论文标题
考虑多曝光图像融合的非线性摄像机响应的色相纠正方案
Hue-Correction Scheme Considering Non-Linear Camera Response for Multi-Exposure Image Fusion
论文作者
论文摘要
我们提出了一种用于多曝光图像融合（MEF）的新型HUE校正方案。到目前为止，已经研究了各种MEF方法来生成更高质量的图像。但是，由于缺乏正确色调的参考图像，因此很少有MEF方法考虑到色调失真，而与其他图像处理不同。在提出的方案中，我们从输入多曝光图像中生成HDR图像作为色相校正的参考。之后，根据RGB颜色空间中的恒定平面，通过使用HDR One的色调信息来删除由MEF方法融合的图像中的色相失真。在模拟中，提出的方案被证明可以有效纠正由常规MEF方法引起的色相延伸。实验结果还表明，所提出的方案可以产生高质量的图像，而不论输入多曝光图像的暴露条件如何。
We propose a novel hue-correction scheme for multi-exposure image fusion (MEF). Various MEF methods have so far been studied to generate higher-quality images. However, there are few MEF methods considering hue distortion unlike other fields of image processing, due to a lack of a reference image that has correct hue. In the proposed scheme, we generate an HDR image as a reference for hue correction, from input multi-exposure images. After that, hue distortion in images fused by an MEF method is removed by using hue information of the HDR one, on the basis of the constant-hue plane in the RGB color space. In simulations, the proposed scheme is demonstrated to be effective to correct hue-distortion caused by conventional MEF methods. Experimental results also show that the proposed scheme can generate high-quality images, regardless of exposure conditions of input multi-exposure images.
