Research On Application Of SVR In Infrared Gas Detection Technology

Accurate detection of gas concentration is of great significance to people’s life and industrial production. Compared with the traditional detection methods, infrared gas detection technology has become a hot spot in the field of gas detection because of its advantages such as high sensitivity, wide measurement range and quick response. However, the technology is easy to be affected by the environmental factors, and the direct use of the technology will have a large measurement error. Single wavelength dual optical path difference absorption model can eliminate some errors to a certain extent, but the compensation ability is limited.In order to improve the measurement accuracy, this dissertation proposes a measurement error prediction model based on SVR (Support Vector Regression, SVR) on the basis of the differential absorption model. In the study of the correlation theory of SVR, it found the regression performance of SVR is closely related to the selection of parameter values. This dissertation puts forward the method of parameters selection based on Glowworm Swarm Optimization algorithm. Because of the fixed step size, the basic algorithm has problems that the convergence is slow and solution accuracy is not high. This dissertation designs step factor to adjust the step size adaptively. The validity of the improved method is verified by the test of the standard function and regression experiment based on housing data sets.Finally, this dissertation constructed the infrared gas detection system, and detects gas concentration by using the error prediction model. The Experimental results show that the error prediction model based on SVR can improve the measurement accuracy of infrared gas detection technology.