Star-point moving target detection and tracking on star space background

For the condition where star-point target detection and tracking real-time. This paper proposed a method of star-point target detection and tracking. The method includes two major steps, target detection and target tracking. In the detection stage, based on difference of inter-frame image in sequences, an adaptive threshold has designed. After morphological filtering and regional connectivity for segment image, using centroid extraction method to identify targets quickly. In the tracking stage, for the characteristic of star point target’s movement is linearity, the tracking method using least squares algorithm to estimate target’s moving trajectory. Based on recording target’s moving trajectory, an algorithm of estimating moving step length has proposed, it forecast targets’ centroid coordinate real-time. To track targets in accurate, a doubting grade checkout method has proposed too. The complete set of proposed methods used to apply in real star space sequence image. Extensive experiment results shows that, the methods are efficient, and it meets the requirement of detecting moving star-point target out of cluttered background real-time.