Star identification algorithm is the key of posture measurement for star sensor. In order to solve the problems of low identification speed and poor robustness, this paper presents a new method which is based on the star identification algorithm of the BP neural network by referring to the idea of Grid algorithm. The grid algorithm can divide a star map into a grid, which can be transformed into a matrix. In this paper, the method of meshing is improved so that the newly generated matrix is the input sample. About the output sample, each star will be numbered and each number represents a star. So the output sample can be represented by the number in binary system. The classifier uses K-MEANS algorithm to achieve clustering of unsupervised similar samples. Finally, the simulation experimental results show that the success rate of the method is ninety-nine percent and recognition speed is greatly improved. After adding a variety of noise tests, there is still a high recognition rate. The method in this paper can improve the recognition speed and robustness.
|