Development Of Weather Phenomenon Recognition System Based On Forward And Backward Light Scattering Principle

Weather phenomenon recognition is of great significance for transportation,agriculture,and human daily activities.In particular,the rapid identification of common weather phenomena such as fog,haze,rain,and snow,and the accurate measurement of visibility,play an important role in the fields of aviation,highway,and shipping.At present,the automatic observation technology of weather phenomena based on optical principles has been widely used,but the current recognition of weather phenomena mainly depends on the precipitation recognition matrix,and the identification information for weather phenomena such as haze is missing;Although the image method based on artificial intelligence can recognize precipitation and visual range obstacle weather such as haze,the recognition effect is related to the selection of image features,and the accuracy of the algorithm needs to be improved.Based on the above issues,this paper studied the differences in particle scattering characteristics of different weather types,and designed and established a weather phenomenon recognition system based on forward and backward light scattering.The specific research work and results are as follows:(1)Based on Mie scattering theory,a theoretical model for distinguishing weather phenomena using forward and backward light scattering is established,and a weather phenomenon recognition system that can distinguish rain,snow,fog,and haze is designed.Firstly,based on Mie scattering theory,the change of aerosol particle scattering phase function with different particle size distributions and the change of forward and backward scattering light intensity ratio of different types of particles are analyzed.According to the coincidence degree of different phase function curves,40° is selected as the forward scattering angle of the system;Since the light intensity ratio of 120° and 40° has a good distinction between particles and a wide range of light intensity ratios,120°is selected as the system backscattering angle.Secondly,according to the selection of scattering angle,the optical and mechanical structure of the weather phenomenon recognition system is designed.In optical design,the light source is collimated and emitted at a divergence angle of 6.2°.The light emission spectrum,filter bandpass filtering,and detector response of the light source are around 850 nm,ensuring effective signal reception.The mechanical design has established a light source emission,forward and backward reception,and main connection structure that meets the requirements of optical design;Then,the finite element analysis of the main related structures of the system is completed.Perform static stress analysis on key components,calculate temperature distribution at the transmitter window under variable conditions,and perform modal analysis on the main body of the system.The finite element results show that the maximum deformation of the component is 41.45μm.The window can prevent frosting under the power adjustment of the heating piece of 3-10W.The resonance position of the system is most likely to occur on the connecting rod,with a natural frequency of around 65Hz,and relevant indicators meet the use requirements.Finally,an electronic module was established to meet the system control requirements,enabling the extraction,processing,and transmission of effective signals.(2)The host computer software of the weather phenomenon recognition system is designed and compiled to achieve real-time serial data collection,parameter curve drawing,database storage,and other functions.The system software is written using Lab View language,and the main functional modules include user registration and login,serial communication,visibility curve drawing,database parameter storage,etc.Use the VISA serial port to complete serial communication and driver calls.With a sampling period of 30 seconds,real-time display of multiple parameters such as temperature,humidity,visibility,weather type,and real-time curve drawing of visibility can be performed through the humancomputer interaction interface on the front panel.Historical data can be queried by expanding the time tree and clicking on the cursor.The OLEDC management program of a PC is used to connect to the Access database,store the measurement data transmitted through the serial port,and create an installation package using the Lab View application generator,making it easy for users to use the software.(3)The system integration,debugging,and calibration have been completed,and the results show that the measurement error of atmospheric visibility within 10 kilometers does not exceed 10%,enabling accurate identification of weather phenomena such as fog,haze,rain,snow,and clean days.Assemble and integrate the system and conduct circuit performance testing.The measured signal waveform conforms to the theory,and the system operates normally.Then calibrate the system transfer coefficient for measuring visibility at the forward receiving end,and the linear coefficient R2 between the forward scattering signal and the extinction coefficient measured by the standard visibility meter is above 95%.Since there is an approximate assumption in the forward scattering method for visibility measurement,the spray device is used to simulate the haze environment to detect the measurement error,and the contrast with the standard instrument shows that the system’s visibility measurement error within 10km does not exceed 10%.Finally,an outfield experiment was conducted in a relatively stable environment to divide the range of backscattered and forward scattered light intensity ratios for different weather types.The results showed that the system had good measurement performance for atmospheric visibility within the range of 50 to 15 km,and the range of light intensity ratios measured in fog,haze,rain,snow,and clean days was 0.16 to 0.19,0.21 to 0.45,0.11 to 0.15,0.75 to 0.9,and 0.47 to 0.58,respectively.Compared to traditional visibility meters or weather phenomenon meters,the system designed and developed in this paper can achieve more types of weather phenomenon recognition.