Translator Disclaimer
14 August 2003 Robust ground control point-matching algorithm for multichannel radar image geometric correction
Author Affiliations +
A new algorithm for multichannel radar image mutual superimposing (registration) and geometric correction is proposed. It uses ground control points (GCPs) selected manually or automatically for warp and base images and the branch-and-bound technique for initial estimation of geometric transform parameters. We show that due to taking into account the GCP selection errors for both warp and base images, the proposed modification of branch-and-bound algorithm allows to additionally improve the accuracy of warp parameter estimation. To reduce the computational complexity of branch-and-bound algorithm an approach based on optimal selection of center position for warp image coordinate system is proposed. It is shown that the coordinate system shifting towards the center of warp image GCP bounding box permits to accelerate the search process. For additional speeding up the branch-and-bound algorithm, bounded alignment approach is applied. For establishing the correspondence between GCPs and for detection of erroneous points, a novel method of statistical analysis of GCP superimposing errors is put forward. The designed algorithm is tested for both simulated and real data demonstrates high efficiency and robustness.
© (2003) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Gennady P. Kulemin, Andrei A. Kurekin, Alexander A. Zelensky, and Vladimir V. Lukin "Robust ground control point-matching algorithm for multichannel radar image geometric correction", Proc. SPIE 5097, Geo-Spatial and Temporal Image and Data Exploitation III, (14 August 2003);

Back to Top