Xinxin Fang / Liaoning Technical University;School of Geomatics
Weidong Song / Liaoning Technical University;School of Geomatics
激光 戴 / 辽宁工程技术大学;测绘与地理科学学院
Accurate rural road vector data are the basis of regional planning and economic development. Existing methods use remote sensing images to correct existing vector data. The internal accuracy of existing vector data and the difference in the internal characteristics of road images are not considered. This leads to incompatibility between accuracy and automation. In response to this problem, this paper proposes a rectifying method for rural vector data based on characteristic differences. A model-driven method for areas with obvious geometric and textural road features for which images with high vector-direction accuracy are available is applied; using this method, the vector data for road sections are automatically rectified. Next, in areas where road features are not sufficiently obvious or the vector direction accuracy is poor, an interactive human–machine data-driven method is proposed to rectify the remaining rural vector data. The experimental results show that when dealing with tasks involving multiple scenes and multiple data types, the degree of automation and the efficiency of the proposed algorithm are superior to those of existing methods.