基于PSF相关的星点定位算法

为提高非高斯分布星点定位的精度，提出了一种新的PSFC (point spread function correlation)星点定位算法，具有很好的抗噪性能和精度水平，易于工程实现.该算法利用互相关中的定义，通过确定系统所测定的PSF与星像灰度值之间的最大相似度来定位星点位置，PSF的测定是在作互相关计算之前全视场范围内可按照整像素和亚像素两种方式完成的.仿真实验结果表明：在带有噪声的星像图条件下，本文PSF相关法的定位精度最高可达到0.034像素，比质心法的精度提高1个数量级，比曲面拟合法的精度提高约20%.

Star Locating Algorithm Based on PSF Correlation

In order to improve the star locating accuracy with non Gauss energy distribution for a star tracker, a new locating algorithm based on PSFC (point spread function correlation) was proposed. The new method has good anti noise performance, high accuracy and engineering realization. The PSFC locating algorithm is realized by using cross correlation definition and determining the maximum similarity of the measured star PSF and the star gray images of the star tracker. The measured star PSF are prepared before cross correlation computation by all the FOV sub pixel measurements. The simulation results show that the accuracy of PSF correlation method considering high noise background can reach 0.034 pixels, and is higher than the traditional centroid method of 1 orders of magnitude , and is higher than the Gauss surface quasi legal of about 20%.