夜视环境下红外与可见光图像真彩色快速融合方法研究  

Research on the fast fusion algorithm of of true colour of infrared and visible images under night vision environment

作　　者：谢一博 程进 周顺[1] 侯劲尧 刘卫国[1] XIE Yi-bo;CHENG Jin;ZHOU Shun;HOU Jing-yao;LIU Wei-guo(School of Photoelectric Engineering,Xi′an Technological University,Xi′an 710021,China)

机构地区：[1]西安工业大学光电工程学院,陕西西安710021

出　　处：《激光与红外》2024年第1期136-147,共12页Laser & Infrared

基　　金：陕西省教育厅科研计划重点项目(No.21JY017);陕西省自然科学基础研究计划项目(No.2022JQ-676);中国博士后科学基金项目(No.2022M712493)资助。

摘　　要：为解决红外图像和可见光图像在夜视环境下成像效果不理想的问题,搭建一种由红外相机与RGB相机组合而成的具有结构简单、实时性高等特点的双目异型成像系统。利用双线程方式实时采集红外与可见光图像,针对非局部均值滤波(Non-Local Mean,NLM)提出了一种基于熵的自适应h求解方法,能够较好的消除红外图像的噪声;之后将红外图像与可见光图像的特征点匹配后,引入基于斜率一致性的方法进行图像配准,最后采用改进基础图像融合规则的双尺度融合算法(Two-scale Image Fusion,TIF)以及多种融合算法将红外与可见光图像进行融合,获得目标明显、信息丰富且真实色彩的(Infrared Radiation and RGB,IR-RGB)融合图像。TIF算法能够高效快速地将红外图像与可见光图像融合,在保留了可见光图像对周围环境的真实颜色信息特征的同时,又保留了红外图像温度提取的特点。分析数据可知,TIF算法的IR-RGB图像的熵值提升了约5.14%;其边缘强度均达到最大39.991、22.433、52.880;相比普通的像素级融合规则,该方法的运行速度提升了10倍的量级。该研究在夜视环境下通过红外与可见光的实时成像在目标识别和监测等实际应用中具有非常重要的意义。In this paper,a binocular heterogeneous imaging system with simple structure and high real-time characteristics is built by combining infrared camera and RGB camera in order to solve the problems of unsatisfactory imaging effect of infrared image and visible image in night vision environment.Firstly,the infrared and visible images are acquired in real time using a dual-threaded approach,and an entropy-based adaptive h-solving method is proposed for Non-Local Mean filtering,which is capable of better eliminating the noise of infrared images.Then,the feature points of infrared and visible images are matched,and a slope consistency-based method is introduced for image alignment,and finally images are fused by combining the Two-scale Image Fusion(TIF)algorithm using improved base image fusion rules and multiple fusion algorithms to obtain the Infrared Radiation and RGB(IR-RGB)fusion image with obvious targets,rich information and true to color.The TIF algorithm can efficiently and rapidly fuse infrared images with visible images,preserving the true color information characteristics of visible images for the surrounding environment while retaining the temperature extraction characteristics of infrared images.According to the data analysis,the entropy value of the IR-RGB images of the TIF algorithm is improved by about 5.14%;their edge intensity all reach a maximum of 39.991,22.433,52.880.Compared with the common pixel-level fusion methods,the speed of the proposed method is improved by an order of magnitude of 10 times.The research is of great importance in practical applications such as target identification and monitoring by means of real-time imaging of infrared and visible light in a night vision environment.

关 键 词：图像处理 成像系统 双尺度图像融合 互补信息 目标识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN219[自动化与计算机技术—计算机科学与技术]