基于BP神经网络的疲劳判定

Fatigue Recognization Based on BP Neural Network

在现有的数据条件下,疲劳判定的本质问题是分类问题,BP神经网络是一种经常应用于人脸识别领域的分类算法,其不需要一套人为确定的规则,并可以根据有代表性的样本自我学习。论文采用BP神经网络算法进行飞行员的疲劳识别,首先定位出飞行员面部和头部特征点,然后根据特征点计算出相应的特征向量,把这些特征向量输入BP神经网络分类器,最后利用神经网络模糊分类的优点,进行飞行员疲劳判定。

On the basis of current data,fatigue recognization is classification in nature.BP neural network refers to classifica?tion algorithm applied in the field of face identification and can conduct self-training according to typical samples with no need ofman-made rules.In this paper,BP neutral network algorithm is adapted for fatigue recognization of pilots.Features of pilot??s faceand head is firstly located,then the corresponding feature vectors computed in terms of feature is input into BP neural network clas?sifier.Finally,strengths of fuzzy classification of neutral network is utilized for pilot??s fatigue recognization.