低频重分布与边缘增强的红外图像增强算法  

Infrared image enhancement algorithm based on low frequency redistribution and edge enhancement

作　　者：邓春华 周勇[2] DENG Chun-hua;ZHOU Yong(School of Software and Blockchain,Jiangxi University of Applied Science,Nanchang 330100,China;School of Computer Information Engineering,Jiangxi Normal University,Nanchang 330027,China)

机构地区：[1]江西应用科技学院软件与区块链学院,江西南昌330100 [2]江西师范大学计算机信息工程学院,江西南昌330027

出　　处：《激光与红外》2023年第1期146-152,共7页Laser & Infrared

基　　金：国家自然科学基金(No.61562063);江西省教育科学“十四五”规划项目(No.21YB286)资助。

摘　　要：针对现有的红外图像增强方法存在欠增强、过增强以及微小细节丢失等缺陷,提出了低频重分布与边缘增强的红外图像增强算法。用基于改进引导滤波的Retinex将红外图像分解为低频和高频图像。为了充分利用像素级的动态空间,对低频图像进行均匀重分布,以提升图像的亮度和清晰度;用提出的方向梯度算子对高频图像进行边缘提取,再对高频图像进行边缘增强,进一步提升图像的对比度。将经增强处理的低频和高频图像作Retinex反变换,得到增强效果的红外图像。实验结果显示,相对于部分现有方法,本文方法的增强图像的信息熵和Brenner指数更高,而NIQE指数更小,因此本文方法能更有效地提升红外图像的对比度,在增强图像纹理细节的同时更好地保持图像的自然度。To address the shortcomings of existing infrared image enhancement algorithms such as under-enhancement, over-enhancement and loss of small details, an infrared image enhancement algorithm based on low frequency redistribution and edge enhancement is proposed. The infrared image is decomposed into low-frequency and high-frequency images by Retinex based on improved guided filtering. In order to take full advantage of the dynamic space of pixel level, the low-frequency image is uniformly redistributed to improve the brightness and clarity of the image;and the proposed directional gradient operator is used to extract the edge of the high-frequency image, with which the edge of the high-frequency image is enhanced, so as to further improve the contrast of the image. The enhanced infrared image is obtained by the Retinex inverse transformation of the enhanced low-frequency and high-frequency images. The experimental results show that compared with some existing methods, the information entropy and Brenner index of the enhanced image are higher, and the NIQE index is smaller, so the method is more effective in improving the contrast of infrared image and better enhancing the texture details of the image while maintaining the naturalness of the image.

关 键 词：引导滤波 红外图像增强 低频重分布 边缘增强 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]