Research On Gas Monitoring Technology Based On Infrared Hyperspectral Imaging

The gas passive infrared hyperspectral imaging monitoring technology has many advantages such as remote detection,real-time on-site,no radiation source,direct imaging and the like,and is one of the effective methods for realizing non-contact imaging gas detection.However,due to the large amount of noise in hyperspectral imaging data,the large number of bands,and the fact that the gas characteristics are weak signals,it is difficult to accurately and rapidly monitor the gas.At present,the main gas detection methods are the spectral fitting and data training of pixel elements,which takes time and cost.Therefore,gas recognition of infrared hyperspectral imaging data in open environment,low concentration,and complex backgrounds has great application value and practical significance for rapid monitoring and display of two-dimensional distribution of target gas space.The purpose of the research on gas monitoring technology based on infrared hyperspectral imaging is to identify the gas composition and quickly monitor and display the two-dimensional distribution of the target gas space.Wavelet threshold denoising,automatic peak position detection,rapid identification of gas species;proposed principal component analysis(PCA)and radiation temperature difference extraction method,extraction of target gas and background,visualization of monitoring results;and final realization of hyperspectral gas comprehensive analysis system design.The main content of the paper is as follows:First,the infrared radiation principle and infrared radiation transmission model are used and studied to derive the relationship between the measured gas cloud transmittance and radiation temperature change;design and implementation of HITRAN-based gas spectral characteristics simulation software to simulate the permeability of multiple gases Spectral;For the radiation temperature spectrum,select the appropriate wavelet basis and wavelet parameters for wavelet threshold denoising;use baseline correction and spectral derivation to automatically identify the absorption peak position,determine the presence of the target gas;use the reference spectrum for spectrum matching to detect the target gas distributed.Secondly,in the case of known gas species,a method based on principal component analysis(PCA)and the two-dimensional distribution of extracted gas space based on radiation temperature difference and the background of the measured area is proposed.The extracted feature image is smooth filtered,image segmentation,morphological processing,color image fusion,and the background image of the measured area with two-dimensional distribution of superimposed gas space is finally obtained.Finally,the infrared spectrum processing function and gas spatial distribution detection program for hyperspectral images based on radiation temperature difference were realized through Visual Studio 2013 and OpenCV.A hyperspectral gas comprehensive analysis system was designed.The detection speed meets the requirements of field level monitoring and can be dynamic.The coordinates of the visual field are displayed,infrared spectrum analysis is performed,the software interface is clean,the operation is simple and user-friendly.