摘要
通过分析遥感影像中直角型房屋特点,本文提出一种新的房屋提取方法:利用Freeman编码推导出常见直角型房屋代码,使得每一种形状房屋都具有惟一编号,且编号不随房屋大小、方位及位置改变而改变,利用Hough变换提取房屋角点并剔除伪角点,最后利用模式匹配方法获得各个角点之间的连接关系。实验结果表明该方法具有较强的抗噪性,能有效提取影像中直角型房屋。
A semi-automatic extraction of right-angle houses from the high-resolution satellite imagery is proposed. The main process is explained as following, firstly, common-shaped houses are been coded with serial numbers by using of the Freeman chain code, every kind of house corresponds to unique code. Secondly, we extract the house comers through Hough translation and get rid of faked comers, so the targeted house code is been gained. At last, we extract the right-angle houses from remote sensing imagery.
出处
《测绘科学》
CSCD
北大核心
2009年第6期203-205,共3页
Science of Surveying and Mapping
基金
国家自然科学基金(40771159)
