Adaptive Lane Departure Warning and Nighttime Forward Collision Warning System
| Autor: | Wei-cheng Chien, 簡偉丞 |
|---|---|
| Rok vydání: | 2011 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 99 Currently, land vehicles are the most popular transportation devices in those few years, the amount of vehicles is rapidly increased, and then results in much more traffic accidents. The main factor of traffic accidents is the distraction of drivers, such as lane departure, road departure, rear collision, intersection collision, etc. Accroding to the reports of traffic accidents in many countries, the lane departure and rear collision are two important accident types; thus in this thesis we proposed an adaptive lane departure warning system and a forward collision warning system for night driving. In the adaptive lane departure warning system, we first generate a horizontal second-difference map. Then, we search the lane marks based on a pre-defined lane model in the map, Third, we judge the lane departure accroding to the detected lane marks. If the system fail to detect lane marks, the proposed adaptive learning module is launched to set proper range of line parameters for following frames. With such a learning strategy, the system will look for lane marks in the following frames as quickly as possible. In this study, we have implemented this system on a embedded system, and improved the system execution performance. The system performs an accepted frame rate of 12 frames per second with ARM 9 500MHz CPU. In nighttime forward collision warning system, we detect the tails of preceding vehicles. Then, we pair the lights using the features of: the horizontal distance, the vertical heights, the trajectory, and correlation of a pair of lights. Finally, we estimate the time to collision (TTC) of the verified light pair. The system was experimented on a personal computer and performed a high execution rate of 30 frames per second. The detection rate of preceding vehicles is also highly at 94.30% in average. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|