Reliable Field of View Coverage in Video-Camera based Wireless Networks for Traffic Management Applications

Abstract

This paper proposes a new method for real-time situations through off-line prediction methods [13].distributed image acquisition through wireless networks of Instead, new concepts are needed that operate as traffic monitoring cameras. The goal is to provide autonomous multi-agent systems based on real-time data comprehensive image acquisition in the presence of dynamic acquisition [8,14]. disturbances, which obstruct the field of view (FOV) of the Several methods have been proposed for video camera cameras. The paper presents a collaborative online algorithm, based traffic management. The work in [9] describes which (upon FOV loss) identifies an optimized set of cameras that must be turned on to cover the FOV loss. The set is found Unicam, a video-detection based traffic monitoring using a novel cost function, also discussed in the paper. system. Two techniques, tripline and tracking, are used for Experimental results are offered to illustrate the effectiveness of traffic analysis. A limitation of the method is its restricted the method in recovering the lost FOV communication layer. The approach in [10] addresses shadow-induced errors in image acquisition. It detects and Keywords -sensor networks, video camera, field of view, suppresses shadows using the color ratio between lit and coverage, collaborative online algorithms shadow pixels. A novel traffic congestion monitoring method is proposed in [11]. The method models cars as agents that deposit pheromone at virtual places. I. INTRODUCTION Pheromone evaporates and propagates following a modified version of the state transition model for digital Smart monitoring systems are emerging as a pheromone. A car predicts the traffic on the road ahead breakthrough concept for many important applications, from the information provided by the preceding cars. The like healthcare, traffic management, security, defense, system in [1] uses video cameras mounted on buses to emergency response, environmental and habitat dynamically monitor the traffic conditions along a traffic monitoring, agriculture, and many more corridor. While existing methods offer valuable [1,2,3,4,5,6,13,15]. The systems are expected to collect contributions, there are still many challenges to be huge amounts of information (such as images, sound, addressed, like the high computational complexity of video temperature, toxin levels, etc.), communicate through processing, difficult modeling and prediction due to a large wireless networks, and collaborate in achieving both variety of situations, and the presence of disturbances (like global and local goals of an application. moving obstacles) that perturb the efficient acquiring of Modem traffic management systems, such as adaptive images. traffic signaling [13], execute complex decision making This paper proposes a new algorithm for real-time algorithms to optimize goals, like improving the traffic distributed image acquisition through a network of traffic flow, minimizing travel times, and minimizing the vehicle monitoring cameras. Neighboring cameras ar...

The Coverage Problem in Video-Based Wireless Sensor Networks: A Survey

“…However, the activation of redundant sleeping nodes can happen due to an external event. Istin et al [ 83 ] proposed an algorithm to maintain image acquisition even in the presence of dynamic disturbance (as moving obstacles), when there are many deployed sensors. Initially, redundant sensors are turned off, letting only one sensor viewing the monitored field.…”

“…The algorithm proposed in [ 83 ] is distributed and online, with a direct (but not limited) application in traffic monitoring. When a node detect that its FoV is lost due to a moving obstacle (as a car or truck) entrance in the camera viewing, it indicate such event to its neighboring nodes.…”

“…Experiments in [ 83 ] verified the loss recovering in the presence of obstacles, studying simulated environments with variable number of cars, speeds and FoV redundancy. In most experiments, the FoV recovery was higher than 40% with the number of cameras ranging from three to five, and higher than 85%, considering the number of cameras from three to 15.…”

“…The work presented in [ 83 ] considers nodes with rechargeable batteries, but expect that energy consumption is kept in minimal levels. A similar research is taken in [ 84 ], considering coverage preservation and network lifetime prolonging in an environment with moving obstacles.…”

Deployment of Efficient Wireless Sensor Nodes for Monitoring in Rural, Indoor and Underwater Environments

“…Finally, in [57], Codruta Istin et al propose a method for real-time distributed image acquisition through wireless networks for traffic monitoring. They also present a collaborative online algorithm, which identifies an optimized set of cameras that must be turned on to cover the field of view loss.…”