Fast Support Value-Based Star Acquisition Method

Abstract

Star acquisition is one of the most time-consuming routines in star tracker operation. Based on observation that the possibility of star pixels in a star image is proportional to the corresponding support value, a new algorithm is developed, wherein the determination of possible star pixels in a star image is the calculation of support value and thresholding. Based on the mapped least squares support vector machine (LSSVM) with the mixtures of Radial Basis Function (RBF) and polynomial kernels, operators to calculate the support values are obtained. With the deduced operators, the calculation of the support values is just only a convolution operation. In contrast to the traditional star acquisition methods, which make the decision as to whether a pixel is the possible star pixel or not purely based on the pixels intensity, the proposed approach make this decision based on the distribution of the neighborhood intensity, the support value. Computer experiments are carried out for extracting stars from simulated star images. The performance of the proposed algorithm is compared with many traditional methods, including traditional scanning method and Motari’s vector approach. The experimental results indicate that the proposed algorithm is efficient and robust over a wide range of sensor noise.