘　　要：种子区域生长法是图像分割中一种常用的方法,其性能在很大程度上取决于种子点的选取与生长规则.为了更有效地应用种子区域生长法从真实图片中提取钉螺图像,在分析种子区域生长法的基本思想后,提出了一种改进的种子区域生长算法并用于提取钉螺图像.改进种子区域生长法使用了新的区域生长相似性规则和区域生长停止的动态阈值方法,改进算法采用人工方式来选择区域生长的种子点.结果表明,改进的种子区域生长法在复杂背景图片中能获得更好的钉螺图像提取效果.Seeded region growing is a common method of image segmentation.Its performance depends largely on the selection of seed points and growth rules.In order to extract snail images with such methods more effectively.The article analyzes the basic idea of the seeded region growing method（SRG） and presents an improved seeded region growing method（ISRG） which is used to extract the snail images.In ISRG,novel similarities rules and a new dynamic threshold method are adopted.For the complex and various habitat of snail,ISRG selects the region growing seed points manually.It is demonstrated by experimental results that ISRG can achieve better snail image extraction results under the complicated real-world scene.

关 键 词：种子区域生长 图像提取 动态阈值 钉螺 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]