A Research On Image Stitching Technology In WMSN

As an emerging technology, development wireless multimedia sensor networks network hasattracted the attention of a growing number of researchers. Introducing multimedia sensors (suchas CMOS camera and miniature microphone) into wireless sensor networks, wireless multimediasensor networks acquire multimedia information via wireless self-organizing network. Wirelessmultimedia networks is of great importance in the applications of video surveillance,environmental monitoring, process control, object tracking, thus attracts many scholars devotethemselves into different research fields , and have produced a lot of meaningful research results.Although the adoption of low-power processors, sensors and wireless communication module, aswell as large-capacity battery, the lifetime of nodes is still below people's expectations for thereis large amount of information to collect and calculate by wireless multimedia sensor networks.Therefore, how to improve battery life and network lifetime is a problem urgent to solve. Up tonow, many researchers have illustrated how to conserve energy and improve the network lifecycle from perspective of compression to remove the image spatial redundancy, routing, andcoverage. In recent years, as an important branch of image processing technology, Imagestitching technology have gained quite a lot achievement and developed several stable, maturealgorithms. Brown, Lowe, et al proposed a fully automatic image stitching technology, which isa breakthrough in intelligent image-stitching. Wai Chong Chia, Li-Minn. Ang, et al firstlyintroduced image stitching technology into the image fusion of the image sensor network, andachieved good results.This thesis extends the approaches of image stitching technology proposed by Wai, ChongChia, Li-Minn. Ang et al, achieves further removal of redundant image data and proposes a novelalgorithm to calculate the correlation coefficient of nodes and perform node clustering.Simulation results show that the proposed redundant data removal and node clustering algorithmeffectively reduces the redundant data in the node image acquisition, and extends the nodesurvival time. This thesis also presents the design and implementation of a image stitchingtechnology oriented prototype system in wireless multimedia sensor networks. The abovealgorithm are implemented on the prototype system, and verified feasible and effective. Asummary and outlook for the above two algorithms and the design of the prototype system arepresented in the last part of this thesis.