Study On The Application Of Binocular Vision Technology In Road Safety Data Detection

With the rapid growth of highway mileage and vehicle holdings in China,road safety has increasingly attracted extensive attention of the whole society.In all relevant road safety data that affect road safety,small-scale data has a very important place,like the size and height of the sign and the width of the road.Measurement of these small scale data is still in the manual measuring phase,which is inefficient and dangerous and cannot satisfy the needs of modern road safety data.Achieving a rapid,efficient and highly automated measurement of the small-scale road safety data has important social and economic significance to the safety of road.This paper studies the application of binocular vision technology in road safety data detection,based on binocular vision technology and vehicle-camera rapid acquisition technology.When using binocular vision technology and vehicle-camera to collect images to detect road safety data,the measurement range needs to cover a large depth range and field size.In this application scenario,the conventional binocular vision measurement model has unstable measurement accuracy,and the measurement error in some depth is obvious.Based on the analysis of the causes of error,this paper proposes a calibration method based on depth distribution to improve the measurement accuracy and stability of binocular vision system in the whole scene.In addition,this paper improves the other deficiencies of the conventional binocular vision measurement method in the large and deep view field,and designs a set of binocular vision measurement system applicable to the scene.The system realizes high precision and high automatic detection of road safety data.After doing researches,this paper gets the following results:(1)Studying and analyzing principle and process of binocular vision technology,putting forward the shortcomings and limitations of conventional binocular vision measurement method in road safety data detection scenarios.(2)Aiming at the problem of uncertainty of measurement accuracy in the road safety data detection scenario which has large depth and deep field,establishing a binocular vision measurement error transfer model,and proposing a calibration method based on thedepth distribution on the basis of the propagation rule of binocular vision measurement error.The method achieves the accurate and stable measurement of binocular vision system in the scene with large depth and deep field.(3)Realizing a set of binocular vision measurement system after optimize design which can be used for road safety small scale data detection,based on depth distribution calibration method.(4)Verifying the stability and accuracy of the system through static and dynamic experiments,experimental data shows that the system has increased the automation of the road safety data detection,and has increased the efficiency and accuracy,the average error is less than 1.1%.