Study on Errors Correction of Infrared Methane Sensor Based on Support Vector Machines
Infrared absorption spectrum theory is introduced and the exiting problems of absorption model are indicated. In order to improve the capability of the methane sensor, the errors correction model was set up by adopting nonlinear regression model based on Support Vector Machines (SVM) which is powerful for the problem characterized by small sample, non-linearity, and local minima .Gaussian RBF kernel was adopted in the model. The experimental results show that errors of the concentration of methane is greatly reduced by adopting the errors correction model of Support Vector Machines, and the model can eliminate all kinds of influence such as temperature, humidity, and has high precision, wide measuring scope, meet the requirements of mine.
infrared absorption model kernels Support Vector Machines errors correction
Li Zhang RuLin Wang KuiKui Liu
Institute of Mechnical and Electronic Engineering,China University of Mining and Technology (Beijing Institute of Mechnical and Electronic Engineering,China University of Mining and Technology(Beijing)
国际会议
长沙
英文
1423-1427
2009-10-10（万方平台首次上网日期，不代表论文的发表时间）