Research On The Fusion Algorithms Of Infrared And Visible Image

Image fusion is an important branch of information fusion technology.It is mainly divided into two types: the fusion of images obtained by the same sensor at different moments and the fusion of images obtained by multiple sensors.And the latter is also called multi-source image fusion.With the continuous development of technology,complex engineering applications require more comprehensive information about the related scenes to enhance the understanding of the images.While the image captured by a single sensor expresses the information of the scene from only a certain aspect,and it usually cannot meet the engineering requirements.Therefore,in the field of computer vision,multi-source image fusion technology has attracted more and more attention from scholars from all walks of life.Infrared and visible image fusion technology is an important application in this field.The imaging principle of infrared and visible images is different,so that the information carried by the two is different.According to their respective characteristics,through the comprehensive utilization,the advantages of the two images are complementary,and the two images are merged into an image that is favorable for the accurate judgment of the relationship between the target and the environment.Infrared and visible image fusion technology plays a prominent role in target detection and localization,and has specific applications in target recognition,remote sensing measurement,battlefield sensing and a range of specific environmental monitoring.This paper mainly focuses on the fusion of infrared image and the visible image in RGB.The specific requirement is to highlight the high-radiation targets in the infrared image and improve the background detail expression ability of the fused image under the premise of retaining the color information in the visible image,which is beneficial to the judgment of the relationship between the target and the environment.In response to this requirement,this paper first introduces the infrared and visible image preprocessing work,and analyzes the advantages and disadvantages of several typical fusion algorithms.Then,according to the characteristics of the infrared and visible images,it is proposed to register the two images by using the edges of the two.For the extraction of edges,this paper introduces an edge detection algorithm based on morphological.For image registration,this paper analyzes and compares the image registration algorithm based on SURF feature points and the image registration algorithm based on maximizing correlation coefficient.Experiments show that the two images obtained by the latter have higher similarity,and the registration effect is more excellent.Thirdly,according to the analysis of typical fusion algorithm,this paper introduces the average gradient into the fusion rule,and proposes a new fusion algorithm based on HIS transform and wavelet transform.The algorithm performs HIS transform on infrared and visible images respectively,and obtains two sets of intensity(I),hue(H)and saturation(S)components.And the fusion based on wavelet transform is applied to the intensity(I)component to obtain a new I component.It performs HIS inverse transform with the hue(H)and saturation(S)components of the visible image to obtain a desired fused image.Finally,a contrast experiment of infrared and visible image fusion is carried out.And the fusion images obtained by the algorithm proposed in this paper and several existing algorithms are compared and analyzed.Experiments show that the fusion algorithm proposed in this paper is clearer than the fusion image obtained by other algorithms and the fusion effect is better.And the infrared and visible image fusion software was developed using this algorithm.