Research On Pedestrian Ranging Technology Based On Vehicle Binocular Camera

In recent years,people’s living conditions are getting better and better,in the rapid economic development today,cars,gradually into human life,to provide great transportation convenience for human beings.However,when people enjoy the convenience of transportation,they also bear a lot of risks.Too fast speed and improper driving behavior of cars will lead to dangers and cause great harm.Therefore,the intelligent transportation system emerges at the right moment.In this paper,through reading a lot of information to understand the vehicle environment perception,on the basis of comparing the advantages and disadvantages of various vehicle environment perception sensors,the pedestrian ranging system based on vehicle binocular camera is proposed.According to the image quality,according to the actual situation of binocular image preprocessing,image for low light shooting environment,the acquisition of image luminance details,low loss and contains noise,etc.,for image enhancement,an improved adaptive correction method for the gamma function,correcting image of light weight,Calibration parameters are used to correct the low brightness image to remove the noise and improve the ranging accuracy.In order to improve the efficiency of the whole process of distance,without damage to the precision of matching,to reduce the use of time,this paper build correct mapping table,stereo calibration steps before giving up stereo matching,matching the original image directly,after the objective functions of left pedestrian area location,using BM stereo matching algorithm to match the position of the image on the right side of the traveler,By using the corrected mapping table,the coordinate after corrected distortion is directly obtained,and the parallax is obtained.Finally,the depth distance of the pedestrian target is obtained according to the binocular ranging principle.This paper respectively in the laboratory environment and vehicle-mounted binocular camera under the condition of the pedestrian ranging system was tested and the experimental results show that the pretreatment method is simple and effective,can effectively remove noise,reduce the distance measurement error,error reduced obviously in 3 to 30 m ranging experiments,maximum relative error is 4.0%,has practical significance.It can effectively reduce the matching time by constructing the correction mapping table.Compared with the normal BM matching algorithm,the matching time can be reduced by 0.05s,and the ranging efficiency can be improved.