Research on the fusion method of spatial data and multimedia information of multimedia sensor networks in cloud computing environment

Abstract

In the process of fusing spatial data and multimedia information of multimedia sensor networks in cloud computing environment, prone to transmit the repeated information many times, the energy consumption of multimedia sensor networks is increased, therefore, this paper proposes a new fusion method for spatial data and multimedia information in cloud computing environment, considering the difference between spatial data and multimedia information characteristics, fusion process is performed respectively. The principle of Laplacian pyramid transform spatial data fusion is adopted and combined the spatial real image features in multimedia sensor networks, to propose a region based fusion method using image information as spatial data fusion characteristics, and the spatial data of different polarization modes are fused. For the integration of multimedia information, the SIFT algorithm is used to extract and match feature points for the image collected in adjacent multimedia sensor networks, to determine the overlap regions between images of larger relevance, the boundary of best image overlap region is selected via optimization approach, the weighted fusion method is utilized to fuse the overlap regions of multimedia information. The experimental results show that the proposed method has high fusion performance and high fusion accuracy.