Image Fusion Based On Adaptive Pulse Coupled Neural Network

The basic idea of image fusion is to integrate multiple images into a image.The pixels of this image are clear.And this image can contain as much information as possible about the input image.It is not only the result of multi-image complementation,but also a high reliability and high quality image.Image fusion technology is convenient to the human eye observation,is also convenient for the rest of the image processing steps,has been applied to more and more fields,and the technology is also put forward more and higher requirements.At the present,pulse coupled neural network is widely used in image processing because of its unique biological background.But the traditional image fusion methods using pulse coupled neural network have much disadvantages such as low quality,low efficiency and poor effect.The complex parameter setting problem has always been a research hotspot.Therefore,the main contents of this paper are as follows:The six parameters of the pulse coupled neural network are optimized.Among them,the connection strength ? is used to set the definition of each coefficient to represent the connection weight of the neuron through the window.The time attenuation coefficient of the feedback input ?F and the time attenuation coefficient of the connection input ?L adaptive set the neuron transitive relations the through the window of correlation of the other coefficients.The time attenuation coefficient of the threshold ?? adaptive set the refractory period of neurons according to the level of activity in the neighborhood.The range coefficient V? of the threshold is derived from the adaptive optimization of bacteria fitness function in the bacterial foraging algorithm,and the quality of the image is evaluated as the fitness function of the bacteria.The external stimulus Sij reflects the activity of neurons by calculating the spatial frequency of the four directions.On this basis,an image fusion algorithm based on adaptive pulse coupled neural network is proposed by combining the optimized pulse coupled neural network with multiband wavelet transform.Using multi-band wavelet transform make the image decompose into high frequency and low frequency coefficient,the high frequency coefficient is fed into the adaptive pulse coupling neural network,and the high frequency fusion coefficient is obtained according to the high frequency fusion rule,The improved Laplace energy is used to calculate the low frequency coefficient.and the low frequency fusion coefficients obtained according to the low frequency fusion rules.If it is a color image,it is necessary to decompose the image into R,G,and B components,and finally merge R,G and B components.Using the image fusion method presented in this paper,a comparative experiment was conducted in medical image and multi-focus gray image,and the results of color image fusion are presented.Objective evaluation indicators and visual effect show that the proposed image fusion method mainly improves the information content of the resulting image and enlarges the applicability of image fusion.