Reconstruction Of Gas Distribution Based On Imaging Differential Optical Absorption Spectroscopy

The atmosphere pollution has been a serious problem.It serious influence people's health and the development of economics.For understanding the distribution of the gas contaminant is a precondition of controlling atmosphere pollution,accurate spatial distribution of polluted gas is very important for pollution control.Because traditional methods which based on electrochemistry are easy to be effected by air quality in the ambient environment,they are hard to show us gas distribution in a large scale.The method of optical remote sensing is an effective means to obtain the spatial distribution of gas.The combination of differential absorption spectroscopy and computed tomography is a method to obtain the spatial distribution of gas according to the value of path integral concentration.Differential optical absorption spectroscopy(DOAS)technology has advantages of large scope of measurement,non-contact and real time et al.Pollution control ask for knowing of the gas contaminant distribution.An extensive studied method for reconstructed gas contaminant distribution that combines DOAS technology with computed tomography(CT)technology.Long path differential optical absorption spectroscopy(LP-DOAS)technology demands reflectors,which limit its' measure height.Multi-axis differential optical absorption spectroscopy(MAX-DOAS)does not need reflectors.For MAX-DOAS can measure only one path integrated concentration(PIC)every time,the temporal resolution is low.Imaging differential optical absorption spectroscopy(IDOAS)use 2D CCD to be its'image device,and measures dozens of PICs of azimuth at the same time.The time resolution of IDOAS of the measured results is greatly improved compare with MAX-DOAS.In this paper,a data acquisition system which based on IDOAS and the calibration method for the system is designed.A set of calibration methods for the system is given.The system is used in feasibility experiment and field campaign for collecting path integrated concentration.Based on the path integrated concentration,computed tomography(CT)algorithms are used to reconstruct spatial distribution of gas pollutant.The algorithms are then evaluated.In this paper,the reconstruction of stack plume from power station is mainly researched.At first,classical CT algorithms are reviewed and briefly discussed.In chapter 3,the algorithm called sequential coordinate-wise algorithm(SCA)which is a kind of non-negative least squares problem is introduced for reconstructing stack plume.The algorithm is then evaluated according to the influence of box discretisation,plume position and error of measurement.To further improve the stability of the reconstruction,compressed sensing is introduced to the field of gas reconstruction for the first time.Based on the compressed sensing,classical algorithm of low third derivative(LTD)model is improved by using total variation(TV).Two algorithms called POCS-LTD and BOCS-LTD were designed.Numerical simulation showed that the algorithms are insensitivity to box discretisation and plume position.The algorithm showed excels traditional LTD method in anti-error capabilities and reduced artifacts of reconstructed images.The feasibility experiment was carried out on the top floor of villa NO.3 of Hefei Institutes of Physical Science.The molecular density of the gas in the quartz glass container was analyzed.The gas distribution in the container was reconstructed using the POCS-TVM method.The peak value of the reconstructed concentration coincides with the maximum concentration by linear regression,and the gas position coincides with the actual position of the container,thus verifying the feasibility of the IDOAS technique for gas reconstruction.The field experiment was carried out outside a power plant in Huai Nan.The spectra were collected from the stack plume of the power plant.The spectra acquisition system is built.The direction of the two IDOAS is calibrated according to the latitude and longitude of the IDOAS and the sun as the common reference.By using the path integrated concentration retrived from the collected spectra,the images of the gas distribution were reconstructed by the SCA method and the third derivative method based on total variation,and the effectiveness of the algorithm is verified.