The Driver Of The Vehicle Safety Testing Methods

Nowadays with a fast development of traffic, the number of motor vehicles is increasing day by day. However, it brings a series of problems, among which the most prominent are the highly frequent traffic accidents that leads to a great threat to our property and life safety. To a large extent, traffic accidents attribute to some dangerous driving behaviors of the driver, such as drinking alcohol before driving or driving in weariness. Therefore, eliminating hidden trouble before any accident by way of security detection has become a vital task.This thesis focuses on correlative problems of driving security detection, involving criteria of driving security determination and some techniques of the determination process.The whole thesis discusses a machinery-vision-based method and puts forward a new determination criterion that combines the data of eyes detection and driving line detection, which gets improvement of stability and veracity.All the involving technologies are: face detection, eyes detection and the determination of eyes state.Finally, a new eyes detection method is suggested. Based on the grayness integral and template matching methods, the author advances an arithmetic that can get rid of eyeglasses glisten and enhances the veracity. It is also proved that this arithmetic has a good effect in the condition without spectacles and can be universally used.