应用灰度特征的行星陨石坑自主检测方法与着陆导航研究

Automated Crater Detection Method Using Gray Value Features and Planet Landing Navigation Research

针对陨石坑阴影不明显的行星着陆导航图像,提出了基于图像灰度值特征的陨石坑自主检测方法,通过兴趣区快速确定陨石坑边缘分布,解决了一般陨石坑检测方法对图像太阳高度角的依赖问题。利用检测出的陨石坑视线信息,采用非线性预测滤波方法估计着陆器着陆阶段的位置和姿态;鉴于视觉着陆导航对照了场景状况的高度依赖性,提出了一种利用分形理论和相邻点分级方法的天体随机地景建模方法。最后验证了所提方法的有效性。

An autonomous crater detection method based on the image gray value feature is proposed in this paper. In this method, region of interest （ROI） is used to identify crater＇s edge distribution from real planet landing navigation images conveniently, thus avoiding dependence of image on solar elevation. Then an algorithm for determining both vehicle attitude and position by use of the nonlinear predictive filter is also described. Considering that landing terrain plays a significant role in celestial body landing navigation, a method of terrain map simulation is proposed. By sorting surface points into levels, new points completely depend on their higher-level neighbors according to the fraetal theory. A celestial body random terrain map modeling method is proposed. Finally the paper presents the validation of the algorithms in the end.