摘要
随着我国月球探测计划的开展,基于视觉的探测器月球表面软着陆的相关技术研究正在进行,陨石坑是月球表面最常见的物体。基于图像的陨石坑识别技术作为探测器自主障碍检测中的一项关键技术,引起了各航天大国的高度重视。提出了一种基于特征点的陨石坑检测算法。该方法可以分为三个部分:特征点检测、陨石坑区域初选、陨石坑拟合。首先通过特征点检测初步确定陨石坑所在区域,然后通过区域生长的方法分别提取陨石坑亮、暗两区域,最后通过椭圆拟合的方式获得陨石坑所在椭圆。实验研究表明,该算法可以有效地检测出半径小于15个像素大于5个像素,有较强明暗对比的陨石坑。在结束语中,作者提出了未来该算法改进的四个方向。
As the development of China's Lunar Exploration, Vision-Based technology of lunar probe autonomous landing is researching. Craters are eonnnonly found on the surface of moon. The crater detection from surface images, as a key technology of Autonomons Hazard Avoidance, has been researched by many scientists from different countries. In this paper, a new algorithm based on feature point is demonstrated. The algorithm is divided into three parts, feature points extraction, candidate area of crater decision and crater detection. Firstly, candidate areas of crater are decided by feature point extraction. Secondly, the light and shaded parts of crater are extracted by region growing. Finally, the craters are detected by ellipse detection. Experiment result shows this algorithm is effective for some craters detection. Those craters have strong intensity variations and their radiuses are longer than 5 pixels and shorter than 15 pixels. In the conclusion, authors offer four improvement directions of this algorithm in the future.
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2009年第3期1243-1248,共6页
Journal of Astronautics
基金
南京航空航天大学博士创新基金(BCXJ07-06)
江苏省研究生创新基金(CX07B-113z)
关键词
陨石坑检测
特征点检测
区域生长
椭圆提取
Craters detection
Feature points detection
Region growing
Ellipse fitting