Research On Low Power Consumption Image Fusion Algorithm For Wireless Multimedia Sensor Networks

Data fusion has been widely concerned as a key technology in Wireless Multimedia Sensor Networks.As a branch of the data fusion,image fusion can not be ignored.Due to the dense deployment of nodes in Wireless Multimedia Sensor Networks,the energy is limited,so the adjacent nodes have the potential to collect redundant image information.If these redundant information has not been dealt with,it will greatly increase the communication overhead in the network,waste the node energy,so as to shorten the working life of the network.Therefore,the image fusion technology is needed to deal with the redundant information.Most of the traditional image fusion methods are the pursuit of high quality,high definition,but the image fusion method in wireless multimedia sensor networks is the pursuit of low power,low complexity,and the quality of the image is ensured.According to the characteristics of image fusion in Wireless Multimedia Sensor Networks,two related algorithms are proposed.An image fusion algorithm combined with the Nonsubsampled Contourlet transform(NSCT)and Laplacian Pyramid transform is proposed.Firstly,the image sources are decomposed by using Laplacian Pyramid transform,and the high frequency subbands are fused by the rules of the neighborhood window consistency.Then the low frequency subbands are decomposed by using NSCT to get low frequency and high frequency coefficients,which are fused separately by improving the energy contrast and larger absolute value,then the inverse NSCT is used to obtain the low frequency subbands.Finally,the inverse Laplacian Pyramid transform is used to reconstruct the fused image.An image fusion algorithm based on Wavelet transform is also proposed.Firstly,the Wavelet transform decomposition process was simplified to reduce the decomposition of the steps,so the computational complexity is reduced.Then the part of the low frequency coefficients are fused by the methods of standard difference weighted based on window area,and the part of the high frequency coefficients are fused by the rules of the neighborhood window consistency.Finally,the inverse Wavelet transform is used to obtain the final fused image.Compared with the traditional Pyramid method,Curvelet transform method,the fusion effect of the two proposed algorithms is very good,and the computational complexity is low,which made good unity in the aspect of subjective evaluation and objective evaluation.The experimental results show that the proposed two fusion algorithms are suitable for image fusion in Wireless Multimedia Sensor Networks.