Precise Point Positioning: Evaluation and Airborne Lidar Calibration

This report presents the results of investigations to determine accurate positions of aircrafts in airborne surveys (airborne gravity and airborne lidar) using precise point positioning, and also introduces a new so called “stepwise geometric misalignment determination” method to retrieve the airborne lidar system misalignment angle by automating the matching of lidar data with ground truth. Kinematic GPS positioning has been widely used, but the available commercial software systems are normally only suitable for the short or medium range kinematic baseline. However, in polar areas, airborne surveys have baselines ranging from a few hundred kilometers to even more than one thousand kilometers due to logistic limitations. It is a challenge to the traditional kinematic GPS software based on double differenced models, such as GPSurvey or GrafNav. Since Zumberge demonstrated the perfect performance of point positioning for kinematic applications, the precise point positioning attracted a lot of attention and opened a new alternative door to kinematic positioning. In this report different tests have been done to evaluate the ability and accuracy of the software TriP in the kinematic and static case by using internal consistency (residuals, RMS, repeatability etc.), known coordinates, ground truth and double-differenced solutions. The kinematic GPS positioning accuracy using four different software systems has been investigated and tested by comparing the degree of agreement between ground truth and the height of airborne lidar footprints derived from combining flight trajectory, orientation and lidar range. The conclusion is that the TriP software is robust and reliable, and that TriP runs much faster (10 times) than GPSurvey 2.30. A static positioning accuracy of mm to cm level could be achievable depending on the observation session length, and kinematic positioning accuracy can reach cm to dm level. Furthermore, a new method for airborne lidar system misalignment calibration was described in detail. The proposed method was a so called ‘stepwise geometric misalignment determination’ based on the relationship between the point clouds on regular objects (e.g. flat top buildings) and the ground truth of the objects used for calibration. In order to extract the footprints on the objects, filtering was implemented before the calibration. Three example tests have been made and verified that the proposed method is feasible and effective.