基于知识库的遥感图像桥梁识别研究  

Knowledge-based Bridge Recognition Method in Remote Sensing Image

作　　者：张永梅[1] 杨晓易[1] 杜晓伟[1] 

机构地区：[1]北方工业大学信息工程学院,北京100144

出　　处：《微计算机信息》2011年第7期217-219,共3页Control & Automation

基　　金：基金申请人:张永梅;项目名称:基于视觉XXX与多元特征的XXXX技术研究(保密);基金资助项目(作者申报的项目)名称:国家863计划;基金颁发部门:总装备部(2007AA0879)

摘　　要：利用桥梁的特征,实现高分辨率遥感图像中桥梁目标的识别方法。首先建立桥梁特征的知识库,在低层处理中运用链码边界跟踪与边界填充的方法提取出水域;在中层处理中使用Hough变换寻找桥梁线并提取其特征参数;最后在高层处理中对目标进行验证,并在整个过程中加入知识进行指导。该识别软件具有较高的识别准确率。Based on the information and features of bridge, a recognition algorithm is realized in this paper. First the knowledge base of bridge and its scene is built, then in the low -level processing, extract the water contour by fast tracing boundary contours approach and contour filling algorithm based on chain codes description. In the middle-level processing Hough transform is used to locate the targets and get the parameter of bridge. In the high -level processing, targets are verified, and each step is guided by its corresponding knowledge to enhance. The software has higher bridge recognition.

关 键 词：知识库 桥梁识别 链码 HOUGH变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]