Research On Image Fusion Method Based On Visual Imaging Mechanism Of Snakes

Image fusion technology plays an increasingly important role in modern applications and computer vision.It is one of the key technologies in the field of image processing.Infrared and visible image fusion can obtain rich details from visible image and significant objects from infrared image,and generate images with richer information.Therefore,in image fusion technology,infrared and visible image fusion has become an important research direction.According to biological research,some special organs of snakes can complete the reception and processing of infrared information and visible information when they capture prey,so the infrared and visible image fusion method to simulate the visual mechanism of snakes comes into being.Because human eyes can recognize targets faster in color images than gray ones,so far all the fusion methods of snake simulation are to generate false color images.However,the pseudo-color images generated by these methods lack details and have unnatural colors,which sometimes seriously affect the visual effects of images.Therefore,when building a model to simulate the visual mechanism of snakes,this paper carried out relevant research on these problems.The main research work is as follows:(1)The visual imaging mechanism of snakes,receptive field of ganglion cells and the corresponding mathematical model were studied.The classical fusion method based on ON-centered perceptual field is studied,and two improvement methods based on ON-centered perceptual field are proposed for the problems that the pseudo-color fused images generated by this method lose details and have unnatural colors,but seriously affect the visual effect of the images.So that the generated fusion image retains more detail information.The source image is enhanced and fused by center receptive field,and then the fused image is corrected by gamma to retain more details Finally,five groups of source images were used to compare the other eight fusion methods.Four objective evaluation indexes such as subjective evaluation and spatial frequency were compared respectively.Through the analysis of the evaluation indexes,it is found that the integrated performance of the proposed fusion method on the four evaluation indexes is better than other comparison methods.(2)The six dual mode cell imaging mechanisms of rattlesnakes and their mathematical models,as well as the standard and simplified mathematical models of Pulse coupled neural networks(PCNN)were investigated.A new fusion method based on snake dual-mode cells and PCNN model is proposed to address the problem of poor visual effect of images generated by the previously proposed fusion method for simulating snakes.The source image was enhanced by snake dual mode cells and center surround against field,and then further enhanced and fused by PCNN to make the fusion image more natural and clear.Finally,five groups of source images were used to compare with the other five fusion methods,and from the subjective evaluation and four objective evaluation indicators,it was found that the proposed fusion method has better performance and the obtained fusion image is clearer.