BROWSE
Detail
Conference Paper
Joint Radiometric Calibration and Multi-view Matching with Ground Control Points
Cited - time in 
Share 
Share
- Authors
- Changjae Oh, Bumsub Ham, Hyon-gon Choo, Jinwoong Kim, Kwanghoon Sohn
- Issue Date
- 2012-01
- Citation
- International Workshop on Advanced Image Technology (IWAIT) 2012, pp.1-6
- Language
- English
- Type
- Conference Paper
- Abstract
- In this paper, we estimate the corresponding points with images taken at different view point and exposure setting by using the reliable matching points called ground control points (GCPs). In real world, Lambertian model, commonly used assumption in image processing, is not valid, so that it is hard to find the correspondences correctly. There are two approaches for addressing this problem: First, a pre-processing is applied for calibrating color information, followed by finding the matching points. Alternatively, the method robust to color variation is directly used. However, both approaches give unsatisfying results. We propose a method in which calibrated images as well as the matching points are jointly estimated. It can be done by recursively updating the matching points and a camera response function (CRF) which is used to calibrate image color. Since the accuracy of matching points is crucial in radiometric calibration, GCPs are used to calculate the CRF. We demonstrate the performance of our method by comparing the accuracy of a CRF and the matching points with those of conventional methods both quantitatively and qualitatively.
- KSP Keywords
- Camera response function(CRF), Color information, Conventional methods, Image processing(IP), Lambertian model, Multi-view, Pre-processing, Real-world, Recursively updating, View Point, accuracy of matching