Visibility Measurement Method Based On Video Images Under Non-uniform Conditions

Visibility is one of the basic factors of meteorological observation.It is closely related to people's daylife.In the aspect of traffic safety,low visibility may lead to serious traffic accidents.So it is a key meteorological factor in the fields of highway,civil aviation and port meteorology.On the other hand,visibility can directly reflect the distribution of atmospheric aerosols.And it is one of the important indicators to measure air pollution.At present,imitating the principle of visibility measurement with human eyes,digital photography method using image contrast to measure visibility has become a new visibility measurement technology.Digital photography visibility system has the advantages of simple installation and convenient maintenance,and it is expected to become a new generation of visibility measurement system.In this dissertation,the error of dual-luminance-difference method was calculated theoretically.The error sources of atmospheric irradiation under nonuniform conditions were analyzed.An improved method with verified was proposed for visibility measurement under high visibility condition.Theoretical analysis shows that the error caused by non-uniform sky background and non-ideal black body is inversely related to visibility.If the visibility is higher,the tolerance to error is lower.The experimental data also show that the average error of the measured visibility is larger when the visibility is higher.The sources of error are as follows.Firstly,the analysis result shows that the visibility measurement error is introduced because of the brightness difference caused by the distribution of solar irradiation when the sky background and the target object are not on the same horizontal line.The magnitude of the error is related to the position of the Sun.In the region near the Sun,the brightness gradient is large.So the error of measurement is large.Secondly,aerosols along different line of sight paths differ in their scattering of light,resulting in non-uniform brightness of the gas column between the targets and the camera.In addition,environmental factors such as surface reflectivity and other lights may increase or decrease the brightness of the gas column of the targets and sky backgrounds,which may affect the measured visibility.It was proved by the data simulation that the small change of the observation condition may cause the great change of the visibility measurement value at high visibility.The dual-luminance difference method is easily affected by the non-uniform condition of the sky,especially in high visibility.To solve this problem,the image frequency contrast method was used to correct the visibility measurement of dualluminance difference method under high visibility condition.After filtering by homomorphic filter and Gaussian high-pass filter,the image with only high frequency information is obtained.The brightness of the image is normalized and averaged to obtain the visibility index.And the visibility index is converted into atmospheric visibility value by fitting function.This method is verified by the data from existing digital photography visibility measurement system and Vaisala atmospheric transmittance meter.These results make us fully understand the measurement principle of video visibility.They give us a more comprehensive understanding about the influence factors of video image measurement visibility.The accuracy and precision of video visibility measurement are improved by the new method.And it is expected to provide technical support for the follow-up video image visibility measurement system for highway visibility measurement.