高动态范围图像原始数据在获取和合成过程中会出现异常值, 进而影响色调映射后的低动态范围图像显示效果。为了解决此问题, 提出了一种基于图像引导滤波的图像色貌模型(iCAM06)色调映射算法。该算法使用具有较强平滑作用的引导滤波代替双边滤波, 将输入数据分成细节层和基本层, 分别对细节层和基本层进行处理。为了使映射后的图像更加符合人眼感知效果, 对色适应矩阵CAT02进行了调整。实验结果表明, 与相关算法相比较, 该算法减少了由于数据异常造成的显著噪声点, 避免了伪影和偏色现象, 使细节得到了较好的保持。信息熵、方差和清晰度等客观指标也证明了该算法的有效性。

During the acquisition and synthesis procedures for a high dynamic range image, outlier points may occur and affect the display performance of the tone mapped image. To solve this problem, an improved image color appearance model(iCAM06) tone mapping algorithm based on image guide filtering is proposed. The algorithm uses the guided filtering with strong smoothing effect instead of the bilateral filtering to convert the input data into detail layer and base layer, which will be dealt with different strategy respectively. To make tone mapped images more similar to human eye perception, an modified chromatic adaptation matrix CAT02 is used. The results show that, compared with the related algorithms, the proposed algorithm reduces the noise caused by abnormal data significantly, avoids the artifact and partial color phenomena, and preserves more details. Information entropy, variance and the definition of objective indicators also show the effectiveness of the proposed algorithm.