加入地面控制点进行几何精校正后，采用后续的非线性最小二乘算法不断优化生

After addition of ground control points geometric precision correction, the subsequent use of nonlinear least squares algorithm constantly optimize sparse point cloud generation, and finally through intensive matching algorithm by matching pixel dense point cloud generation, complete with relative coordinates to restore the three-dimensional point cloud.

After adding the ground control point for geometric fine correction, the subsequent nonlinear least square algorithm is used to continuously optimize the generation of sparse point cloud, and finally the dense point cloud with relative coordinate series is generated by dense matching algorithm, and the recovery of three-dimensional point cloud with relative coordinates is completed.

After adding ground control points for geometric precision correction, the non-linear least square algorithm is used to optimize and generate sparse point cloud. Finally, dense point cloud is generated by pixel matching through dense matching algorithm to complete the recovery of 3D point cloud with relative coordinates.