Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal CAL1362

TitleObject and Change Detection with fused Optical and Radar data based on high resolution Digital Surface Model
Investigator Reulke, Ralf - HU-Berlin / DLR, Computer Vision / DLR-OS
Team Member
Dipl.-Inf. Rueß, Dominik - DLR, RM-OS
SummaryOptical and Radar data fusion is a widely used approach for remote sensing applications. Due to different observation conditions and imaging geometry, co-registration and data interpretation in highly undulated regions is a challenging task. A significant improvement is expected when high resolution DSMs are introduced to the fusion of both sensor types. Both Radar and optical data offer advantages to many remote sensing fields. Due to cloud penetration and night coverage images from active Radar systems can be acquired much faster than data from passive optical sensors. Some Radar wavelengths can also penetrate the ground to a certain extent. However, better human interpretability and multi spectral analysis are provided by means of optical images. The high availability makes Radar naturally predisposed for the task of change detection, which is usually based on a combination of data, acquired at different points in time. Also it is evident that the different orbits or flight trajectories lead to different viewing conditions and to different occlusions as well as ground projections in the combined data from scenes with significant height differences. Additionally, the increasing resolution of today’s Radar and optical sensors introduces different projection errors in the image for drastic local height changes from objects like buildings or trees. The fusion of these data sets with high resolution DSMs and known sensor orientation for both Radar and optical images will lead to a significant improvement of the derived data products and information. With DSMs viewing conditions and imaging geometry from each data set can be corrected and occluded regions can be detected in a geo-rectified frame. As a result, co-registration can be improved and a refinement of the DSM with 3D approaches is possible. In the proposed project, the benefits of the fusion of optical and Radar data with high resolution DSM should be investigated, related to change detection in urban and rural areas. It is assumed that there is a large number of optical and SAR data of the same scene available, which were acquired at various points in time. Quite often for certain events (such as disasters) one only has Radar data for evaluation. It is the question to be examined, what changes can be detected in urban areas with Radar data only or what is the information loss in comparison to the use of optical data.

Back to list of proposals

© DLR 2004-2016