This paper provides geometric accuracy improvements for very high resolution WorldView‐2 (WV2) Ortho Ready Standard Level‐2A (ORS2A) satellite images in the panchromatic (PAN) band. The entire process can be divided into two steps: (1) sensor orientation, which was conducted on 40 ground control points (GCPs) to determine which sensor model was best depending on various numbers of GCPs; and (2) orthorectification, whose main objective was to compare the accuracy of the process when using different freely available global digital elevation models (DEMs). The research showed that a sensor model based on rational polynomial coefficients (RPCs) with a shift or zero‐order bias correction attained the best geopositioning accuracy during the orientation step with the use of seven GCPs. The highest accuracy was achieved for images orthorectified with the ALOS World 3D‐30 m DEM (AW3D30).

中文翻译：

---

使用免费的DEM数据提高WorldView-2图像的几何精度

本文为全色（PAN）波段的高分辨率WorldView-2（WV2）正交标准2A（ORS2A）卫星图像提供了几何精度改进。整个过程可分为两个步骤：（1）对传感器进行定向，该定向是在40个地面控制点（GCP）上进行的，以根据各种GCP的数量确定哪种传感器模型最好。（2）矫正，其主要目的是在使用不同的免费提供的全球数字高程模型（DEM）时比较过程的准确性。研究表明，基于有理多项式系数（RPC）的位移或零阶偏差校正的传感器模型在定向步骤中使用七个GCP可以达到最佳的地理定位精度。