Automated Updates of Coordinates of Ground Control Points Through Tiepoints from Multiple Satellite Images

For utilization of satellite images, enhancing the geometric accuracy and the provision of accurate ground control points (GCPs) are essential. However, maintaining and updating numerous GCPs is time-consuming and costly, presenting considerable limitations. To improve these challenges, this study proposes an automated method to accurately adjust GCP height values using rational function model (RFM) bundle block adjustment with multiple satellite images. Tiepoints over multiple images were constructed through automated matching between satellite images and GCP chips. We converted true GCP height values into heights with errors. The GCP height values were iteratively adjusted through bundle block adjustment using tiepoints over multiple images. The estimated height values were compared with the true GCP height values. Experiments compared and analyzed the accuracy of height value adjustments based on the number of satellite images used, imaging geometry, and the weights assigned in the model. Results with 13 high-resolution images showed that the root-mean-square-error (RMSE) of GCP height values improved from 8.959 m to 0.863 m after adjustment, achieving an accuracy within 1 m. Moreover, as the number of satellite images used in the bundle adjustment increased, the RMSE gradually decreased, leading to more accurate estimations. When using satellite image datasets with diverse imaging geometries, the RMSE was 0.931 m, whereas datasets with similar imaging geometries resulted in RMSEs of 1.228 m and 1.473 m, indicating lower adjustment performance. The optimal weight setting involved assigning lower weights to the initial GCP heights compared to other parameters, allowing for more significant adjustments. We highlight that tiepoints over multiple images were constructed through automated matching between satellite images and GCP chips. This supports strongly the automated update of GCP’s ground coordinates precisely. Experiment results indicated that the proposed method could be effectively utilized for practical GCP management and that it improves the quality of GCPs in areas where accurate field surveys are challenging.