Determination and Monitoring of Boresight Misalignment Angles during the HyMap Campaigns HyEurope 2003 and HyEurope 2004

Abstract

High geometric accuracy of airborne hyperspectral image data is essential for multi-temporal studies, the generation of mosaics and the comparison and combination with other georeferenced data. Direct georeferencing of airborne line scanner data (e.g. HyMap data) implies the combination of kinematic GPS-positioning and an inertial measurement unit (IMU). With these orientation observations the exterior orientation of the sensor is available for each moment of the data acquisition. For the exact determination of the sensor’s exterior orientation, the so-called boresight misalignment angles have to be known, which describe the angular discrepancies between the sensor and the IMU coordinate system. They can be determined using ground control points (GCPs). If the misalignment proves to be stable for each mounting of the sensor, it is possible to rectify the data without the usage of extra GCPs for every image strip. This would reduce the expenditure of time and simplify the process of georefencing. The misalignment angles are considered within the ortho-rectification software ORTHO developed by DLR, which is an essential part of the automatic processor of the future ARES (Airborne Reflective Emmisive Spectrometer) sensor. Within this study, the long term stability of boresight misalignment angles has been investigated during the HyMap campaigns HyEurope 2003 and HyEurope 2004, where the mounting of the HyMap sensor remained unchanged for over 2 months. Several data takes over a test field around Oberpfaffenhofen with approximately 30 GCPs have been carried out to calculate the misalignment angles. The results were compared to each other and applied to different sets of HyMap data to investigate the stability and the specification of the accuracy of the boresight misalignment angles to fulfill the requested sub-pixel accuracy of the rectified image. Furthermore the study allowed an evaluation of the quality of the GCPs within the test field. The result was a significant improvement of the existing GCPs in addition with the definition and surveying of new GCPs.