Study On Car’s A-pillar Blind Spot Elimination Algorithm

The A-pillar of the car is the supporting pillar connecting the cockpit windshield and the door to ensure the safety of the body structure.However,the position of the A-pillar causes a certain angle of obstruction to the driving vision,which affects the driver’s judgment of the road conditions and eliminates the blind spots can effectively prevent driving accidents caused by blind spots.In this regard,this paper proposes a vision system that eliminates blind spots in the A-pillar of the car through line-of-sight detection and field-of-view estimation,which can eliminate the blind spots of the A-pillar and compensate the visual field without affecting the car’s handling.Because the blind spots caused by the A-pillars are different for people,this paper proposes a method of using fixation points,which can calculate the corresponding blind spots according to the differences of people.Due to the particularity of the camera position mounted on the A-pillar of the car,this paper proposes a multi-camera calibration scheme for non-overlapping fields of view,using the mirror imaging principle and reasonable target position adjustment to achieve accurate calibration of multiple cameras.The methods proposed in this article have proved the feasibility and effectiveness of this vision system through targeted experiments,and provided new ideas for solving the blind spot problem of A-pillars.The research contents of this article are as follows:1.Firstly,we analyzed the occlusion of the car’s A-pillar during driving,and proposed a hypothesis that the relative position of the A-pillar and the driver’s head is constant during driving.The proposed algorithm uses two cameras to simultaneously capture the set field of view,and realizes the positioning of the A-pillar blind spot by registration,thereby realizing segmentation and field of view compensation for the blind spot.The experimental results show that the algorithm can achieve accurate A-pillar blind spot elimination and field of view compensation.2.Based on the analysis of the limitations of the above proposed algorithm,this paper designs an algorithm for estimating the A-pillar blind spot by using the gazing point.The algorithm calculates the blind spot by calculating the position of the gazing point by reconstructing the driver’s three-dimensional line of sight.Due to the similarity between the camera and human eye imaging principles,the line of sight detection can effectively deal with the difference of the driver’s occlusion area in the A-pillar blind spot,that is,the blind zone is calculated by the personnel’s ownparameters,and the blind spot in the true sense of the driver is obtained.Experiments show that the proposed method can effectively calculate blind spot based on personnel differences.3.The rigor of the above method is based on strict camera system calibration.We analyzed the particularity of the camera group position and could not use the traditional camera group calibration method.This paper designed a non-overlapping field camera group calibration scheme based on specular reflection.Experiments show that the calibration scheme can effectively deal with the calibration of non-overlapping field camera groups.