Monocular Computer Vision Aided Road Vehicle Driving for Safety

The meaning of this research is using the techniques of image processing and computer vision to aid road vehicle driving to improve the safety of road driving, to reduce the traffic hazard, and to eliminate the damage of life and property. The primary functions of the proposed system include: detecting the distance to a front obstacle or vehicle by analyzing the image sequence, determining the left/right location of the vehicle in the current lane, and estimating the vehicle speed based on a continuous image sequence. The advanced functions include: detecting the illegal driving, predicting the motion of neighboring objects, detecting the run through a red light, analyzing the physiological situation of drivers, traffic mark recognition, ..., etc.

This system will combine the following subsystems to reach the real-time tracking and supervision of driving situation.

1. Global position system: using digital map, GPS, and difference broadcasting system to track and supervise the vehicles in real time.

2. Data record and image compression: recording the vehicle location, speed, front-vehicle distance, types of illegal driving, the time of illegal driving, and the compressed images.

3. Real-time wireless transmission: transmitting all above data and images to the control center, searching the illegal-driving images, and replaying the driving route.

 

Results: detecting the distance to a front obstacle or vehicle,

determining the left/right location of the vehicle in the current lane,

estimating the vehicle speed

 

Members: H.-C. Wu, C.-W. Lin, Frank Liu

 

A camera is installed in the vehicle and faces to front road.

Road image captured by the camera.

 

Combining the visual analysis and geographical information system

to supervise the situations of all related vehicles.