PRE-FLIGHT AND IN-FLIGHT GEOMETRIC CALIBRATION OF SPOT5 HRG AND HRS IMAGES

SPOT5 Location model calibration addresses five main issues. The first issue is to get best relative and absolute location performances. It consists of relative orientation calibration for HRG, HRS and stellar location unit reference frames. Such a calibration started in-house, using theodolites; it ends in-flight using GCPs. The second issue is to get a model of THR pairs relative shifts good enough to deliver the best 2.5m sampled image. The first ever, true HRG images have been acquired during satellite design, a few months before launch. Such images contributed to THR processing validation and allowed ground calibration of THR detection lines relative shifts, way before launch. In-flight measures confirmed that such ground measures are reliable. Third issue is to turn HRS stereo pairs parallax in a precise enough altitude estimate. That means that HRS location models have to include an accurate model of objectives distortion. A fourth issue consists of HRG’s steering mirror mechanism calibration, in order to get the same location performance, whatever the HRG mirror viewing angle is. Final issue deals with optimisation of time delay between two HRG off nadir images acquisition. Such time delay depends on mirror damping speed. For a given viewing angle, called "Autotest", one can acquire HRG images of a designed pattern located in the focal plane. A straightforward processing of this kind of images indicates if the mirror command can be improved. This new calibration process is available in flight as well as on ground. After the two months commissioning phase, objectives are met: images location models are accurate enough to meet all requirements. Absolute location performances provided by the stellar unit are such that new phenomena, never seen on initial SPOT satellites, appear. Initial results show performances trends that we should be able to model, once confirmed through further images acquisition. We specially think about orbital and/or seasonal effects. Work is still carried on. This paper deals with geometric calibration of SPOT5 location model. Focus is set on commissioning phase results for which calibration has been performed in-flight as well as on ground. That is, relative reference frames orientation, CCD line distortion of HRG and HRS cameras, THR processing and tuning of steering mirror damping duration. To start with, I give a brief on main improvements of SPOT5 location model. 1. FROM SPOT4 TO SPOT5 SPOT5 payloads offers several improvements. Images sampling step can be reduced up to 2.5m thanks to a complete refurbishing of the focal plane. A stellar unit participates in minimisation of the absolute location performance. Two new cameras have been designed in order to provide on track stereo pairs. Following paragraphs focus on the main stakes of location model calibration. 1.1 Enhanced Focal plane Technology improvement allows production of CCD array whose detector size is one fourth of today's SPOT1-4 detector and which numbers is increased almost by ten, on a single line: from 1700 to 12000. As a result, just changing CCD lines leads to the acquisition of 5m sampled images, still covering a 60 km swath. Studies on optimisation of an image chain (Latry 2000), applied to HRV-HRG telescope, shown a way to improve images resolution combining two CCD lines information’s. A quincunx super-sampling principle called "Supermode" appears as a general optimisation of MTF, noise and sampling given the HRG telescope cut-off frequency (Latry 1998). THOMSON manufactured a dedicated specific sensor comprising two CCD lines, with a 3.45 pixel line shift and a 0.5 column shift. To achieve such a regular sampling on the ground, yaw steering and adaptive time delay in between lines acquisition had to be implemented. Yaw steering compensates earth rotation. Time delay compensates lines shifts, especially when off nadir viewing. 1.2 Improved Attitude restitution New challenge given to SPOT5 payload is to provide accurate location model. Accuracy of SPOT location model mainly depends on attitude restitution performances. For SPOT1 to SPOT4, satellite attitude provided with images is just the same as the one computed on board, for platform earth pointing control. Attitude Control System uses gyro meters, to provide

[1]  Christophe Latry,et al.  SPOT5 THR mode , 1998, Optics + Photonics.

[2]  Aime Meygret,et al.  In-flight assessment of SPOT5 image quality , 2003, SPIE Remote Sensing.

[3]  C. Lautry,et al.  Optimized sampling for CCD instruments: the Supermode scheme , 2000, IGARSS 2000. IEEE 2000 International Geoscience and Remote Sensing Symposium. Taking the Pulse of the Planet: The Role of Remote Sensing in Managing the Environment. Proceedings (Cat. No.00CH37120).

[4]  T. Westin Inflight calibration of SPOT CCD detector geometry , 1992 .