Driver fatigue detection method fusing facial features and image pulse heart rate estimation

Abstract

The invention relates to a driver fatigue detection method fusing facial features and image pulse heart rate estimation. The method comprises the following steps: 1) carrying out data initialization on a driver; 2) carrying out face detection on the driver by a video acquisition module, and acquiring and sending a current facial feature image; 3) utilizing the pre-training model to position the face; 4) processing the current facial feature image after positioning, calculating a current eye closing time ratio by adopting a PERCLOS algorithm, and calculating a current yawing frequency by adopting a yawing detection algorithm; 5) calculating heart rate difference values of current left and right cheek heart rates P1 and P2 by adopting an IPPG technology; 6) judging whether the driver is tired by using a fuzzy neural network system, 7) monitoring whether the video acquisition module does not detect the face time t>20 minutes, if so, repeating the step 1) when the face is detected next time, and if not, continuously sending out fatigue early warning.

Description

technical field [0001] The invention relates to the technical field of transmissions, in particular to a driver fatigue detection method which combines facial features and image pulse heart rate estimation. Background technique [0002] 2.2% of global deaths are caused by road traffic accidents every year, and the number of deaths caused by road traffic accidents is expected to increase within 20 years. Driver fatigue driving is an important cause of traffic accidents, which is caused by the driver in the human-vehicle-environment system. When drivers are fatigued, they will show some special physiological and psychological phenomena. According to the length of driving time, fatigue driving can be divided into short-term driving fatigue and long-term driving fatigue. Symptoms of short-term driving fatigue include: blinking frequently, feeling a little tired, and reducing attention to safety; shifting gears is not timely, inaccurate, and lack of concentration; the car does ...

AI Technical Summary

Problems solved by technology

The disadvantages of this detection method are: it will be affected by many uncontrollable factors, such as: complex road conditions, changeable weather, personal driving habits, and different models; Increased the difficulty of detecting the driver's fatigue state

This detection method usually uses the camera to obtain the real-time state detection information of the driver, extracts features through image processing technology, and analyzes the real-time state detection information of the driver to determine the driver's fatigue state, but this method lacks direct quantification standards , without considering individual differences

The shortcomings of this detection method are: in order to obtain a clear image of the driver's state more accurately, expensive special image acquisition equipment is often used, and this method lacks direct quantification standards and does not fully consider individual differences.

The shortcomings of this detection method are: because it needs to directly contact a certain part of the driver to complete the monitoring, the driver needs to wear professional equipment to collect relevant physiological parameters, but many drivers do not wear detection equipment while driving. Habits, to a certain extent, affect the comfort of driving, and then affect the safety of driving. In addition, the reason for further weakening the use of such detection devices is: the cost of such detection devices is too high

However, due to the phase shift of the measured part and the image acquisition device in this method, this relative movement will produce motion artifacts, resulting in motion noise

Considering noise interference in actual scenarios, there is no mature IPPG method for driver fatigue assessment

Images