摘要
提出一种利用高分辨率遥感影像半自动提取建筑物边缘的方法。先对遥感影像进行预处理,对所有边缘进行边缘检测,然后进行边缘跟踪,提取建筑物的主方向线,利用模型进行线段关系判断,再进行线段关系处理、区域分割和区域生长,最后进行区域合并提取出建筑物的轮廓。用上述方法对Quick Bird卫星的高分辨率影像进行了实验,实验结果证明该方法有较高的识别率、较好的准确性,具有一定的实用价值。
A method of high-resolution remote sensing image of buihlings on semi-automatic edge extraction was proposed. Firstly, remote sensing image was preproeessed and all the edge was detected. Moreover, the edge was tracked, linear feature and the main direction line of the buildings was extracted. Then, using the model to judge the relations of line, regional segmentation and growth was realized; Finally, merge regions and extract the building was eontoured. The result showed that this method had a higher recognition level, a better accuracy of a certain practical value using this method to experiment on high-resolution Quick Bird satellite images.
出处
《环境保护与循环经济》
2009年第10期26-28,共3页
environmental protection and circular economy
关键词
航空影像
建筑物轮廓
建筑物识别
边缘检测
aerial images
building contour
building recognition
edge detection
