SOME ISSUES OF THE CALCULATION PROCESS OPTIMIZATION IN TERRESTRIAL LASER SCANNERS CALIBRATION

Abstract

A conventional and widely used mathematical model of laser scanners calibration has such an imperfection as correlation between some of parameters during the adjustment process and quite complex numerical solution. A method of avoiding the correlation between scanner and polygon transformation parameters and calibration parameters as well as calibration numerical solution simplification is considered. It is proposed to implement a separate transformation parameters computation by using Kabsch-Umeyama algorithm that can compute Helmert transformation parameters even in case of gross errors present in data set. But direct transformation before adjustment leads to parametric equations and following numerical solution complication. Therefore it is also proposed to conduct a reverse transformation from polygon coordinate system into scanner coordinate system as it simplifies further equations linearization and allows simplifying the numerical solution in general. As it is more than one station to be scanned during the calibration process, one cannot simply swap coordinate systems. The polygon coordinate system remains as original one but reverse Helmert transformation is conducted. To prove this approach is correct, three datasets using different laser scanners and high-precise total station were obtained. Direct and reverse transformation between scanners and total station coordinate systems was made for all datasets and MSE of transformation was calculated for each case. Thus, the possibility of using the reverse transformation to optimize the process of the calibration parameters determining has been experimentally proven.