Semiautomatic color checker detection in distorted images
暂无分享,去创建一个
In order to map the color space of a raw digital image into a certain rendering space such as sRGB or Adobe RGB, which are commonly used in consumer digital cameras, a calibrated color checker (CC) target is required. Common targets are the GretagMacbeth ColorChecker and the Digital ColorChecker SG. In order to reproduce the scene colors at a certain illumination condition, the target is to be detected in the image and its average color tristimulus values are to be extracted out of the checker fields. With the proper field colors in a certain reference space, finally, a color mapping can be determined. We propose a novel semi-automatic method for the target detection and color extraction process. We consider projective and polynomial geometry and hence, the target can be in any alignment in the scene. In our experiments we show the robustness of our method, even on radially distorted images. Finally, we apply color correction using polynomial least-squares fit in sRGB domain.
[1] Andrew Zisserman,et al. Multiple view geometry in computer visiond , 2001 .
[2] Roger Y. Tsai,et al. A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf TV cameras and lenses , 1987, IEEE J. Robotics Autom..
[3] Mark S. Drew,et al. Constrained least-squares regression in color spaces , 1997, J. Electronic Imaging.
[4] N. Otsu. A threshold selection method from gray level histograms , 1979 .