SAR影像的RPC模型参数求解算法研究

The algorithm of computing RPC model's parameters for SAR image

首次对ERS-SAR卫星影像进行了基于SRTM DEM无需初值的RPC模型参数求解试验,对比了9种形式RPC模型参数的求解精度,并对控制点格网大小及高程分层对参数求解精度的影响作了评价。试验表明SAR卫星遥感影像的成像几何可以用RPC参数进行拟合,可以和光学遥感影像一样进行摄影测量处理,简化SAR影像几何处理的难度;对于ERS-SAR影像,阶数为3并且分母不同的RPC模型可以取代严格成像模型进行摄影测量处理,在利用严格成像模型求解RPC模型参数时候,采用控制点的格网大小为30×30、高程分层为3可以达到精度和效率的平衡。

The rational polynomial coefficient （RPC） model has recently raised a considerable interest in the photogrammetry and remote sensing community. This model is a generalized sensor model tha.t is capable of achieving high approximation accuracy. Unfortunately, in all previous literatures, the computation of the parameters of the RPC model depends on initial values for the parameters. In this paper, an algorithm for the computation of the parameters without these initial values is presented. Finally, the algorithm is tested on ERS-1 imagery. Based on numerous tests, the following conclusion can be drawn. （1） The RPC model can achieve an approximation accuracy that is extremely high for satellite data. The results support the view that the RPC model can be used as a replacement for sensor models for photogrammetric restitution. （2）The high order RPC model may be necessary when satellite data sets are dealed with, because the RPC model resembles the rigorous sensor model very well. （3）The RPC model cases with unequal denominators achieve better accuracy than the cases with equal denominators at check points overall. （4） In establishing the 3D object grid for the RPC model solutions, at least two or more elevation layers are needed. （5） For satellite imagery, when the image grid contains 30 × 30 points and the number of elevation layers is three, the precision and the efficiency are in balance. This study finds that the RPC model is suitable for both push-broom and SAR imagery.