Low-Complexity and Energy-Efficient Algorithms on Image Compression for Wireless Sensor Networks


IEICE TRANSACTIONS on Communications   Vol.E93-B    No.12    pp.3438-3447
Publication Date: 2010/12/01
Online ISSN: 1745-1345
DOI: 10.1587/transcom.E93.B.3438
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Section on Wireless Distributed Networks)
digital image processing,  lapped transform,  wireless sensor networks,  discrete cosine transform,  Lempel-Ziv-Welch,  energy balance,  

Full Text: PDF>>
Buy this Article

This paper proposes two algorithms to balance energy consumption among sensor nodes by distributing the workload of image compression tasks within a cluster on wireless sensor networks. The main point of the proposed algorithms is to adopt the energy threshold, which is used when we implement the exchange and/or assignment of tasks among sensor nodes. The threshold is well adaptive to the residual energy of sensor nodes, input image, compressed output, and network parameters. We apply the lapped transform technique, an extended version of the discrete cosine transform, and run length encoding before Lempel-Ziv-Welch coding to the proposed algorithms to improve both quality and compression rate in image compression scheme. We extensively conduct computational experiments to verify the our methods and find that the proposed algorithms achieve not only balancing the total energy consumption among sensor nodes and, thus, increasing the overall network lifetime, but also reducing block noise in image compression.