基于主元分析的相机间颜色校正

Principal component analysis based color calibration between cameras

针对相机间的颜色校正问题,提出一种新方案以克服因相机所处环境不同而对校正结果带来的影响。利用主元分析确定颜色空间变换关系矩阵,将RGB颜色空间转换到各颜色通道间相关性较小的Lαβ空间,在该颜色空间下,利用图像的颜色统计特性一致性原理对相机间的颜色进行校正。实验结果表明,新方案可有效克服相机所处环境对目标颜色校正的影响,实现较为理想的相机间颜色校正。

For color calibration between cameras,a new method is proposed to overcome the influence of the environments.Principal component analysis is applied to obtain the color space transform matrix.Then the RGB color space is transformed to Lαβ space in which each channel has little relevance,and in this space,color statistical features are used for color calibration between cameras.Experimental results show that the proposed method could overcome the influence of camera environments and the calibration results are satisfied.