景象匹配导航技术以其自主性强,定位精度高,抗电子干扰的优势逐渐成为未来无人机平台最主要辅助导航方式之一。采用基于边缘响应加权Hausdorff距离景象匹配算法作为仿真系统的核心匹配算法,利用小波多分辨分析技术对模拟实时图进行预处理,将改进的蚁群算法用于航迹规划,运用遗传算法的快速搜索优势进行景象匹配定位,构建出无人机平台景象匹配导航仿真系统。该仿真系统可模拟无人机的真实工作环境,并为无人机平台景象匹配导航的设计与开发提供技术支持。
Scene Matching Navigation Technology(SMNT) is getting to be one of the most important ways to navigate the UAV.With the property of autonomous and high accurate of locating,SMNT also has a uniquely outstanding characteristic of anti-electronic obstruction.In this paper,we have designed a simulation system of scene matching navigation for UAV.We adopt a wildly-used scene matching algorithm as the core which based edge-response weighted Hausdorff distances.Multi-resolutions of wavelet are used in our system to process the simulated runtime-image of camera on UAV.Route of UAV was planned by improved ant-gather algorithm and use genetic algorithm to optimize the strategy of search for fast scene matching.SMNT was designed to help researchers research and develop the suitable algorithms for UAV platform efficiently.