二维变分模态分解矿井监控视频图像去噪  

TWO-DIMENSIONAL VARIATIONAL MODE DECOMPOSITION FOR MINE MONITORING VIDEO IMAGE DENOISING

作　　者：闫洪波[1] 赵蓬勃 刘恩佐 刘霈 Yan Hongbo;Zhao Pengbo;Liu Enzuo;Liu Pei(School of Mechanical Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,Inner Mongolia,China)

机构地区：[1]内蒙古科技大学机械工程学院,内蒙古包头014010

出　　处：《计算机应用与软件》2023年第6期211-215,共5页Computer Applications and Software

基　　金：科技部创新方法工作专项资助项目(2017M010660-03)。

摘　　要：煤矿监控已基本覆盖到各矿井下的安全生产中,但矿井下复杂且恶劣的环境严重影响图像的质量,受到大量的椒盐噪声污染,降低了保障安全和消除隐患的能力。为此提出一种变换域图像去噪方法,将图像使用二维变分模态基于不同频率分解成低频和高频子模态图像,低频和高频图像中包含着不同的图像信息,对低频图像采用各向异性扩散滤波进行保边去噪处理,使用自适应中值滤波对高频图像中的大量噪声进行降噪,重构各子模态图像。以实际监控图像为研究对象,在不同噪声环境下进行对比实验,结果表明,与其他方法相比该方法具有优异的去噪效果,并且还能保留原始图像的边缘结构信息。Coal mine monitoring has basically covered the safety production in all mines.However,the complex and harsh environment in the mine seriously affects the quality of the image,which suffers a lot of salt and pepper noise pollution and reduces the ability to ensure safety and eliminate hidden dangers.Therefore,a method of image denoising in transform domain is proposed.The image was decomposed into low-frequency and high-frequency sub-mode images based on different frequencies using a two-dimensional variational mode.Low-frequency and high-frequency images contained different image information.Anisotropic diffusion filtering was used for low-frequency images for edge-preserving and denoising processing.Adaptive median filtering was used to reduce a lot of noise in high-frequency images.Each sub-mode image was reconstructed.The actual monitoring image was taken as the research object,and comparative experiments were conducted under different noise environments.The results show that this method has excellent denoising effect compared with other methods,and it can retain the edge structure information of the original image.

关 键 词：矿井监控视频图像 二维变分模态分解 图像去噪 各向异性扩散滤波 中值滤波 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]