Research the accurate detection of indoor pollution gas. This paper proposed a reinforced neural network indoor pollution gas detection method. The Obtained excitant pollution gas odor characteristic signals were filtered to get rid of the external factors interference, and a mathematical model of reinforced neural network was established. The method of reinforced characteristics was used to test indoor pollution gas. The experimental results show that the algorithm enhances the inspection accuracy of indoor excitant pollution gas.%研究室内污染性气体准确检测问题.室内甲醛挥发速度较快,残留的气体由于含量过少,长时间的挥发会导致污染气体特有的气味特征信号衰减,造成信号检测困难.传统的气味检测方法需要根据明显的气味特征信号建立高斯识别模型,一旦房间静置时间过长,引起气味特征淡化、信号衰减,使模型无法准确检测正确信号,造成识别困难.为解决上述问题,提出了一种加强型神经网络的室内污染气体检测方法.对采集的刺激性污染气体气味特征信号进行滤波处理,去除外界因素对于污染气体特殊信号的干扰,建立加强型神经网络数学模型,运用特征强化的方法对室内污染气体含量进行检测.实验结果表明,改进算法提高了室内刺激性污染气体含量检测的准确性.
展开▼
