三角形星图识别改进算法研究

三角形识别算法是目前采用星敏感器为导航部件所广泛运用的算法，但由于特征维数较低，容易造成冗余匹配和错误识别。为了提高三角形识别算法的识别正确率，提出了改进识别过程的三角形识别算法，先将满足三边判决门限范围内的导航星对记录到匹配数组，对出现的导航星进行计数，利用三角形三边之间两两相交的相关性，将导航星出现次数小于2的星对作为整体剔除，在余下的导航星对中寻找与观测三角形同构的三角形。识别过程中，以星对角距和相对星等差为识别特征，并对识别特征构造了散列函数，减少了识别过程中特征量比较的次数，提高了识别效率，同时，也减小了内部导航星库的存储容量。

Research on modified triangle identification algorithm

Triangle star identification algorithm is widely used in attitude determination system, that adopts star sensor as navigation parts at present. But low characteristic dimensions lead to high percentage of redundancy and mistake. To improve accuracy of triangle identification algorithm, a modified algorithm which based on traditional one is proposed. Record the star pairs that meet interstellar angle threshold in star matrix, and count the guide star’s appearance simultaneously, according to the truth that each star must be attached to two sides of the triangle, eliminate the entire star pair which any particular star appears less than two times, then construct guide triangle from stars left in star matrix. Only take interstellar angle and relative magnitude of stars as pattern, decreases the capacity of the database, enhances the validity of results, for another thing, hashing search method used during star identification speeds up the identification process.