【摘要】 针对二维图像三维重构时对输入数据精度要求比较高的问题,采用Maya建模,运用直接线性变换(DLT)定标系统进行定标实验.对于各种输入扰动进行测试与误差分析,总结出各种输入误差的影响,并使用光线追踪算法进行重构实验,检测参数的误差对最后重构结果造成的影响.相机定位不精确会对重构物体的完整性造成较大的影响,照片边界坐标不精确、旋转角度不精确会根据偏差的程度,在相应的坐标轴方向上对重构结果造成影响.通过分析误差对重构结果偏差的影响,设计调整输入可以减小重构误差的方案.更多还原

【Abstract】 Direct linear transformation (DLT) system was used for calibration with Maya modeling in order to achieve the high accuracy of input parameters in three-dimensional reconstruction from two-dimensional images.The influence of the perturbations in input data was tested and analyzed.Ray-tracing algorithm was adopted to reconstruct model and get the parameters’ influence to the reconstruction results.The inaccuracy of camera positioning had great influence on reconstruction.The inaccuracy of photo border coordinates and rotated angle also caused error on their axis direction in reconstruction.The different influence in reconstruction by different error was analyzed,and a strategy was designed to reduce the error in reconstruction.更多还原