嫦娥三号着陆器下降序列影像匹配方法研究

Research on Image Matching Method of Falling Sequence of Chang'e-3 Lunar

通过对嫦娥三号着陆器下降序列影像特点进行研究,联系国内外影像匹配算法的发展历史与优化趋势,针对所获取的嫦娥三号下降序列影像缩放旋转平移变换多样的特点,比较各种图像匹配的方法,选择一种尺度不变特征转换算法(Scale Invariant Feature form),简称SIFT算法。根据SIFT算法的主要过程,用MATLAB软件对算法代码进行实现,经过调试后对嫦娥三号下降序列影像进行匹配,设计实验从匹配的正确率、匹配的时间和匹配的精度3个层面对匹配方法的性能进行分析,试验结果表明此算法适用于嫦娥三号下降序列影像的匹配。该算法适用于下降序列影响的匹配中,得到的数据为航天探测的后续操作提供精准的数据基础。

Soft-landing of lunar rover becomes a highlighted issue of next lunar probing project. How to obtain the Chang′e-3 landerdescent sequence influence precision matching and then obtain accurate landing point coordinates to achieve its precise soft landing isparticularly important. In this paper, through the study of the image characteristics of the descending sequence of the “Chang′e-3”,contact at home and abroad, according to image features, select a method called“Scale Invariant Feature form”, referred to as SIFT.According to its main process, we design code. The performance of the matching method is analyzed from three aspects of matching accuracy, matching time and matching accuracy after debugging. The experimental results show that this algorithm is suitable for the image matching of the descending sequence of Chang′e-3. The algorithm can be applied to the matching of the strong impact of the falland provide accurate data base for the subsequent operation of space exploration.