Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal MTH0299

TitleSpectral characteristics preserving image fusion of TerraSAR-X and optical data
Investigator Ehlers, Manfred - University of Osnabrück, Institute for Geoinformatics and Remote Sensing (IGF)
Team Member
Dr. Rosso, Pablo - University Osnabrück, Institute for Geoinformatics and Remote Sensing (IGF)
Dipl. Umweltwiss. Klonus, Sascha - University of Osnabrück, Institute for Geoinformatics and Remote Sensing (IGF)
SummaryThe objective of the project is to use TerraSAR-X data to improve the spatial resolution of multispectral data from different optical remote sensing sensors. To this end, we will make use of the newly developed Ehlers Fusion approach, which has already been proven to perform better than the standard pan sharpening techniques such as IHS, PC, Brovey and multiplicative fusion methods (Ehlers 2004 and 2005, Klonus 2005 and 2006) and shows excellent results for the Fusion of data from different optical sensors (Ehlers & Klonus 2004, Klonus & Astrand 2005, Klonus & Ehlers 2007, Ling et al. 2007). In this project we will investigate whether this method can be used to fuse optical and radar data. Additionally we will evaluate to what extent this method results in an improvement of some automated interpretation methods such as segmentation, classification or change detection. To reach these objectives several steps are necessary. First the data will be prepared and the test areas chosen. The choice of the test areas depends on the availability of TerraSAR-X data, (several possible areas were already mentioned). The optical data sets for these areas are already available. Next the Ehlers Fusion will be modified to integrate radar data with optical data. The success of the integration must be validated. For this, the multispectral data will be fused with the TerraSAR-X data using the Ehlers Fusion method. The results will be compared with other standard fusion techniques such as Brovey Principal Component, and with more developed fusion techniques such as Gram-Schmidt, UNB, wavelet based fusion and CN-Spectral Sharpening. The evaluation will be based on the verification of the preservation of spectral characteristics and the improvement of the spatial resolution. The preservation of the spectral characteristics will be assessed through correlation coefficients, RMSE (Wald 2002), differences between the original and the fused images and the Structural Similarity Index (SSIM). To assess the improvement of the spatial resolution, methods to be considered are: the correlation coefficients between high pass filtered radar data and the fused image, edge detection methods, and the analysis of the modular transfer function (MTF). This method needs to be developed. After this the possibility to improve automated image processing methods such as segmentation, classification and change detection will be investigated. Fused images will be analysed with standard techniques for accuracy assessment and computer aided photo interpretation (CAPI). A software will be developed to automate the whole process, including the fusion, the automated classification and change detection. The software will be designed in a way that would allow the user to use this tool without any background information. Finally, a performance analysis will be carried out to evaluate user friendliness, ease of integration into other software products, speed and memory requirements. The project is funded by the German Ministry of Finance and Technology (BMBF) through the German Aerospace Center (DLR) with the promotional reference 50EE0704.

Back to list of proposals

© DLR 2004-2016