Authors:
Manal Alsabhan
and
Adel Soudani
Affiliation:
College of Computer and Information Systems and King Saud university, Saudi Arabia
Keyword(s):
WMSN, Multimedia Sensing, Object Recognition, Low-power, Fourier Descriptors, Zernike Moments.
Related
Ontology
Subjects/Areas/Topics:
Data Communication Networking
;
Energy Efficiency
;
Energy Efficiency and Green Manufacturing
;
Enterprise Information Systems
;
Image Processing
;
Industrial Engineering
;
Informatics in Control, Automation and Robotics
;
Internet of Things
;
Obstacles
;
Power Management
;
Robotics and Automation
;
Sensor Networks
;
Software Agents and Internet Computing
;
Software and Architectures
;
Telecommunications
;
Wireless Information Networks
Abstract:
In Wireless Multimedia Sensor Networks (WMSN), image-based sensing applications face the issue of energy efficiency and the availability of resources. This issue leads to image sensing and transmission severely exhausting the sensor energy, potentially flooding the network with unnecessary data at the application level. Compression of the image fails to solve this issue efficiently, due to the complexities of the algorithm. Thus, the approach of employing image sensing to detect an event of interest locally prior to transmission of the Region of Interest (ROI) would avoid useless data transmission, and consequently save energy. This approach promises to extend the life of the entire network while reducing the sensing time. The main contribution of this work is to establish a low-complexity scheme for image sensing in WMSN. This scheme based on 2D General Fourier shape descriptors for target recognition and notification to the end user. This current paper outlines the specification of
the proposed scheme and its implementation on wireless multimedia sensors. It addresses the performances evaluation regarding time and energy consumption. The results reveal the high levels of accuracy of the proposed approach in efficiently recognizing the target and notifying the end user. It shows a significant performance that overcomes the efficiency of alternative similar sensing approaches that have been proposed in the literature.
(More)