基于亮度一致性的多曝光图像融合

Multi-Exposure Image Fusion Based on Luminance Consistency

针对传统的多曝光图像融合算法存在局部亮度与对应场景不一致的现象,提出一种亮度一致性的多曝光图像融合算法。该算法通过分析多曝光图像序列,提取局部对比度、结构信息、曝光亮度信息三种图像特征构造权值图,一方面保留更多场景细节信息和场景结构信息,另一方面保持融合图像局部亮度与对应场景亮度一致。选取8组多曝光图像序列实验,与另外7种经典多曝光图像融合方法对比分析。实验结果表明:该方法得到的融合图像细节丰富,结构相似性、亮度一致性良好,呈现出自然生动的视觉效果。

A new multi-exposure image fusion algorithm is proposed to address the problem of the local luminance is inconsistent with the corresponding scene in traditional multi-exposure images.By analyzing the multi-exposure image sequence,the algorithm decomposes each image into three conceptually independent components:local contrast,structure information and exposure luminance information.On the one hand,it retains more details of the scene information and scene structure information;on the other hand,it keeps the fused image local luminance consistent with the corresponding scene luminance.Eight sets of multi-exposure image sequence were tested in the experiments,and the result was compared with other seven kinds of traditional methods.The experimental results demonstrate that the fusion image is rich in detail with good structural similarity and luminance consistency,showing a natural and vivid visual effect.