Vehicle Positioning Using Image Processing
An image-processing approach is described that detects the position of a vehicle on a bridge. A load-bearing vehicle must be carefully positioned on a bridge for quantitative bridge monitoring. The personnel required for setup and testing and the time required for bridge closure or traffic control are important management and cost considerations. Consequently, bridge monitoring and inspections are good candidates for smart embedded systems. The objectives of this work are to reduce the need for personnel time and to minimize the time for bridge closure. An approach is proposed that uses a passive target on the bridge and camera instrumentation on the load vehicle. The orientation of the vehicle-mounted camera and the target determine the position. The experiment used pre-defined concentric circles as the target, a FireWire camera for image capture, and MATLAB for computer processing. Various image-processing techniques are compared for determining the orientation of the target circles with respect to speed and accuracy in the positioning application. The techniques for determining the target orientation use algorithms based on using the centroid feature, template matching, color feature, and Hough transforms. Timing parameters are determined for each algorithm to determine the feasibility for real-time use in a position triggering system. Also, the effect of variations in the size and color of the circles are examined. The development can be combined with embedded sensors and sensor nodes for a complete automated procedure. As the load vehicle moves to the proper position, the image-based system can trigger an embedded measurement, which is then transmitted back to the vehicle control computer through a wireless link.
A. Kaur et al., "Vehicle Positioning Using Image Processing," Smart Materials and Systems II, SPIE, Jan 2009.
The definitive version is available at https://doi.org/10.1117/12.815321
Electrical and Computer Engineering
Article - Conference proceedings
© 2009 SPIE, All rights reserved.
01 Jan 2009