中文摘要：

针对低轨变光空间目标和较亮空间目标的跟踪不稳问题，采用主动轮廓模型，提出一种改进的GVF—Snake算法，实现了星像实际轮廓的实时搜索．结合Kalman滤波外推，实现了一种空间目标的自适应跟踪方法．实验表明，跟踪过程中方法能够克服采用固定窗口带来的跟踪误差，提高跟踪的稳健性．

英文摘要：

Aimed at the unstable tracking problem of low-orbit variable and bright space objects, an active contour model is accepted, and a refined GVF-Snake algorithm is proposed to realize the real-time searching of the real contour of objects on CCD image in this paper. Combined with the Kalman filter for prediction, a new adaptive tracking approach is proposed for space objects. Experiments show that the method can overcome the tracking difficulty brought by a fixed window, and improve the tracking robustness.