Research On Image Fusion Method Based On Rattlesnake Vision Imaging System

Image fusion technology is an important technology in the field of digital image processing.In a complex environment,the fusion image of infrared image and visible image can present the target with larger infrared radiation in the infrared image and the environmental details in the visible image theoretically.It is one of the research directions of image fusion technology with broad application prospects.Biological studies have found that rattlesnakes can receive and process infrared information and visible light information at the same time.In response to this discovery,the paper combined with the National Natural Science Foundation of China Youth Fund project " Application and Mathematical Modeling for the Vision System of Rattlesnake"(No.61201368)to carry out the research of "Image Fusion Method Based on Rattlesnake Vision Imaging System".Focusing on biological vision receptive field and its mathematical model,Waxman's typical fusion structure,rattlesnake's six bimodal cells fusion mechanism,and image preprocessing,and carrying out related research work.The main research contents are as follows:Based on the research of the biological vision receptive field and its mathematical model and Waxman's typical fusion structure,an improved image fusion model based on the center-surround confrontation domain is given in this paper.Aiming at the problem of unnatural colors and insufficient environmental details of the fusion image obtained by Waxman's typical fusion structure,the characteristics of the images generated by each node of Waxman's typical fusion structure are studied,and the infrared enhanced visible image,infrared suppressed visible image and visible light enhancement finally obtained by the fusion structure are studied.The corresponding mapping channel of the image is modified to obtain a fusion image with more natural colors and more environmental details.Aiming at the typical fusion structure of Waxman and the fusion image obtained by the fusion model after modifying the mapping channel in this paper,the problem that the infrared target is highlighted in the infrared source image is ignored.The infrared enhanced visible light mechanism and the visible light enhanced infrared mechanism are studied,and the two fusion signals are managed by linearly OR.To further enhance the infrared information,so that the infrared target is better presented in the fusion image.For the convenience of distinguishing the above two improved fusion models,they are respectively recorded as: “this article's improved model one” and “this article's improved model two”.Two improved models improve the color of the fusion image and the effect of infrared target presentation.In the experimental results of the selected ten image pairs,the proportion of the best value of the standard deviation is80% and 20% respectively,and the proportion of the best value of the average gradient is 40% and 50% respectively,the best value of entropy accounts for 80% and10% separately,and the best value of spatial frequency accounts for 20% and 60%separately.The overall performance on the four fusion image quality evaluation indicators is more excellent than comparison experiments.Compared between the two improved models,“this article's improved model one” performs better in the standard deviation and entropy evaluation indicators,and “this article's improved model two”performs better in the average gradient and spatial frequency evaluation indicators.Based on the research of the six bimodal cell fusion mechanisms and image preprocessing of rattlesnakes,an image fusion model based on rattlesnake vision imaging system is given.Aiming at the problem that the typical fusion structure of Waxman only simulates two cell fusion mechanisms in the six bimodal cells of rattlesnakes,and the use of center-surround adversarial domains for simulation and calculation is complicated.The fusion mechanism of the six bimodal cells of rattlesnakes is studied and the corresponding are given.The mathematical model provides an improved image fusion model to effectively use the six bimodal cell fusion mechanisms of rattlesnake to obtain a fusion image that better presents the environment and scenery information.Aiming at the problem that the infrared source image and the visible light source image are not clear enough to retain and focus on the image information during the fusion process,the image preprocessing is studied,and the source image is processed by extracting the common and unique information of the infrared and visible images,to obtain more individual characteristics of infrared image and visible image.And improving the quality of the fusion image.At the same time,the mathematical model of the visual receptive field is simplified to reduce the amount of data calculation and reduce the complexity of the model.For the convenience of distinguishing the above two improved models,they are respectively recorded as: “this article's improved model three” and “this article's improved model four”.The two improved models have good performance in fusion image detail information and color richness,and make better use of source image information.Among the experimental results of the selected ten image pairs,the best value of the standard deviation accounts for 10% and 80% respectively,the best values of average gradient accounted for 20% and 80% respectively,the best values of information entropy accounted for 50% and 50% separately,and the best values of spatial frequency accounted for 20% and 80% separately.The overall performance of the four fusion image quality evaluation indicators is better than the comparison experiment.On the whole,the performance of “this article's improved model four” is generally better than “this article's improved model three” on the four evaluation indicators.On the basis of Waxman's typical fusion structure,six bimodal cell mathematical models of rattlesnake were introduced,and an image fusion structure was built.The infrared image and visible image were enhanced through image preprocessing,and the mathematical model of visual receptive field was simplified to reduce computational complexity.Image fusion model based on vision imaging system is given.Through the evaluation index data of standard deviation,average gradient,entropy and spatial frequency,it is verified that the performance of the proposed method has been improved.