一种SPOT图像公路信息全自动提取新方法  

An Automatic Road Extraction New Method from SPOT Image

作　　者：谭左平[1] 李朝峰[1] 魏敏[1] 

机构地区：[1]江苏无锡江南大学信息工程学院,214122

出　　处：《微计算机信息》2006年第10X期257-259,97,共4页Control & Automation

基　　金：国家自然科学基金仪器专项(No.50427401)资助

摘　　要：提出了一种基于模糊连接度的SPOT图像公路信息全自动提取算法。该方法首先运用Sobel算子对SPOT图像进行边缘检测,然后使用阈值优化道路边缘点从中选取出道路种子点,接着通过计算图像中各点与种子点间模糊相似程度来提取道路信息,最后进行进一步完善处理。实验表明,该方法性能良好,在整个提取过程中不需要人工输入种子点,可以快速地实现道路特征的全自动提取。A fuzzy connectedness-based road automatic extraction method from SPOT image is presented. The method firstly uses Sobel to compute the edge point of road, and secondly chooses road seed points by threshold method, and then extracts road information by computing the fuzzy connectedness between the seeds and pixcls in SPOT image, and finally makes further processes to achieve whole road net. Experimental results show that the method may receive good performances, and moreover road seeds need not choose by manual, and may fast perform road automatic extraction.

关 键 词：SPOT图像 道路提取 道路种子点 自适应 模糊连接度 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]