中文摘要：

将微分几何中的贝特朗曲线性质进行离散化表达,用于道路边缘信息提取,构造的算法能够在有效去除非道路边缘的同时完成对道路边缘侣线的插值,从而使道路边缘更加连续完整。实验和定量化评价结果表明,与经典边缘检测算子相比,此方法具有更高的提取精度。

英文摘要：

Road extraction from high-resolution remote sensing images has very important significance for improving the informatization admministration level of planning, road network maintaining and navigation and traffic analyzing. Bertrand curve property is discretized, and is used to extract the road edge information from high-resolution remote sensing images. The proposed algorithm can effectively eliminate non-road edge, at the same time it can interpolate the companion line of road edge, which makes the road edge more continuous and integrated. The experiments and quantitative assessment results indicate that in comparison with the classic edge detection operators the proposed approach is more efficient.