Research On Non-measuring Digital Camera Calibration Based On Improved Bundle Adjustment With Additional Parameters

In recent years,with the continuous development of modern manufacturing technology,the mechanical structure of ordinary digital cameras is more and more stable,and then gradually become the main equipment for close-range photogrammetry data collection,close-range photogrammetry also displays new vigor.Ordinary digital cameras are non-measuring cameras,which do not have the function of recording internal azimuth elements.Cameras must be calibrated to accurately measure internal azimuth elements,optical distortion and other parameters,otherwise the accuracy of photogrammetry results will be directly affected.Therefore,the method of digital camera calibration has become one of the research hotspots in the field of close-range photogrammetry and computer vision.This paper takes digital camera calibration as the research object.Firstly,the bundle adjustment with additional parameters is researched and implemented,and the characteristics of this method are analyzed.Secondly,an improved bundle adjustment calibration algorithm with additional parameters is proposed to overcome the shortcomings of this method.The three-dimensional direct linear transformation solution is used to provide the initial value for the algorithm,which satisfies the need of high precision initial value for the bundle adjustment method.In the algorithm,the coordinates of the control points are regarded as true values,and the internal parameters and external elements of the camera are regarded as weighted observations,which reduces the number of unknowns,optimizes the adjustment model of additional parameters,and solves the problem that the coefficient matrix of the normal equation is prone to rank defect and can not be solved.At the same time,considering that the digital camera has relatively large scale inconsistency error and non-vertical error,these two kinds of error parameters are introduced into the additional parameter model to improve the calibration accuracy.Finally,taking the indoor 3D control field of Wuhan University as an example,the camera calibration experiment is carried out,and the accuracy of the calibration results is evaluated.In order to verify the reliability of the proposed algorithm,a comparative experiment was conducted between the conventional algorithm and the proposed algorithm.The accuracy of the 3D direct linear transformation method was 0.802 mm,the accuracy of the bundle adjustment method with additional parameters was 0.724 mm,and the accuracy of the proposed algorithm was 0.648 mm.The experimental results show that the proposed method has strong applicability and high calibration accuracy,and can meet the application requirements of photogrammetry.