In large‐scale block adjustment of high‐resolution satellite imagery (HRSI) without ground control points (GCPs), the selection of virtual control points (VCPs) is an important factor in determining the block adjustment accuracy. The traditional VCP generation method is based on a regular grid and uses the initial rational polynomial coefficient (RPC) files but without considering topographic factors. To further improve the accuracy, this paper proposes an approach for the optimal selection of VCPs with planar constraints, which ensures that the VCPs are mostly in areas with small elevation differences and are not influenced by vegetation and buildings. A three‐step approach generates the 3D coordinates of tie points based on forward intersection, calculates their mean values and finally selects optimal VCPs using a threshold value. Experiments conducted using ZY3‐01 satellite images show the accuracy of the proposed method without GCPs outperforms the existing method.

中文翻译：

---

具有平面约束的虚拟控制点的最佳选择，用于卫星图像的大规模块调整

在没有地面控制点（GCP）的高分辨率卫星图像（HRSI）的大规模块调整中，虚拟控制点（VCP）的选择是确定块调整精度的重要因素。传统的VCP生成方法基于规则网格，并使用初始有理多项式系数（RPC）文件，但不考虑地形因素。为了进一步提高精度，本文提出了一种具有平面约束的VCP的最佳选择方法，该方法可确保VCP大多位于高程差较小的区域，并且不受植被和建筑物的影响。采用三步法，基于正向交点生成联系点的3D坐标，计算其平均值，最后使用阈值选择最佳VCP。