基于滤波算法的曝光图像多特征细节识别仿真  

Simulation of Multi-Feature Detail Recognition of Exposed Image Based on Filtering Algorithm

作　　者：李振峰[1] 高丽杰[2] 张太行[3] LI Zhen-feng;GAO Li-jie;ZHANG Tai-hang(Information Center,Zhengzhou University of Science and Technology,Zhengzhou Henan 450064,China;Department of Information Engineering,Zhengzhou University of Science and Technology,Zhengzhou Henan 450064,China;Information Office,Henan University of Chinese Medicine,Zhengzhou Henan 450012,China)

机构地区：[1]郑州科技学院信息化建设与管理中心,河南郑州450064 [2]郑州科技学院信息工程学院,河南郑州450064 [3]河南中医药大学信息化办公室,河南郑州450012

出　　处：《计算机仿真》2023年第12期242-246,395,共6页Computer Simulation

基　　金：赛尔网络下一代互联网技术创新项目(NGII20170318);科技公关项目(182102210552);[重点科研项目]河南省高等学校重点科研项目(教科技(2020)348号)(21A630035)。

摘　　要：为解决曝光图像信息传递效果差的问题,提出对多序列曝光图像进行垂直方向切片后卷积的方法,然后采用Anscombe变换与图像滤波算法结合,去除曝光图像中的除泊松噪声增强曝光图像的边缘与细节信息。仿真结果表明,在增强曝光图像多特征信息能力上,双边滤波较高斯滤波效果更明显,平均梯度提高了10.2%。基于滤波的多曝光图像特征融合重构可以有效的提高曝光图像的细节信息,提高图像传递质量。In order to solve the problem of poor information transmission of exposure images,this paper proposes a convolution method for multi-sequence exposure images after slicing in the vertical direction,and then uses the Anscombe transform and image filtering algorithm to remove Poisson noise from the exposure images and enhance the edge and detail information of the exposure images.The simulation results show that bilateral filtering is more effective than Gaussian filtering in enhancing the multi-feature information ability of the exposed image,and the average gra⁃dient is increased by 10.2%.Multi-exposure image feature fusion and reconstruction based on filtering can effectively improve the details of the exposed image and improve the quality of image transmission.

关 键 词：多曝光图像融合 图像滤波 图像重构仿真 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]