Orthorectification and Geometric Quality Assessment of Cartosat-1 for Common Agricultural Policy Monitoring: Final Results

The European Union uses remotely sensed data in a large operational programme to monitor subsidies given to farmers and to identify irregularities in claims. The trend over the last few years has been in sharp increase in the use of very high resolution sensors, with a number of different sensors being used in a complementary manner. For instance, whilst instruments able to provide imagery with a ground sampling distance (GSD) of <1m make up the primary use (with acquisition in 2006 of around 150,000 km2), sensors acquiring data with around a 2m GSD are in general used as a back-up in case of primary instrument acquisition failure. Cartosat-1 test data, which falls into this 2nd category and potentially could provide useful data in the main programme, was imaged in the framework of the Cartosat-1 Scientific Assessment Programme (C-SAP). This paper presents the final results of sensor model and DEM extraction for Cartosat -1 full dataset (two stereo pairs) for a test site located near to Mausanne-les-Alpilles (France), used since 1997 with a time series of reference data for the checking of farmers’ aid applications. The chosen approach was to: First, determine the best method for image rectification and model adjustment; this was verified by undertaking a series of orthorectifications and comparing results with independent check points. Off-the-shelf software (PCI Geomatica 10 and LPS) was used for the evaluation. Second, upon selecting the chosen model creation approach, undertake the extraction of the DEM using the stereo pairs, and proceed with the DEM quality assessment. Moreover, for the assessment of the DEM generation purpose fully automated models were prepared based on GIS techniques. In brief, the full DEM covering the study site is compared with a reference DEM (of higher quality), and assessed according to the land cover and slope categories thus enabling application specific assessment of the DEM generation and potential for image use. Lastly, the quality assessment of SRTM (open source) for the same AOI was prepared and compared with obtained results form Cartosat – 1 DEM extraction process evaluation.