Over-height trucks are not only a hazard to the over-height trucks themselves, but they pose a threat to the bridges they come into contact with, and most importantly the other drivers on the road way when a collision takes place with a low clearance structure. Therefore, there is a need for an over-height detection system that is affordable yet also reliable. At this time there exist over-height detection systems using laser and infrared beam devices however, they are expensive. This high cost makes it impossible for Department of Transportations across the nation to implement these systems at all low-clearance headroom roadways. In this research a machine vision based system is proposed to detect the height of trucks and provide a warning for over-height vehicles. The height determination will be completed using line detection and blob tracking; these two methods will be overlapped where an upper point of the truck can be compared to a lower point on the ground. These 2D coordinates will then be translated into 3D world coordinates that will provide an approximation of the truck height. If the truck is over the set height then a warning will sound. The accuracy of the test proves that the method is a reliable method of height determination, achieving a 96.59% accuracy rate for measured trucks. The method does have an error rate of 3.3%. The merit of this work is the creation of an automatic image based method which can provide height determination of trucks and is a low cost alternative to the current expensive laser and infrared detection systems.
| Identifer | oai:union.ndltd.org:GATECH/oai:smartech.gatech.edu:1853/44893 |
| Date | 10 July 2012 |
| Creators | Sandidge, Matthew Jay |
| Publisher | Georgia Institute of Technology |
| Source Sets | Georgia Tech Electronic Thesis and Dissertation Archive |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0023 seconds