| The detection of atmospheric aerosol particle size concentration has always been one of the important issues in industrial production,environmental climate research and human health.In this era of rapid development of science and technology,various industries have put forward higher requirements for the detection of atmospheric aerosol particle size.Among the current hundred controversial measurement methods,the linear array CCD measurement method has become one of the current mainstream measurement methods because of its high efficiency,real-time,and high precision.During the thesis,first,the principle of light scattering is introduced in detail.Based on the principle,various measurement methods and theoretical calculation methods are derived.The schematic model of the particle scattering method is briefly described through schematic diagrams or principle formulas.The thesis combines the measurement model of light scattering method and deep learning algorithm to study the measurement of aerosol particle size and concentration.The principle of light scattering is introduced.And use deep neural network to optimize the data.The main research work of this paper is based on the optimization of MIE measurement technology,taking aerosol particles as the research object.Aerosol has an environmental effect,can affect the atmosphere and the ozone layer,have an adverse impact on human health through industrial production,automobile pollution and other means.The optical platform is designed by using VC ++,Matlab and Spyder.The method using a one-dimensional linear array photoelectric sensor as a measuring device is proposed to achieve high-precision measurement of aerosol particle size and concentration.Through the method of instantaneous direct imaging calculation of the flow velocity aerosol particle size,the aerosol concentration detection in the real-time area scale is carried out.For the first time,it’s theoretical model uses the conjugate magnification imaging of the linear array photoelectric sensor and the high power microscope objective lens to avoid the full-scattering Mie optical model with large conventional errors,while overcoming the relaxation effect of the instantaneous sampling of the area array imaging.Less quantity and higher measurement accuracy.The experimental results show that the aerosol particles detectable by the system range from 1 to 10 μm,and the measurement error is better than8%.On the basis,the artificial intelligence deep learning algorithm was introduced,and the BP network and LSTM network were selected as the experimental control group,and the BP network and LSTM network were reproduced through the spyder and matlab platforms.It is optimized for the linear array photoelectric sensor particle size measurement system.Compared with the original data,the relative error is reduced by 10% on the original basis.At last the interface program for designing the particle size measurement system of the linear array photoelectric sensor is introduced. |
