基于深度学习的疲劳驾驶监测融合算法的研究
首发时间:2019-05-24
摘要:现代物流交通行业发达,疲劳驾驶是大量交通事故发生的主要原因,如何精准地检测驾驶员的疲劳状态是当今的研究热点。近年来,机器学习为很多行业的技术都带来了巨大的变革,尤其是卷积神经网络在图像领域的广泛应用。它能够实现自动的特征提取,解决了图像识别中特征提取的难点问题。本文基于深度学习算法的思想展开研究,提取了眼部状态和头部姿态这两个特征,采用特征融合的方法来判断和评估驾驶员的疲劳状态。将最终提取到的特征送入神经网络进行识别,并取得了93%的准确率,证明本文的方法是可行的。
For information in English, please click here
Research on A Fusion Algorithm for Fatigue Driving Detection Based on Deep Learning
Abstract:The modern logistics and transportation industry are developed, and fatigue driving is one of the causes of many vehicle traffic accidents. How to accurately detect the driver\'s fatigue state is a hot research topic today. Recently, machine learning methods have brought great changes to many industries. In particular, convolutional neural networks are widely used in the field of imaging. It can realize automatic feature extraction and solve the difficulty of feature extraction in image recognition. Based on deep learning algorithms, this paper extracted the two characteristics of eye state and head posture, and used the method of feature fusion to judge and evaluate the driver\'s fatigue state. The extracted features were sent to the neural network for identification, and 93% accuracy was obtained, which proves that the method in this paper is feasible.
Keywords: fatigue driving deep learning eye region extraction head pose estimation
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于深度学习的疲劳驾驶监测融合算法的研究
评论
全部评论0/1000