阐述了动态星敏感器星图的特点,指出了目前国内外星体目标提取方法的不足。针对这些不足,提出了一种新的星体目标提取方法。根据星敏感器运动特性,自适应选取处理窗口的大小及方向,自适应窗口可以适应不同形式的星敏感器运动。在选取的窗口内进行星体目标提取,当星体存在断裂情况时,利用数学形态学的方法对断裂星体进行断点判定及主动生长。对提取的星体目标利用灰度加权的质心法得到其中心位置。最后给出了利用上述方法对2(。)/S动态模拟星图进行星体目标提取及定位的结果。实验结果表明,该方法可适用于高动态星敏感器,并能有效提高星体定位精度。
The character of dynamic star sensor's sky image and the disadvantages of existing star extraction methods at home and abroad were presented. To solve these problems, a new star extraction method was put forward. Firstly, according to the character of the star sensor's movement, the size and direction of the region of interest (ROI) were chosen. The adaptive ROI could fit different types of movement. Then, stars were extracted in the ROI, and mathematical morphology methods were used for the broken stars. By the mathematical morphology method, broken points were judged and broken stars were grown autonomously. The locating of stars was calculated by centroid algorithm. Finally, the star extraction and locating results of a simulated sky image in 2 (°)Is dynamic condition were presented. The experimental results indicate that the method can be suitable for high dynamic star sensor and improve the precision of star locating effectively.