基于深度相机的疲劳预警检测算法研究

Fatigue Warning Detection Algorithm Based on Depth Camera

为解决复杂环境下传统彩色图像对危险驾驶行为无法正确识别的难点,本文提出一种基于深度相机的疲劳预警检测方法。首先通过深度相机获取驾驶员的红外图像与深度图像,对获取的红外图像使用局部二值模式(local binary pattern,LBP)特征算子检测定位人脸区域;在人脸区域使用随机森林和全局线性回归相结合的方法训练出模型,并检测定位人脸的68个特征点,进而确定眼睛和嘴巴的闭合状态。为了增强疲劳检测的准确性,判断驾驶员佩戴眼镜情况,采用改进的疲劳检测算法判定驾驶员的疲劳状态,同时采用图像处理方法对眼部状态与嘴部状态进行疲劳检测。检测结果表明,本算法能够有效识别白天和夜间眼睛与嘴部的疲劳状态,具有较强的实用性。

In order to solve the difficulty that traditional color images can’t correctly identify dangerous driving behaviors in complex environments,a fatigue warning method based on depth camera is proposed in this paper.Firstly,the infrared image and the depth image of the driver are obtained by the depth camera,and the local binary region mode feature operator is used to detect the localized face region for the acquired infrared image;the random face forest is combined with the global linear regression method for training in the face region.The model is detected,and 68 feature points of the face are detected to determine the closed state of the eyes and mouth;in order to enhance the accuracy of the fatigue detection,and whether or not the driver wears the glasses,the fatigue state of the driver is determined by the improved fatigue detection algorithm.At the same time,the image processing method is used to perform fatigue detection on the eye state and the mouth state.The test results show that the algorithm can effectively identify the fatigue state of the eyes and mouth during the day and night,and has strong practicability.