星点亚像元细分定位精度分析

摘要

Precise positioning of the star could provide key data for subsequent navigation calculation. Various star point’ s numerical simulations were characterized to vertify the positioning method, according to the principle of Gauss distribution. Centroid algorithm and Gaussian surface fitness algorithm were introduced in this paper, the effect factors on star position accuracy were studied, which include the gray system error, random noise, segmenta-tion threshold, magnitude and coordinate distribution. The similarities and differences of the two star position loca-ting method were indepth analysed. According to the study of simulation experiment, distance between star and ref-erence point is farther, positioning error is larger, and the positioning error rises with random noise increasing. The experimental results show that, compared to the Gauss surface fitness method, centroid method has better practica-bility and its positioning accuracy has less dependence on pixel information of star point.%星点的精确定位为后续的导航解算提供数据支撑。基于高斯分布原理进行星点模拟，为星点定位提供图像数据。分别介绍了质心法和高斯曲面拟合法的亚像元细分定位原理，并研究系统误差、随机噪声、分割阈值、星等和星点坐标分布5个因素对星点定位精度的影响，分析两种方法的异同点。对模拟的星点进行定位分析发现，星点距离参考点越远定位误差越大，定位精度随噪声增大而降低。实验表明，相比高斯曲面拟合方法，质心法具有更好的实用性，其定位精度对星点的像元信息依赖性较小。