基于北斗地球同步轨道卫星多星成像的图像融合

Image Fusion Based on Multiple Satellite Imaging of Beidou Geosynchronous Orbit Satellite

利用大型轨道卫星系统(global nanigation satellite system,GNSS)导航卫星作为发射信号源的双基地雷达成像方法是一种新型的微波成像手段。由于单颗卫星成像存在空间分辨率不高、信噪比偏低等问题,采用了多颗北斗GEO卫星,获取了多幅卫星成像图,然后将最大互信息准则应用于多幅卫星成像图的配准,利用Powell算法将最大互信息作为目标函数搜索卫星成像图几何变换的最优配准参数;最后使用高频带融合方法对配准后的卫星成像图进行融合以提高图像质量。将配准后融合的卫星成像图与Google实景地图比较的结果表明,该多星成像融合算法使得信号增益提高了约20%;且避免了同一目标点在不同位置多次叠加产生的成像混淆,从而达到了提高成像精度的效果。

Bistatic radar imaging using global navigation satellite system（GNSS） navigation satellites as signal transmission sources is a new method of microwave imaging.To overcome the harm of the low spatial resolution and low signal-to-noise ratio of single satellite imagery,multiple Beidou GEO satellites were used to obtain multiple satellite images.Then,the maximum mutual information criterion was used to register multiple satellite images and used maximum mutual information as an objective function to search the optimal registration parameters in satellite images transformation（Powell algorithm）.Finally,the satellite imagery after registration by the high-band fusion method was fused to improve the quality of image.In the comparison with fused satellite imagery and Google real image maps,the conclusion shows that the signal gain has been improved by about 20% through multi-satellites imaging fusion algorithm.It also shows that multiple overlays of identical target points at different locations have been avoided,thus,the effect of improving imaging accuracy was achieved.