基于SWT和边缘检测的红外与可见光图像融合算法  

Infrared and Visible Image Fusion Algorithm based on SWT and Edge Detection

作　　者：陈钰 郭立强[1,2] CHEN Yu;GUO Li-qiang(School of Computer Science and Technology,Huaiyin Normal University,Huaian Jiangsu 223300,China;Huaian Key Laboratory of Big Data Intelligent Computing and Analysis,Huaiyin Normal University,Huaian Jiangsu 223300,China)

机构地区：[1]淮阴师范学院计算机科学与技术学院,江苏淮安223300 [2]淮阴师范学院淮安市大数据智能计算与分析重点实验室,江苏淮安223300

出　　处：《淮阴师范学院学报（自然科学版）》2023年第3期202-209,共8页Journal of Huaiyin Teachers College;Natural Science Edition

基　　金：江苏省计算机学会教学类专项资金项目(JSCS2022051);江苏省高等教育学会高校实验室研究委员会年度课题(GS2022YB08);江苏省大学生实践创新训练计划项目(202210323014Z)。

摘　　要：针对传统SWT算法融合图像边缘不连续的问题,提出了一种基于SWT和边缘检测的红外与可见光图像融合算法.采用二维平稳小波分别对红外与可见光图像进行变换,得到高低频系数.对高频系数采用交叉双边滤波器进行滤波.根据红外图像和可见光图像不同的成像特点,红外图像采用Canny算子,可见光图像采用Prewitt算子和Roberts算子共同处理,将边缘信息均值融合到低频系数内.采用脉冲点火频率、对比度拉伸分别融合高、低频系数.经过小波反变换得到融合图像.实验结果表明,本算法得到的融合图像边缘较连续,保留的信息量丰富,主观视觉效果较好.In view of the trouble of discontinuous edge fusion of traditional SWT algorithm,a fusion algorithm of infrared and visible images based on SWT and edge detection was proposed.Firstly,two-dimensional stationary wavelet was used to transform the infrared image and visible image respectively to obtain the high and low frequency coefficients.Secondly,the high frequency coefficients are filtered by cross bilateral filter.Thirdly,Canny operator was used for infrared image,and Prewitt operator and Roberts operator are used for visible image according to the different imaging characteristics of infrared image and visible image.The mean edge information was fused into the low frequency coefficient.Finally,high and low frequency coefficients were fused by pulse ignition frequency and contrast stretching.Finally,the final fusion image was obtained by inverse wavelet transform.According to the experimental results,the fusion image obtained by this algorithm has continuous edges,rich information retention and good subjective visual effect.

关 键 词：图像融合 边缘检测 平稳小波变换 交叉双边滤波器 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]