基于立体影像的有理多项式参数自动精化

Automatic refinement of rational polynomial parameters based on stereo imagery

卫星影像几何模型从严格传感器模型转换为有理函数模型时,像点拟合精度都能达到较高精度,但其定位误差会增大,通常需要地面控制点或基于RPC的区域网平差予以解决。本文提出基于立体影像RFM建模及RPC精化方法,无需外部地面任何控制数据,解决单航带影像模型转换后定位误差变差问题,并利用实际卫星影像数据进行试验验证。试验结果表明,该方法既不受地面控制数据限制,又能解决三维误差变大问题,可实现对立体影像RPC参数自动精化,有效保证模型转换后定位精度的一致性。

When the geometric model of satellite imagery is converted from the strict sensor model to the rational function model(RFM),the fitting accuracy of the image point is high while its positioning error increases.Usually,the errors can be solved using ground control point(GCP)or block adjustment based on the rational polynomial coefficient(RPC).In this paper,the methods of RFM modeling and RPC refinement method are proposed to solve the problem of low positioning accuracy after the conversion of the single-airband imagery without any external ground control data.Finally,the accuracy is verified using actual satellite imagery data.The experimental results show that the new method can solve the problem of low positioning accuracy after model conversion without the limitation of ground control data.It can realize the automatic refinement of RPC parameters of the stereo images,and also ensure the consistency of positioning accuracy after model conversion.