一种基于BP网络的星图识别算法

A Star Pattern Identification Algorithm Based on BP Neural Network

为了提高星图识别算法的抗噪性能,提出一种基于逆向传播(Back Propagation,BP)网络的识别算法。该算法通过将星图转换成"0"、"1"和"2"的网格矩阵,提取行列数值和星数形成匹配向量,利用多个BP识别子网进行训练完成匹配识别。通过仿真试验得出以下结论:对星等位置和星等添加噪声,当位置噪声标准偏差为2像素时,BP网络算法的识别率和识别时间相对传统栅格算法分别提高2%和60ms,对噪声有较强抗干扰能力,表明BP网络算法具有更快的识别速度。

In order to improve the anti-noise property of grid algorithm,an identification algorithm based on back propagation(BP)network was proposed.It extracts the number of rows and columns to form matching vector by transfroming the star pattern into a grid matrix of'0','1'and'2'.Then multiple BP recognition subnets were used for training to complete the matching recognition.Simulation results indicate that,when the standard deviation of position noise is 2 pixels,the recognition rate and the recognition time of BP network algorithm is improved by 2%and 60 ms compared to grid algorithm after adding noise to both star locations and star magnitude.This algorithm based on BP shows strong anti-noise ability and faster identification speed.