According to the invention, whether a vehicle is about to drive out of a garage is detected through a pressure detection unit, three camera shooting devices are triggered to be started, the first camera shooting device and the second camera shooting device are used for shooting pedestrian conditions on the left side and the right side of a road adjacent to a garage exit respectively, and the thirdcamera shooting device is used for shooting vehicle conditions of the road inside the garage exit. Firstly, an image is subjected to deformation preprocessing, then the moving direction of a pedestrian and the distance relative to the garage exit are obtained through a pedestrian detection unit, and the moving direction and a license plate number of a vehicle are obtained through a vehicle detection unit. When the pedestrian detection unit detects that pedestrians exist on the two sides of the garage exit and move towards the garage exit at the same time and the pedestrians are located withina preset safe distance range, the system reminds a vehicle going to be driven out of the garage through an indicator lamp and a loudspeaker box at the garage exit and pushes a message to vehicle-mounted WiFi equipment of the vehicle at the same time, so that a vehicle owner is timely informed that a pedestrian is about to approach the garage exit, and the vehicle owner avoids timely.
本发明通过压力检测单元检测是否有车辆即将从车库中驶出并触发打开三个摄像装置,其中,第一和第二摄像装置分别用于拍摄与车库出口相邻道路左侧、右侧的行人状况,第三摄像装置用于拍摄车库出口内部道路的车辆状况。首先对图像进行变形预处理,然后,通过行人检测单元获得行人的移动方向和相对于车库出口的距离,通过车辆检测单元获得车辆的移动方向和车牌号。当行人检测单元检测到车库出口两侧有行人,同时该行人朝车库出口位置移动并且该行人位于预设的安全距离范围之内时,系统通过车库出口位置的指示灯和音箱对即将从车库中驶出的车辆进行提醒,同时向车辆的车载WiFi设备推送消息,以此及时告知车主有行人即将靠近车库出口,使车主及时避让。
Garage vehicle outgoing pedestrian safety prompting system based on convolutional neural network
基于卷积神经网络的车库出车行人安全提示系统
2021-03-30
Patent
Electronic Resource
Chinese
Garage departure safety prompting system based on convolutional neural network
European Patent Office | 2021
|Vehicle blind area pedestrian prompting method and vehicle
European Patent Office | 2023
|