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 non-linear 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.