Research On The Information Fusion Technology For Individual Soldier Based On Heterologue Image

In recent years,with the development of information fusion theory,scholars have further studied the field of image fusion.In the field of heterologous image fusion,the fusion technology of visible and infrared images has also been improved,which lays a technical foundation for the intelligent perception of military information in the future.With the help of this technology,when soldiers are fighting in a complex environment,individual soldiers can obtain battlefield scene details and other information by using visible images,and can also find the outline information of concealed targets by using infrared images.The combination of the two will be beneficial to enhance the perception ability of individual image acquisition equipment for the complex battlefield scene,so as to effectively improve the overall control ability of battlefield environment perception for individual soldier.Based on this,a deep research on heterologous image has been made in this paper.First,the article explores different registration and fusion methods of homologous and heterologous images.On this basis,in view of the existing problems in the field of visible and infrared image registration and fusion,such as large difference in gray scale and poor robustness,this paper proposed improved methods respectively,which improved the quality of visible and infrared image registration and fusion,providing an important theoretical basis for the future information combat.The work of this paper includes image preprocessing,image registration,image fusion,etc.,as follows:(1)The combination of wavelet edge detection and SI-PIIFD is used to reduce the influence of gray difference on image registrationDue to different imaging principles,the original gray values of visible light and infrared images collected by hardware equipment are still quite different even if they are taken for the same scene.Therefore,the registration of heterologous images is more difficult than that of images from homologous sensors.To solve this problem,this paper improves the traditional PIIFD algorithm.Firstly,the conspicuous edge features of source image pairs are extracted by WT.Secondly,Speeded-Up Robust Features(SURF)detector is used to detect the feature points of the edge images,then the feature descriptions should be acquired by scale-invariant partial intensity invariant feature descriptor(SI-PIIFD),which is improved in this paper.Finally,the gaussian field estimator is adopted for feature point matching,and image alignment is accomplished by affine transformation matrix to achieve the registration of visible and infrared images.The simulation experiment verifies that,compared with the traditional PIIFD registration algorithm,the proposed algorithm combining the wavelet edge detection with SI-PIIFD registration greatly improves the correct matching rate and matching precision index,thus verifying the effectiveness and accuracy of the algorithm proposed in this paper.(2)The robustness of fusion algorithm is enhanced through robust principal component analysis(RPCA)and pulse coupled neural network(PCNN)The existing algorithm of PCA combined with 2nd Generation WT has low robustness and fusion performance in the field of image fusion,based on this,a fusion method based on RPCA and PCNN is proposed in this paper to overcome the above problems.The proposed algorithm firstly converts visible and infrared images into high and low frequency signals by the second-generation wavelet transform,and then uses different fusion strategies to fuse low and high frequency signals.For low frequency signals,this paper uses robust principal component analysis to restore the low-rank matrix and uses the weighted average fusion strategy for fusion.In this paper,the high frequency signals are fused into the pulse neural network to obtain the wavelet coefficients after fusion.After that,this paper will carry out the inverse transform to acquire the fused image.Through simulation experiment,compared to the fusion algorithm of traditional PCA combined with lifting wavelet,the proposed algorithm has a significant improvement in spatial frequency,peak signal-to-noise ratio and other indicators.It further verifies that the proposed algorithm can extract the target information more clearly and improve the fusion image quality,thus proving the advanced performance of the proposed algorithm.