基于单目测距和逆向路径规划的自动平行泊车

Automatic Parallel Parking Based on Distance Measurement of Monocular Vision and Converse Path Planning

根据生活实际的需要,文章提出一种基于单目测距和路径规划的自动泊车方法。首先确定泊车目标位置相对于车身的坐标,按照图像与实景比例在摄像机拍摄的图像中生成车身轮廓线框。然后驾驶员进行车身线框与泊车位的匹配,最后计算机实行逆向路径规划,采用碰撞分析,规划出最短泊车路径。此方法有效地解决了传统泊车系统利用超声波实时测距需要左右车位有车的局限性,并通过Matlab siumlink搭建模型进行运动仿真,检验了此方法的可行性和稳定性。

According to the actual needs of living, this paper puts forward a method for automatic parking which is based on monoculor vision distance and path planning.Firstly, the position of the parking target is determined relative to the coordinate of the car body, and the frame of the body contour is framed according to the image and the actual proportion in the image of the camera.Then the driver carries out the matching of the body line frame and the parking position.Finally, the computer performs the reverse path planning and adopts the collision analysis to plan the shortest parking path. By using this method, it effectively solves the traditional parking system using ultrasonic distance measuring real-time needs about the limitations of parking. By using Matlab siumlink, it constructs a model to perform dynamic simulation, so as to test the feasibility and stability of this method.