机载激光雷达景像匹配辅助导航方法  

Airborne LiDAR Scene Matching Aided Navigation Method

作　　者：王晓龙 赵龙[1] Wang Xiaolong;Zhao Long(Beihang University,Beijing 100191,China)

机构地区：[1]北京航空航天大学,北京100191

出　　处：《航空科学技术》2024年第12期40-46,共7页Aeronautical Science & Technology

基　　金：国家自然科学基金(42274037);航空科学基金(2022Z022051001)。

摘　　要：以可见光视觉传感器方案为代表的景像匹配导航方法容易受光照、天气等外部环境变化影响,无法完成图像信息的有效采集,导致导航系统误差发散。为解决该问题,本文提出了一种机载激光雷达景像匹配辅助导航方法。首先,对激光点云数据进行预处理,拼接激光点云数据并提取点云强度信息构建激光点云强度图像;其次,基于异源图像特征匹配方法提取激光点云强度图像与先验基准地图中的特征信息进行特征匹配;最后,从特征匹配结果中提取匹配定位信息,将其引入惯性/景像匹配融合导航框架,进而校正惯性导航系统发散误差。经真实飞行数据测试验证,本文方法能有效抑制惯性导航系统的误差发散,其导航精度≤2m,并且可实现全天时条件下应用。Scene matching navigation methods based on visible light visual sensor solutions are often vulnerable to external environmental changes such as illumination and weather,which leads to ineffective image information collection and divergence of navigation system errors.To address this issue,this paper proposes an airborne LiDARbased scene matching aided navigation method.Firstly,the LiDAR point cloud data is preprocessed,stitched,and the intensity information of the point cloud is extracted to construct a LiDAR point cloud intensity image.Secondly,by employing a method for heterogeneous image feature matching,feature information from the intensity image and the prior map is extracted for matching.Finally,the extracted matching positioning information is integrated into the inertial/scene matching fusion navigation framework to correct the divergence error of the inertial navigation system.Through experimental validation with real flight data,the proposed method can effectively suppress error divergence of the inertial navigation system,with a navigation accuracy within 2m,and can be applied under all-weather conditions.

关 键 词：激光雷达 图像匹配 景像匹配导航 惯性导航 融合导航 

分 类 号：V249[航空宇航科学与技术—飞行器设计]