基于双层案例推理模型的高分辨率遥感影像道路提取方法

Road information extraction method of high resolution remote sensing image based on bi-level case reasoning model

为实现高分辨率遥感影像道路信息的自动提取,引入案例推理思想,提取道路影像的光谱、形状、纹理和影像间拓扑关系特征信息,构建多层案例库,在传统案例推理模型的基础上进行改进,提出多层案例推理模型.结合影像预处理、权重分配和分层检索,实现多层案例推理模型的高分影像道路信息提取.通过道路信息提取、案例自学习、案例适用试验及与SVM方法对比分析,证明该方法基本实现道路信息提取的目标.

In order to realize the automatic extraction of road information for high resolution remote sensing images, the case-based reasoning theory was introduced, and the spectrum, shape, texture and topological relationship information of high-resolution remote sensing images were extracted to construct multi-layer case base and improve it based on traditional ease reasoning model, then the multi-layer case-based reasoning model was proposed. The high level image road information extraction from muhi-layer case-based reasoning model was implemented by combining with image preprocessing, weight allocation and layered retrieval. Experimental re- suits show that the proposed method basically realize the goal of road information by comparing with SVM method, road in- formation extraction, case self-learning and case testing.