New Parametrisation of Bundle Block Adjustment for Self‐Calibration of Terrestrial Laser Scanner (TLS)

ABSTRACT Similar to camera calibration, three types of parameters: calibration parameters (CPs), exterior orientation parameters (EOPs) and object points (OPs) must be estimated through the bundle block adjustment (BBA) of the designed network for TLS calibration. The high precision of estimated parameters is guaranteed as the consequence of decorrelated parameters. In this research, the innovative parametrisation for exterior orientation of the bundle block adjustment using various Structure from Motion (SFM) approaches is proposed for TLS self‐calibration networks. Rather than applying a simplified version of collinearity conditions from a single scan station, which results in highly correlated parameters, the new parametrisation is established to decorrelate the parameters and increase their precision. The reliable formulation is built on the construction of modified and constrained collinearity conditions via coplanarity conditions without dependency on priori information of the object points. Under these conditions, the relative orientation parameters are gradually added to verify the precision and correlation of the parameters of the BBA. The entire methodologies concentrate on zero‐order design (ZOD) of the TLS self‐calibration network—the optimal datum definition. Apart from those two criteria, the noteworthiness of the implied self‐calibrated network is to validate additional criteria of the network—reliability, efficiency and uncertainty—for the entire 18 identified CPs of TLS self‐calibration under the developed parametrisation. The outstanding data collection experiments were conducted using five different products of global manufacturers: Leica ScanStation P50, Leica ScanStation C5, Maptek I‐Sites 8820, Leica RTC360 and Trimble SX10. Accordingly, the robust validation stage considering the accuracy assessment tests was concluded after thematic and thorough data analysis for each sensor.