Research On Key Technologies Of Visual Identity Of Parking Behavior In The Interactive Area Of AGV Intelligent Parking System

At present,with the rapid growth of car parc in the country,the parking problem has become increasingly prominent.The AGV intelligent parking system can effectively alleviate the parking problem.However,the parking status of the interactive area of the system is various,which will affect the normal operation process and working efficiency of the automatic parking robot.The parking behavior does not affect the normal operation of the automatic parking robot is defined as the standard parking behavior,otherwise it is defined as parking violation.In order to better identify various parking behaviors effectively,and provide effective information for the automatic parking robot to carry the vehicles to meet the normal operation of the parking system,this thesis study several key technologies,which are following as:1.To obtain the angle and the position information of vehicle body after it is static.After the vehicle enters the interaction zone,the parking system firstly detects the vehicle,and judges its stationary state,and then it extracts and calculates the static vehicle body information,and acquires the oblique angle and the coordinates of the four vertices of the vehicle body in the interaction zone.2.To determine whether the parking behavior is standard or not,and remind the owner to correct it.The obtained information of oblique angle and the coordinates of the four vertices of the vehicle body are compared with the specified angle range and the specified boundary coordinates of standard parking.Judging by the comparison results.3.To get the coordinate information of the wheel after the vehicle is stationary.This thesis has designed an algorithm of the wheel position from a vertical view based on the vehicle side information,that is,it firstly extracts the wheel contour and the exterior vehicle body contour from the side view,and obtains the center and diameter of the wheel,the total length of the vehicle body and the distance from the wheel center to the front bumper of the vehicle.It then acquires the ratio of the distance and the total length of the vehicle body.Finally,it calculates the coordinates of the wheel based on the contour of the vehicle body in the vertical view.4.To judge whether the person leaves the interaction zone or not.The method of combining the HOG and the SVM and the background differencing are jointly applied for pedestrian recognition.Whether the minimum outer rectangle coordinates of the pedestrian contour are within the boundary range specified in the interaction zone can be treated as thejudging criteria.Finally,from the above mentioned,this thesis is designed under the VS2010 and OpenCV2.4.8 platforms,and uses the parking video data collected in the field for experimental verification.The experimental results show that this thesis can accurately visualize the parking behavior in the interactive area with the premise of guaranteeing the real-time of the algorithm,and meet the working requirements of the AGV intelligent parking system in the interactive area.