A Research Of Image Fusion Based On Spatial Correlation In Wireless Multimedia Sensor Network

Wireless Multimedia Sensor Networks（WMSNs） are distributed sensor networks whichbecome a concern recently. They combine the wireless sensor network and multimedia processingtechnology which could collect multimedia information of images, video streams, sounds and so on,based on traditional wireless sensor network. Compared with traditional sensor network, moremonitoring information collected by camera nodes makes users understand WMSNs well. In denseWMSNs, there must be lots of redundant information among camera nodes whose field are overlap.For WMSNs the nodes with limited energy and constrained processing power, redundantinformation is transmitted will shorten the life of network. As a result, it’s very important toeliminate redundant information of network and prolong the network life by studying correlationbetween camera nodes.In WMSNs, due to limited range of focus, it’s quite difficult for ordinary optical imaging systemto collect a clear image of all the objects of different depth for the area of interest. To solve thisproblem, we fuse images in the vertical direction. So, a scheme for fusion of multi-view andmulti-focal images based on wavelet transform is proposed for WMSNs. The image fusion schemeusing a network structure of clustering, camera nodes which monitor the same screen are requiredhave greater relevance. Image fusion is conducted using the method based on wavelet transform toobtain a high-quality and clear image and reduce the amount of data transmission networkIn WMSNs, the problem of camera nodes of limited field view, limited energy and uneven taskamong camera nodes monitor the same area of interest still exists. In this paper, we propose ascheme of image stitching based on balancing energy consumption for WMSNs in clusteringnetwork to solve the above problem. A panorama image of the area of interest is obtained by imagestitching on head nodes. The scheme applies method of balancing energy consumption that themonitoring tasks are evenly distributed to mutual cooperation camera nodes to avoid these problemsof some camera nodes died due to overweight tasks and shorten the life of the network. It canreduce the amount of data transmission and improve the overall performance of the network.