从恒星背景中检测出空间目标是空间目标天基光学测量相机所要解决的关键技术难题之一。提出了一种基于星图识别的空间目标检测算法,利用背景星图的平移、旋转、比例伸缩不变性,应用星图识别原理,根据识别恒星在不同帧图像中的坐标变化估计背景运动参数,进行探测图像序列的配准,然后在配准后的图像序列中检测出空间目标。通过仿真验证,该方法对于运动恒星背景中运动小目标的检测具有比较好的效果。

Space object detection in the star background is one of the key techniques in space-based optical measurement for space object. A new space object detection algorithm is put forward which is based on the principle of star map recognition. At first the moving parameters of the detecting star background is evaluated according to the coordinates variety of stars recognized in different frame of images using star map recognition algorithm based on the inalterability for translation, and revolution, and proportion dilation of star map. Then the image registration is accomplished using the evaluated moving parameters of background. At last, the space object is detected in the registered image series by use of the target detection algorithm in static background. It is demonstrated that this method have a good effect on the moving small space object detection in the moving star background through simulation test.