Image Processing-based Automobile Headlight Calibration System Research

In recent years, with the general increase of the car ownership, traffic safety has become one of the problems of concern. Headlight orientation as a seized item of headlight testing, with an improper direction, will become one of the main risks that affect the vehicle safety. In view of the existing detection apparatus and methods for illumination direction detection limited in a specific environment and equipments requirements, complex operation, and high detection cost, cannot meet the increasing needs of automobile safety detection. This paper presents a headlight illumination direction detection method that based on the image processing technology, and designs an automatic headlight illumination direction calibration system.Firstly, the headlight projection image processing, and presents a headlight segmentation algorithm to obtain the headlight region of the image. Then presents two the feature detection methods to detect the headlight regional image:respectively the designed centroid detection algorithm of the high beam and the light-dark cut-off line inflection point detection algorithm of the low beam. And, proposed a method for fitting straight line correction in inflection point detection algorithm, which use for fitting the light-dark cut-off line to make sure the exact location of the inflection point.Secondly, establish the calibration model, which using the detected headlight feature and the camera imaging principle, to calculate the headlight illumination direction angle. Based on the different environment, propose two methods that there are the wall-based headlight calibration method and the road-based headlight calibration method to calibrate headlight, which can calibrate the high beam and low beam of the headlight, to obtain the illumination direction of horizontal and vertical deflection angle. And use the simulation experiment for error analysis of the potential error factors in calibration.Finally, based on feature detection result and the calibration model, design an automatic headlight irradiation calibration system. And using the system, the driver can complete headlight irradiation calibration automatically. At the same time in order to improve the stability of system,this paper presenting the multi-frame statistics method to obtain stable detection feature coordinates, and also setting the calibration result evaluation method to overcome the disturbance of large information system calibration results, improve the consistency and accuracy of the system.This automobile headlight calibration system is not rigidly adhere to the requirements of the specific equipments and environment like the vehicle inspection stations, and saves testing cost, operates simple, at the same time provides a feasible way for the intelligent detection of headlight orientation.