A High-speed Trapezoid Image Sensor Design for Continuous Traffic Monitoring at Signalized Intersection Approaches
Project Title
A High-speed Trapezoid Image Sensor Design for Continuous Traffic Monitoring at Signalized Intersection Approaches
University
University of Idaho
Principal Investigator
Suat Ay, Ph.D.
U of I Electrical & Computer Engineering
PI Contact Information
U of I Electrical & Computer Engineering
Funding Sources and Amounts Provided
US Department of Transportation/TranLIVE — $48,000
University of Idaho — $48,000
Total Project Cost
$96,000
Agency ID or Contract Number
DTRT12GUTC17; KLK910
Start Date
8/18/2013
End Date
9/30/2014
Description of Research Project
The goal of this project is the monitor traffic flow continuously with an innovative camera system composed of a custom designed image sensor integrated circuit (IC) containing trapezoid pixel array and intelligent future extraction engine.
Implementation of Research Outcomes
A unique image sensor was designed and sent for fabrication Jan. 3, 2014. Imager design composes of 40 x 151 scaled pixels for focal plane trapezoid image sensor array that could be able to monitor objects 250m from the stop line. Software for background and moving objects extraction were developed and demonstrated on a platform using an existing rectangular imager. The fabricated imager was received the first week of May 2014. Die photograph is shown in the figure below:
The developed software will be modified for the trapezoid pixel structure and once the image sensor is received from the fabrication, the software will be integrated with the image sensor and the complete traffic monitoring system will be ready for field testing.
The camera system is specifically designed for continuously monitoring incoming traffic at an intersection. It will extract the vehicle location, speed and acceleration information continuously.
This sensor and camera system will allow dynamic traffic control with reduced communication bandwidth and complexity proving only necessary information without requiring high performance backend processors.
A senior undergraduate design team composing of two electrical engineering, one computer engineering, and two computer science students were recruited during fall 2013 and spring 2014 semesters to develop algorithms for the trapezoid imager.
Impacts and Benefits of the Project
Continuous monitoring of vehicle’s speed, location, and acceleration in the trapezoid field of view (FOV) of the camera is done in image sensor level intelligently without requiring complicated digital post processing of video streams and high communication bandwidth associated with it.
Continuous monitoring will allow dynamic traffic control. Dynamic traffic control has potential benefits for national economy and environment by reducing fuel consumption of city vehicles and reducing CO2 emission.
Web Links
Final Report: UI_TranLIVE_FinalReport_High-SpeedTrapezoidImage
Team New Perspective: http://mindworks.shoutwiki.com/wiki/Team_New_Perspective
Keywords
- Signalized intersection
- traffic surveillance
- traffic signal control systems
- image sensor
- trapezoid
- CMOS image sensor
- trapezoid image sensor