Automatic calibrating and tracking system of driver sight line

Abstract

The present invention relates to an automatic calibrating and tracking system of driver sight line capable of combining a driver head position with a vehicle key part information and automatically calibrating the driver sight line. The driver sight line calibrating process of the system not has the calibrating point setting and sight line justifying process and the automatic calibrating and tracking system automatically calibrates the driver sight line according to the driver head position information and the relative position of the correlative components such as a vehicle rearview mirror and an instrument panel during the driving process. In order to increase the identification effect of the head feature, a light source system is automatically controlled according to the variance in the outside light intensity, namely when in the daytime and the light ray is relatively good, the visible light is used to identify and when in the night and the light ray is relatively weak, an infrared light source is used. The influence on identifying the head feature by vehicle head lights is avoided by the infrared light source in the night. The automatic calibrating and tracking system has important value on aspects such as driver attention research and fatigue drive based on sight line.

Description

technical field [0001] The invention relates to a driver's line of sight automatic calibration and tracking system, in particular to an algorithm and a device for collecting key component position information of people and vehicles and determining the driver's gaze point position through statistical and learning methods. Background technique [0002] Driver's gaze tracking is of great significance to the study of driver's attention and fatigue. However, due to the traditional line-of-sight technology has certain restrictions on head movement, light or application scenarios, such as the most widely used pupil cornea reflection method based on infrared light sources, in order to achieve higher accuracy, the human head must not Move (fix the head with chin rest and other methods); in order to ensure the stability of light and avoid the influence of sunlight, infrared light sources are generally used and carried out indoors; the target of sight tracking is also relatively certai...

AI Technical Summary

Problems solved by technology

Another major problem is that the above methods require a rigorous calibration process, such as requiring the observer to fixate on 6-12 different calibration points sequentially displayed on the screen in order to determine the vector between the pupil and the corneal reflection point and the fixation point on the screen. Correspondence between

This calibration process generally takes about 1-2 minutes to complete well, which is rather cumbersome.

[0003] In a real driving environment, the above method does not work well

For example, it is impossible to calibrate in the car, the driver's head has a large range of motion (observe the left and right rearview mirrors, and sometimes look backwards), the light is not uniform, especially the sunlight can weaken the reflection of the infrared light source on the cornea effect, so widely used infrared-based corneal reflectance methods are no longer effective in real driving environments

Images