Rapid star extraction and identification from star images obtained by drift-scan CCDs

Abstract

In the real-time orbit determination systems of synchronous satellites, the digital signal processor（DSP）system is introduced to accelerate the extraction and identification process and realize the real-time orbit determination of synchronous satellites. A simplest features fusion algorithm which can sharp star targets is proposed to reduce the time consumed by star extraction by computing and fusing the simplest features which are the maximum pixel gray value and the regional contrast. An optimized triangle identification method is presented to decrease the identification time by exploiting negative model on the basis of previous matching results and matching all stars extracted with reference stars in the catalog by adopting binary chop after analyzing the corresponding relativity between the background stars and navigation stars. Experimental results with the star image resolution of 1528×1528 show that our algorithm can significantly enhance the star extraction and identification speed, the extraction time is declined to 34ms, and the precision of star extraction via the proposed method is higher than that generated with the traditional star extraction method. The identification time is 58% of that consumed by traditional triangle identification algorithm, and the identification accuracy remains at 98.4%, thereby validating the efficiency and reliability of our algorithm.