A new star identification algorithm based on improved hausdorff distance for star sensors

High accuracy attitude for spacecraft can be determined by star sensors. The key technology is star pattern recognition. Based on the Hausdorff distance (HD) algorithm, a robust identification algorithm has been used for matching point-sets, which does not require absolute point correspondence. HD identification is not suitable for large attitude angle changes around the boresight, which usually results in a low recognition rate and low speed of identification. A new image similarity measure combined with an improved HD algorithm is proposed for recognizing stellar maps. An improved HD based on scalar distances is implemented to guarantee an acceptable success rate of recognition for large attitude changes around the boresight, which is noise resistant. Another improved HD based on vector distances is constructed to guarantee the recognition speed using a star dimensional configuration. Appropriate gray and matching thresholds are selected to improve recognition speed. Results of the semiphysical simulation show that the proposed algorithm is better than the HD identification method in terms of noise resistance, recognition rate, and speed of operation.