基于高分辨率遥感影像数据的道路网络提取算法  

Road Network Extraction Algorithm based on High- Resolution Remote Sensing Image Data

作　　者：彭雅琪 吴国昊 PENG Yaqi;WU Guohao(Zhejiang Institute of Surveying and Mapping Science and Technology,Hangzhou,Zhejiang 310030,China)

机构地区：[1]浙江省测绘科学技术研究院,浙江杭州310030

出　　处：《测绘标准化》2022年第4期36-40,共5页Standardization of Surveying and Mapping

摘　　要：为了能从高分辨率遥感影像数据中有效提取道路中心线,满足地理信息库建设等方面的数据需求,本文提出了基于高分辨率遥感影像数据的多特征融合道路中心线提取算法。首先,基于影像分割提取道路的空间特征与光谱特征;其次,通过多特征融合算法融合空间特征与光谱特征,通过对面向对象角度构建的形状特征进行网络优化,获取精细化道路网络;最后,引入计算机视觉中的张量投票算法准确提取道路中心线。将本文算法与已有道路网络提取算法进行对比,结果表明,本文算法提取结果在检测质量、正确率和完整率3个评价指标方面均优于已有算法,验证了其有效性与优越性。In order to effectively extract road centerline from high-resolution remote sensing image data and meet the data requirements of geographic information database construction, this paper proposes a multi-feature fusion road centerline extraction algorithm based on high-resolution remote sensing image data. Firstly, spatial features and spectral features of roads are extracted based on image segmentation;Secondly, spatial features and spectral features are fused by multi-feature fusion algorithm, and refined road network is obtained through network optimization based on shape features constructed from the perspective of object-oriented;Finally, the tensor voting algorithm in computer vision is introduced to accurately extract the road centerline. The proposed algorithm is tested and compored with the existing road network extraction algorithms, the experimental results show that the extraction results of the algorithm in this paper are better than the existing algorithms in terms of detection quality, accuracy rate, and integrity rate, which verifies its effectiveness and superiority.

关 键 词：道路提取 道路网络 多特征融合 张量投票 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置] U495[自动化与计算机技术—控制科学与工程]