Reliable Infrastructural Urban Traffic Monitoring Via Lidar and Camera Fusion

Paper #:
  • 2017-01-0083

Published:
  • 2017-03-28
Abstract:
Infrastructural urban traffic monitoring plays an important role in the present traffic system as it constantly provides reliable traffic data to regulatory organization and individual drivers. Although a number of commercial systems already exist for such purposes, many of them have various issues including expensive installation and operation, power supply constraint and unreliable performance. Therefore, the present traffic system will need a low-cost and reliable solution for infrastructural traffic monitoring. This paper presents a novel design of infrastructural traffic monitoring sensor system that is self-powered, maintenance-free, easy to install and of low-cost operation. The system consists of an infrared camera, a time-of-flight (TOF) laser range finders, and a processing unit to perform vehicle counts, traffic flow speed estimation, and classifications. Using the two types of sensors, the system implements three approaches for traffic monitoring. The first approach uses single beam Laser Range Finder (LRFs) and perform traffic monitoring by using TOF information. TOF information sees the change of measured distance resulting from passing vehicles. This allows the LRFs to count the vehicles accordingly. For the second approach, LRFs are used only to transmit laser beam onto ground and the IR camera tracks the locations of these laser points. Significant location change of the laser points in camera images indicate the vehicle passes. For these two approaches, two parallel sets of sensors are used for speed estimation. Different response time from two sensor sets combined with known-ahead detecting distance provides vehicle speed knowledge. Vehicle classification is made by length measurement for each vehicle based respond time and estimated speed. The third approach utilizes the IR camera only. Using background subtraction and edge detection algorithms, the system recognizes the passing vehicles in each frame. Given frame rate of the camera, speed estimation is performed by counting the frame number it takes for the vehicle to travel certain distance. Classification is made based on the edge weight of a vehicle, which is a good reflection of the vehicle size. A comparative study shows that at least two out of three approaches are effective under various circumstances such as daytime and nighttime. The power analysis of the system concludes that the system has low power consumption and can be supported solely by solar panel. A prototype has been built and a one-hour field test at a public road around Virginia Tech campus shows promising results by achieving good accuracy and proves the reliability of the system.
Access
Now
SAE MOBILUS Subscriber? You may already have access.
Buy
Attention: This item is not yet published. Pre-Order to be notified, via email, when it becomes available.
Select
Price
List
Download
$22.00
Mail
$22.00
Members save up to 36% off list price.
Share
HTML for Linking to Page
Page URL

Related Items

Training / Education
2011-04-20
Article
2016-02-02
Technical Paper / Journal Article
2007-04-16
Technical Paper / Journal Article
2007-11-28
Technical Paper / Journal Article
2008-04-14