论文部分内容阅读
卫星星图中恒星像点质心定位是星敏感器定姿的基础。质心提取的精度将直接影响到星敏感器定姿的精度。本文首先介绍和分析当前常用的几种星像点质心提取算法,如灰度加权法、高斯积分法、高斯细分法、高斯曲面解析法等。并在此基础上,利用我国首颗民用立体测绘卫星——资源三号卫星APS星敏感器下传的星图,对比各种算法星象点质心提取精度。试验以恒星星对角距为准确参考来评价星像点质心提取精度,结果表明,相对于传统灰度加权法,高斯细分法(3×3窗口)和高斯曲面解析法(3×3窗口)的结果更优,所得到的角距误差平均值分别为-3.92″、-3.77″,标准差分别为16.80、16.96″。
Satellite image star in the position of the centroid positioning of the star sensor is the basis of attitude. Centroid extraction accuracy will have a direct impact on the accuracy of the star sensor pose. This article first introduces and analyzes several common centroid centroid extraction algorithms, such as gray weighted method, Gaussian integral method, Gaussian subdivision method, Gaussian surface analysis method and so on. On the basis of this, we use the star map of APS star sensor, the first civilian stereoscopic survey satellite in China, to compare the centroid extraction accuracy of various algorithm astronomical points. The accuracy of star centroid centroid extraction was evaluated by accurate reference of the angular distance between stars and stars. The results show that compared with the traditional gray weighted method, Gaussian segmentation (3 × 3 windows) and Gaussian surface analysis (3 × 3 windows ), The average value of the angular error obtained is -3.92 “and -3.77” respectively, and the standard deviations are 16.80 and 16.96 ", respectively.