摘要
以数字表面模型(DSM)影像为辅助,提出了一种结合基于对象和基于知识的建筑物提取方法。首先利用DSM辅助多尺度分割,有助于分割完整的建筑区域;然后利用3种规则进行建筑物提取;最后利用国际摄影测量与遥感学会(ISPRS)标准数据集影像进行实验验证。结果表明,该方法可有效提取建筑物,得到高精度的提取结果,错提率与漏提率分别为6.26%与7.43%。
With assistance of digital surface model(DSM), we proposed a building extraction method by using the combination method of object-based and knowledge-based in this paper. Firstly, we segmented images by multi-resolution segmentation method with the help of DSM to get complete building area. Then, we set three rules to extract buildings, and used the experimental images of ISPRS datasets to verify the method. The results show that the proposed method can effectively extract buildings with a high accuracy, and the commission rate and omission rate are 6.26% and 7.43%.
出处
《地理空间信息》
2019年第1期38-40,124,共4页
Geospatial Information
基金
中央高校基本科研业务费专项基金资助项目(2042016kf0002)
关键词
DSM
建筑物提取
多尺度分割
DSM
building extraction
multi-resolution segmentation
