Research On Accuracy And Self-calibrating Bundle Adjustment Of Close-range Photogrammery For High Speed Railway

The railway tracks are the infrastructure facilities that lead directly high-speed vehicles running safely and smoothly.The good geometric state of the track is important guarantee for the realization of high-speed vehicles running smoothly.Therefore,it is necessary to strengthen the examination and maintenance of railway tracks.At present,the method of track static smoothness inspection is of high precision,but the efficiency of inspection needs to be further improved.In addition,the increased speed of trains brings great challenge to the efficiency and accuracy of track geometric state inspection.The close-range photogrammetry can be applied to track geometric state inspection and it is promising as a new technology for improving efficiency and accuracy.In this thesis,the vehicle-borne close-range photogrammetric technology is used to collect high-resolution track images.Using the acquired close-range photogrammetric images,a method to improve the accuracy of the self-calibrating bundle adjustment of close-range photogrammetry for high-speed railway track is explored.Encryption point coordinates and exterior orientation elements are calculated accurately,w hich provides basic data for track static smoothness inspection.There is a phenomenon that appears ill-posed and rank defect of normal equation in close-range photogrammetric images adjustment.In this paper,the self-calibrating bundle adjustment consists of two kinds of virtual observations,including virtual observations of image coordinates and of additional parameters.The weight estimation of these two kinds of virtual observations is discussed by using a posteriori variances estimation.It can avoid ill-posed normal equation effectively and ensure the stability and convergence of the self-calibrating adjustment results.For unknown lens distortion of non-metric digital camera used by the vehicle-borne close-range photogrammetric platform,the test of significance and the correlation analysis of additional parameters are carried out by the use of the hybrid additional parameter model and the standard self-calibration model.Those parameters which will bring out high correlations among the unknowns are removed.Meanwhile,by comparing precision of new two additional parameter models,it proves that the image systematic errors can be compensated efficiently by the use of additional parameter model with 10 parameters,and a stable and reliable result of the self-calibrating adjustment can be ensured.For high-speed railway in long-wide area,this paper designs kinds of the distribution solutions of control points.Through short-distance and long-distance of close-range photogrammetry for high-speed railway contrast experiment,it proves that control points are evenly distributed at the edge of high-speed railway,which can be more suitable for high-speed railway of close-range photogrammetry in long-wide area.The precision of the whole surveying area can be ensured and the result of the self-calibrating adjustment can be improved.