A Research On Dynamic Multi-Focus Image Fusion Technology

Image fusion refers to the multiple sensors gather different information at different perspectives or at different times for the same target,and generates a new image with more information.Image fusion can meet people's different needs of image information,improve the system to identify the accuracy of the target recognition.The dynamic multi-focus image fusion is an important component of image fusion technology,it is proposed to solve the problem that optical sensors with different focal length in the same scene can't generate clear dynamic images due to the depth of the scene gap,the camera focal length of the different target.This makes the application of dynamic multi-focus image fusion technology in the field of medical,civil,military and other fields have played a relatively broad application prospects.In this paper,we learn the dynamic multi-focus image fusion technology in detail,the main work is as follows:Firstly,this paper introduces the principle and technical characteristics of dynamic multi-focus image fusion,including the principle and characteristics of multi-focus image fusion,the main algorithm of multi-focus image fusion,and the main technology of image fusion from three levels of image fusion.Advantages and disadvantages and the comparison of different technologies is analyzed.After that,this paper introduces the subjective evaluation standard and the objective evaluation standard of the image fusion result.Then,the principle and implementation of non-subsampled Contourlet transform(NSCT)image fusion algorithm are introduced in detail.On this basis,an NSCT multi-focus image fusion algorithm is proposed to improve the fusion performance by improving the fusion rule.The improved rule select larger average gradient in the low frequency subband,and in the high frequency subband adopts a kind of mixed fusion rule.The mixed fusion rule is to select larger gray value in the highest decomposition layer and using the rule of fusion based on region intensity.After experiment and analysis,the algorithm achieves better fusion performance.Finally,an image fusion algorithm based on lifting scheme wavelet transform and pulse coupled neural network(PCNN)is proposed in this paper.In the application of image fusion based on lifting scheme wavelet transform and PCNN image fusion algorithm,not only the use of lifting the format of wavelet transform is simple and easy to implement the characteristics,and because the lifting format wavelet transform still has multi-scale,multi-resolution characteristics,Therefore,it is possible to select the appropriate fusion criterion for the components on different scales,in which the low frequency components are fused by comparing the regional energy,and when the high frequency components are fused,the PCNN is selected to make full use of the image High-frequency information in this advantage.The experimental results show that the algorithm can achieve a good fusion effect.