Joint Stereo Positioning Based on SAR/CCD Satellite Images Introducing Virtual Observation Weights

Joint stereo positioning based on multi-source satellite images is essentially a hybrid model of different type observations. The reasonable determination of the weight ratios of various observations is an important prerequisite for obtaining reliable positioning results, which has not been considered in previous research. So in this paper, it is to be explored that how to construct a robust adjustment model and give full play to the advantages of various information sources in the adjustment to further improve the joint positioning accuracy and stability. We adopt the RPC block adjustment modeling based on the image compensation. Then a weight assignment method of virtual observation values is presented based on the error source analysis and the prior accuracy information of SAR/CCD images. Experiments under different image combinations have verified that our method is a useful proposal for the joint positioning of heterogeneous images. It is shown that the image priori plane positioning accuracy is an important basis for the weight determination of the virtual observation values of the image compensation parameters. Under reasonable weight allocation, the optimal uncontrolled plane/elevation accuracy of heterogeneous joint positioning can reach within 2m.