Fatigue detection method for tracking eye movement rate based on one-dimensional projection

Abstract

The invention provides a fatigue detection method for tracking an eye movement rate based on one-dimensional projection. A single-target image sequence shot by a camera is used as input data. The method comprises the following steps: firstly, extracting detected eye images by using an eye detection module, and outputting the eye images as modules according to a sequence; secondly, performing gray normalization processing on the eye image, performing vertical projection along the x-axis direction, and calculating a variance value of one-dimensional projection; finally, calculating the variance value change rate in the eye closing process to serve as the eye movement rate in the eye closing process, and judging whether the detected target is fatigued or not according to whether the eye movement rates are larger than 0.06 or not continuously twice. According to the method, on the basis that the eyelid movement rate in the eye closing process in the fatigue state is obviously smaller than that in the non-fatigue state, compared with a traditional method based on the blinking frequency or the PERCLOS value, the detection time is shortened to be within several seconds from the minute level, and the early warning time is greatly shortened. Meanwhile, redundant information of the image is eliminated through one-dimensional projection, the calculation amount is reduced, and the requirement of the system for hardware is lowered.

Description

technical field [0001] The invention relates to a fatigue detection method based on one-dimensional projection tracking eye movement rate, which belongs to the field of computer digital image processing. Background technique [0002] Fatigue detection has a wide range of applications in dangerous driving warning and high-risk work mental state detection. At present, the detection methods for fatigue are mainly classified into two categories, one is fatigue detection based on physiological characteristic parameters, such as analyzing the electroencephalogram signal (EEG), electrocardiogram signal (ECG) or electromyography signal (EMG), the fatigue state The physiological information under the normal state is compared with the physiological information under the normal state to find out the physiological characteristic information of the fatigue state. This method has a fast detection speed and high accuracy. The application environment is subject to many restrictions. The o...

AI Technical Summary

Problems solved by technology

At present, the detection methods for fatigue are mainly classified into two categories, one is fatigue detection based on physiological characteristic parameters, such as analyzing the electroencephalogram signal (EEG), electrocardiogram signal (ECG) or electromyography signal (EMG), the fatigue state The physiological information under the normal state is compared with the physiological information under the normal state to find out the physiological characteristic information of the fatigue state. This method has a fast detection speed and high accuracy. The application environment is subject to many restrictions

The other is to extract the features of the facial image of the target based on digital image processing, and judge the blinking frequency of the eyes or the PERCLOS value to judge the fatigue according to the facial feature information. or PERCLOS value to judge, the detection time is relatively long

Images