基于支持特征点的SAR辅助导航反演点置信度估计

合成孔径雷达(SAR)由于全天候、全天时等特点使得它正成为飞行器导航的理想传感器之一.SAR辅助导航是利用SAR实时获取的目标散射图像与飞行器中存储的参考图像进行匹配运算,进而修正飞行器的位置数据.但由于SAR为侧视雷达,当地形存有较大起伏时,实时图中某些位置的像素将产生偏移,使整个反演精度下降.基于此,以F.Leberl构象模型为基础,提出了基于支持特征点的反演控制点置信度估计方法.实验结果显示,根据置信度对适合精确匹配区域的数据进行优化选取,能较大提高反演飞行器空间位置的精度.

Confidence estimation of SAR navigator control point based on support feature point

Synthetic aperture radar(SAR) will be a perfect sensor for automatic aircraft navigation because it can work all-weather,day-or-night.Synthetic aperture radar seeker(SARS) gets new spatial coordinate of aircraft via matching a real-time SAR image with a reference image in the aircraft.SAR is the side-looking radar,and the changing of terrain can cause the image pixel displacement,which cause the decline of the inversion precision.A confidence estimation method of the inversion control point is proposed according to F.Leberl conformation model.The results of the experiments show that higher precision of aerocraft space position of the inversion can be achieved if we optimize the data adapted to the matching area according to the confidence.