Research On The Key Technologies For Application Of Video Sensor Networks

As essential technology of the Internet of Things, Wireless Sensor Network (WSN)has important research value and strategic significance. Traditional wireless sensornetwork’s applications have been gradually exteneded from the traditional data services tovoice, images, video and other multimedia services. Video sensor network as a typicalapplication of Wireless Multimedia Sensor Networks (WMSN), has gradually becomeworldwide research interests, and has important theoretical and applied research value.This paper, with study and analysis of Video Sensor Networks（VSN), focuses onresearch of the key technologies involved in video sensor network applications, such as thekey issues of energy efficient routing algorithm over VSN, precision study of positioningand tracking, as well as video and image optimization for VSN transmission. Under thepremise of ensuring the accuracy and effectiveness of the communication, the paperprovides a theoretical guidance for the video sensor network applications.A novel energy-efficient multi-path routing algorithm for wireless sensor network isproposed in this paper. It consists of three elements:(i) gradients to disseminate data overmultiple paths from a source to a sink,(ii) rules to set up disjoint multiple paths, and (iii)policies to select the working path. By limiting the maximum number of the gradients, theexploratory data messages forwarded are decreased efficiently. By using the rules forsetting up disjoint multiple paths, the sink node can get a sufficient number of disjointmultiple paths, which enables our algorithm to recover from a routing failure quickly.Finally, we evaluate its performance and compare to other related algorithms fromperspectives of average dissipated energy and the delay to set up multiple paths.Positioning and tracking technology is an important issue in the video sensor networkapplications, for the premise of limited energy and limited bandwidth, providing accuratelocation information is critical problem in video sensor networks. A key issue in wirelesssensor networks is to achieve required information accuracy under some physical resourceconstraints especially in terms of energy consumption. Intuitively, more accurateinformation can be achieved by using more jointly sensing nodes at the expense of moreenergy consumption. However, both our analytical and simulation results show thatinformation accuracy may also decrease as the number of jointly sensing nodes increases inthe presence of noise, and the increment of information accuracy reaches zero when thisnumber approaches infinity. In this paper, we discusse a model for information accuracy and the number and position of jointly sensing nodes in WSNs based on a quantitativedefinition of information accuracy by exploiting the spacial correlation between senseddata. The analytical results are validated by simulation.Video sensor network usually uses fairly low-resolution images due to the limitedtransmission bandwidth in transmitting images. It is potential to enhance the capturedlow-resolution images using image resolution enhancement technique that is able toproduce a high-resolution image from its low-resolution counterpart.The key challenge ofimage resolution enhancement is to preserve the edge structure in images. In this paper, anew image resolution enhancement approach is proposed, in which the intensity of theunknown pixel is estimated by using a bilateral weighted average of its neighboring pixels,the neighboring pixels with nearer distance have larger contributions. Furthermore, theneighboring pixels belonging to direction with smaller variation have larger contributions.Experimental results are provided and summarized to show that the proposed approachoutperforms several conventional edge-directed image interpolation algorithms. The resultsshow that the proposed approach yields low computational complexity, which implies thatthe approach is more suitable for real-time applications.In summary, this paper studied three key applied technologies in video sensornetworks, which include energy efficient routing algorithm design, positioning and trackinganalysis, video and image optimization. The analysis and simulation results show that theproposed algorithms are capable of improving the image quality and life-time of videosensor networks.