Accuracy Assessment of Establishing 3D Real Scale Model in Close-Range Photogrammetry with Digital Camera

Abstract

Three-dimensional (3D) real-scale models delivered from digital photogrammetric techniques have rapidly increased to meet the requirements of many applications in different fields of daily life. This paper deals with establishing a 3D real-scale model from a block of images (18 images) captured using a Canon EOS 500D digital camera to cover a test field area consisting of 90 artificial target points, 25 of which are ground control points (GCPs). At the same time, the remains are checkpoints (CPs). The analytical photogrammetric processes, including the calculation of interior orientation parameters (IOPs) of the camera during the camera calibration process and exterior orientation parameters (EOPs) of the camera in each capture. The model's object space (ground) coordinates are calculated simultaneously based on the collinearity equation using the bundle block adjustment method (BBA). Assessment and validation of the accuracy of the results is an important task in this study that was implemented to determine and analyze the errors of 3D coordinates through linear regression analysis (LRA). Root mean square error (RMSE) is the statistical parameter used in the statistical analysis of results. The standard error is another statistical parameter used to evaluate the accuracy of camera locations and rotation angles (EOPs). The total RMSE (RMSE) xyz of GCPs is ± 2.530 mm, while the total RMSE (RMSExyz) of CPs is ± 2.740 mm. The overall accuracy of the work is 5.000 mm.