17-09-2016, 03:03 PM
Spatial Correlation Based Image Compression and Image Mosaicking for
Wireless Multimedia Sensor Networks
1455168629-chaitra1.pdf (Size: 3.45 MB / Downloads: 5)
Abstract
WMSN consist of camera node which can capture the image
continuously and send observed image with the other cameras of the
overlapped Field of View (FoV) to sink.
The overlapped FoV leads to redundant information transmission in
WMSN. Battery power is a scare resource in WMSN which will
consume more power in WMSN node.
To avoid redundant transmission Entropy, Joint Entropy and Mutual
Information is performed to estimate Entropy Correlation Coefficient
(ECC) which describes the correlation characteristics of images
observed by cameras with overlapped sensing area.
Two types of redundancy are removed. First technique is to remove
redundancy using maximum correlation coefficient value and Second
technique is to select minimum entropy correlation coefficient
Then the overlapping images are mosaic to get 3D image with greater
information. Mosaicked image may have photometrical disparities.
The photometric disparities are removed using probabilistic color
correction algorithm. Using mean shift algorithm color corrected
regions are segmented and then extracted using fusion algorithm.
Introduction
Wireless Multimedia Sensor Networks (WMSNs)
They are highly distributed and self-organized networks.
Used to pass multimedia data like images,videos and audio.
Used in many applications such as video surveillance, Environmental
monitoring, home automation, and industrial process control etc.
In a WMSN, a number of camera sensor nodes are deployed in a field
of interest with one or more data sinks located either at the center or
out of the field.
The camera sensor nodes capture the information at different
locations in the overlapped FoV and send their observations to the
sink is in the raw data form.
This redundant information is leads to limited storage,energy
consumption,transmission bandwidth and resource.
Motivation
In the WMSN,multiple cameras sensors are need to capture, compress
and transmit a large amount of data from each sensor node to sink.
Hence maximal image compression cannot always minimize the
overall energy consumption of the networks which consumes more
energy than communication.
This is motivated to apply a spatial correlation on the redundant data
to the each cluster head and send to the sink node,hence this reduce
the transmission energy and power.
For sufficient utilization of constraint resources on generating high
resolution images with wide field of view,a cooperative image
mosaicking algorithm among multiple camera sensor node is proposed.