In personally view, the boresight calibration method from McMullen Nolan Group because of its specific data acquisition method has got a great achievement on the accuracy; and there is not too much could be improved for the boresight errors only to achieve the standard accuracy of MNG i.e. ±15mm absolute horizontal 2σ 95% which satisfies the accuracy requirements for most projects. The improvement that could do is to develop a new algorithm in the purpose of achieving automatic adjustment of boresight values. That will not only reduce the time waste by the manual adjustment (to be more efficient) but also minimize the risk of human errors (which may slightly enhance the accuracy). The new algorithm can reference to the planar adjustment method by Riegl. This thesis had an attempt on this improvement but do not get any results due to the limitation of the software.
In contrast, the self-calibration is quite successful with MNG Riegl VQ250 experiment. The scanner is proved that it had internal errors (average 25mm over 20m) and the concept of the design calibration method is proved that it is feasible locally. However, these errors are not constant and may relate to the speed of the scanner rotating rate. The parameters of error function have not been solved. Thus, the correction to the measurement couldn’t be determined. The further improvement only can be done once the factory test being released. Otherwise, numerous tests need to be processed with distinct control parameters to verify if it is one relating factor. If the error is random, there will be no reliable method for applying an appropriate correction. Then the ideal situation is the manufacture can service locally.