Fast recognition on star pattern with method of three FOVs based on rough location and azimuth

Abstract

In order to quickly navigate with the known rough azimuth and location information by three FOVs positioning and orientation device,a fast recognition method for local star pattern is proposed. First,we analyse the reasons of the low efficiency when performing the stars identification using global recognition database.Next,we note that it is a preferred method to perform star identification in one FOV firstly and identify the remaining stars between different FVOs,and give the angular distance error threshold. Then,we propose a method to generate the local recognition database,which can reduce the redundancy of identification information and improve recognition efficiency. The simulaion and field experiment results show that the correct iden-tification rate using this recognition method is improved up to 99. 19%,and rocognition time is about 24. 3ms,which can satisfy the system requirements for fast and efficient navigation. It also proves the correctness of the preferred recognition order.