月球探测器着陆中基于被动图像的着陆场搜索及斜坡估计  

Passive Image-based Safe Landing Site Searching and Slope Estimation in Probe Landing

作　　者：丁萌[1] 曹云峰[1,2] 吴庆宪[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016 [2]南京航空航天大学高新技术研究院,南京210016

出　　处：《宇航学报》2009年第6期2258-2264,共7页Journal of Astronautics

基　　金：江苏省研究生创新基金(CX07B-113z);南京航空航天大学博士创新基金(BCXJ07-06)

摘　　要：在无人探测器月面软着陆技术研究中,因为月球表面障碍物对探测器着陆安全构成的威胁,所以首先需要解决的是安全着陆区域确定及其坡度估计的问题。从被动图像角度出发,研究了基于灰度变化的着陆区域确定和基于两视图几何的候选着陆区域坡度估计的方法。首先,研究了一种基于多尺度窗口区域内灰度变化标准差的安全着陆候选区域确定方法;其次,为了计算基础矩阵F和单应矩阵H,本文利用SIFT算法和RANSAC策略进行特征点检测和跟踪;最后,本文利用平面单应H和基础矩阵F的相容性原理,提出了一种斜坡坡度估计方法。和同类方法相比,该方法无需探测器位姿转移参数。利用该方法得到的斜坡估计结果和其他方法的结果比较后可知:本文提出的方法检测出的坡度误差能够满足着陆过程中斜坡检测的要求。In the research of lunar probe autonomous landing, the first step is to search safe landing site. In this paper, landing site searching based on Intensity variation and slope estimation algorithm based on two views geometry was developed. The paper can be divided into three sections. Firstly, Multi-scale Window Intensity Standard Deviation was introduced for Safe Landing Candidate Region searching. Secondly, Scale Invariant Feature Transform and Random Sample Consensus were employed for leatare point extraction and matching robustly and accurately. Lastly, this paper used principle of homograph matrix H compatible with fundamental matrix F to estimated slope. Compared with similar algorithm, this algorithm did not need movement parameters of probe attitude and position. Relative experiment demonstrates that result of this method can satisfy requirements of slope estimation in probe landing.

关 键 词：被动视觉 着陆场搜索 斜坡估计 相容性 

分 类 号：V249.122.9[航空宇航科学与技术—飞行器设计]