基于梯度的多曝光图像融合

Gradient-based Fusion of Multi-exposure Images

针对运动场景拍摄的不同曝光度的图像序列,提出基于梯度的多曝光图像融合方法,将不同曝光度的图像细节呈现到同一张普通图像上。该方法是直接加权融合的方法,先用梯度信息来定义权值,然后用多分辨率加权融合。实验表明,本文的方法取得较好实验结果,消除运动物体对图像融合产生的影响,同时增强了图像的细节表现能力,具有较好的实用价值。

This paper proposes a gradient-based technique of multi-exposure images fusion for the image sequence captured at the motion scene,which makes all images＇ detail visible in one image.This method directly fuses different images.First of all,defines weight with the gradient information,and then fuses all images seamlessly with the multi-resolution spline scheme.Experiments show that these methods can deal with the problem of multi-exposure images fusion,eliminate the impact of moving objects,and enhance the image detail.