Atmospheric Visibility Measurement And Inversion Algorithm Research

Visibility refers to the farthest distance that a person with normal vision can recognize an object from the background under certain weather conditions at that time.For a long time,China has become more and more mature in the measurement technology of visible light visibility,and there are few reports on the measurement technology of infrared visibility in the infrared band.Due to the demand of infrared photoelectric engineering applications,this paper develops a set of infrared atmospheric visibility measurement system based on the definition of atmospheric visibility.My main work is to design the control part of the measurement system and write a set of control software.In this paper,the shift subtraction method and the wavelet new threshold denoising method are introduced in the image processing of imaging lidar.Compared with the traditional background subtraction method,the displacement subtraction method effectively reduces the laser beam noise.After the new threshold,the imaging laser The correlation between the radar inversion results and the visibility meter measurement results was improved by 8%,and the correlation between the MIE scattering lidar inversion results and the visibility meter measurement results was improved by 6.2%.The experimental results show that the infrared atmospheric visibility has the same trend as the visible light visibility,and the correlation between the two is 80%.Finally,based on the analysis of atmospheric visibility measurement in coastal areas,three inversion models were used to invert the atmospheric visibility in Maoming,Guangzhou.The experimental results show that the results of the three models have a good correlation with the measured results,and the correlation coefficients are all Up to 85%or more.The statistical analysis shows that the relative error of the support vector machine and the nonlinear regression model compared with the measured data is less than 0.35,and the inversion result based on the support vector machine is more consistent with the measured results,that is,the model built by the support vector machine.Is the preferred model.