Enhanced compressed sensing for visual target tracking
in wireless visual sensor networks
Guo Qiang*
Criminal Investigation Police University of China, Shenyang, Liaoning, China
Abstract. Moving object tracking in wireless sensor networks (WSNs) has been widely applied in various fields.
Designing low-power WSNs for the limited resources of the sensor, such as energy limitation, energy restriction,
and bandwidth constraints, is of high priority. However, most existing works focus on only single conflicting opti-
mization criteria. An efficient compressive sensing technique based on a customized memory gradient pursuit
algorithm with early termination in WSNs is presented, which strikes compelling trade-offs among energy
dissipation for wireless transmission, certain types of bandwidth, and minimum storage. Then, the proposed
approach adopts an unscented particle filter to predict the location of the target. The experimental results
with a theoretical analysis demonstrate the substantially superior effectiveness of the proposed model and
framework in regard to the energy and speed under the resource limitation of a visual sensor node.
© 2017
SPIE and IS&T [DOI: 10.1117/1.JEI.26.6.063028]
Keywords: unscented particle filter; compressive sensing; wireless sensor networks; visual object tracking.
Paper 170591 received Aug. 9, 2017; accepted for publication Nov. 28, 2017; published online Dec. 19, 2017.
1 Introduction
Visual sensor networks have been an attractive field of study
in several distinctive applications, such as remote environ-
ment monitoring, traffic and safety monitoring, health assis-
tance, and tracking.
1–4
A wireless sensor network (WSN)
composed of a large number of small sensors with low-
power transceivers is an effective tool for gathering data in
a variety of environments.
5
Wireless visual sensor networks (WVSNs) are greatly
emerging recently, among which camera nodes and many
relay nodes are deployed to catch and convey the visual
signal. Currently, WVSNs are mainly used for surveillance
applications in government security, law enforcement and
military, traffic monitoring, etc.
6
The role of visual sensors
is similar to the generalized WSNs. However, the volume of
visual data in WVSNs is too large compared with traditional
scalar signals. Therefore, visual data processing in the
network is more complicated, and hardware with high
configuration is required. This burden of transmission of
visual data affects its wide application. Moreover, the noisy
condition in surveillance is also a challenge to the signal
channels. Consequently, signal compression and tracking
reliability are the restraining factors when designing WVSNs
applications.
In recent years, much research has been focused on
visual-target-surveillance systems by WSNs. Kumar et al.
7
presented a survey of recent advances in WSNs. Wang et al.
8
presented a multiview visual-target-surveillance system in
WVSNs. The classification and tracking of moving objects
are alternatively processed for online learning and localiza-
tion. Bhuvana et al.
9
put forward an object tracking method
with surprised observations built upon the information filter
in WVSNs with multiple cameras.
Although much progress has been made, challenges still
remain. Memory limitation, as well as certain types of band-
width and energy constraints, hinders the full application
and further development of a robust tracking framework
in WVSNs. More recently, because of strong capabilities
of signal acquisition techniques, compr essed sensing (CS)
has been applied successfully in image processing tasks.
Unlike traditional WSNs using an information theoretic
approach for decoding, incorporation of CS theor y in
WSNs leads to economic power cost, lower time delay,
more efficient energy management, and higher success rate
of data transmission.
10
Experimental evidence in Ref. 11
proves that signals in WSNs are sparse, which also provides
a theoretical guarantee. In Ref. 12, Fayed et al. proposed an
adaptive block CS technique to represent the captured video
frames, which makes wireless transmission efficient for
energy saving and minimized memory storage. Zhang et al.
13
proposed optimal cluster-based mechanisms by a modi fied
multihop layered model for energy balance. Al sheikh et al.
14
introduced a sparsity inducing algorithm for data aggregation
of nonsparse signal in WSNs. However, the reconstruction
of visuals signal in the network is not sophisticated, fast, or
efficient enough, and traditional CS without custo mization
hinders the performance of WSNs significantly.
Motivated by these, the improved compressive sensing
approach based on memory gradient pursuit (GP) and
unscented particle filter (UPF) is proposed and implemented
on WVSNs. To summarize, the main contributions of our
work are as follows: first, an improved compressive sensing
approach based on a fast customized memory GP algorithm
is presented; it will efficiently extract the moving objects
in the video sequence without compromising the energy
constraint. Second, by integrating a CS-based background
model, the improved CS algorithm is integrated into an
*Address all correspondence to: Guo Qiang, E-mail: royinchina@163.com 1017-9909/2017/$25.00 © 2017 SPIE and IS&T
Journal of Electronic Imaging 063028-1 Nov∕Dec 2017
•
Vol. 26(6)
Journal of Electronic Imaging 26(6), 063028 (Nov∕Dec 2017)
