|
Robust and efficient detection of point targets in the starry sky background is very important in the space situational awareness system. A method for the extraction of the space target motion trajectory is proposed using a long exposure single-frame image of the starry sky background. First, pre-processing operations are performed such as image denoising, threshold determination, and image segmentation on the acquired image, and the centroid coordinates of each target point are extracted, and then the line segments combined by the nearest neighbors of each discrete point are computed, and the information in a designed data structure is stored. The stored structure information includes the endpoint coordinates, angle, length, etc., then we take angle as the horizontal axis, and the number of corresponding line segments under this angle for the vertical axis. Since the space target moves almost at almost fixed direction and nearly constant linear speed in the limited time of period, through the chart, line segments with almost the same angle will be collected. In these line segments ends exists the space target at different time. For example, if there are N line segments, there will be 2N space point target positions where N-1 points forms the inner 2N-2 positions and the other 2 points are the end point positions. By the help of length information of the data structure the speed of the space target can be calculated in pixels. By the connection of these N+1 points or N line segments a whole line is formed. Besides the position information, the angle of the line segments which are movement direction of the space target is obtained and the connection line is the target motion trajectory. The experimental results show that the algorithm is simple, direct, and fast. It can effectively extract the trajectory of single or multiple space targets in a long-exposed single-frame image with relatively few stars in the scene. The detection time of the algorithm is proportional to the number of stellar point targets and space target points in the image. The detection efficiency and accuracy of the algorithm will decrease as the stellar point targets increase and will increase with the number of space target points in each image.
|
