论文部分内容阅读
为了提高在全天球自主工作模式下星图识别的成功率和鲁棒性,本文结合kNN算法和有向图理论的思想,构造出具有强约束的有序星点集模式,提出基于有序星点集的全天球自主星图识别算法。该方法首先利用k近邻算法的思想,以导航星点为中心,对位于其一定范围内的导航星进行了分类划分。然后基于有向图理论,以距离中心点星最近的导航星为基准,按照顺时针顺序对分类的导航星进行排序,构造出具有强约束特性的有序星点集作为星图识别的特征。实验结果表明:在存在星点位置误差和伪星点的情况下,本文提出的基于有序星点集全天球自主星图识别算法具有很强的抗噪声能力、抗伪星点干扰能力和鲁棒性。在星点质心位置达到3像素时,基于有序星点集星图识别算法成功率仍然可以达到99.8%,比三角形识别算法和栅格识别算法的识别成功率高16%以上;在存在3颗伪星点的情况下,基于有序星点集星图识别算法成功率为98.4%,比三角形识别算法和栅格识别算法高10%以上。
In order to improve the success rate and robustness of the satellite image recognition in the all-sky autonomous working mode, this paper constructs an ordered star-point mode with strong constraint based on the idea of kNN and directed graph theory, All-star self-contained star map recognition algorithm of star point set. The method first uses the idea of k-nearest neighbor algorithm to classify navigation stars within a certain range based on navigation stars. Then, based on the directed graph theory, the navigation star closest to the center star is taken as a reference, and the sorted navigation stars are sorted in a clockwise order to construct an ordered star set with strong constraint characteristics as a feature of the star map recognition. Experimental results show that in the presence of star position error and pseudo-star point, the proposed algorithm based on ordered star point set of all-celestial self-contained star map has strong anti-noise ability, anti-pseudo-star interference ability and Robustness. When the centroid of the star reaches 3 pixels, the success rate of the algorithm based on the ordered star set can reach 99.8%, which is more than 16% higher than that of the triangle recognition algorithm and the grid recognition algorithm. When there are 3 In the case of pseudo-star, the success rate of the algorithm based on the set of ordered star sets is 98.4%, which is more than 10% higher than that of the triangle recognition algorithm and the grid identification algorithm.