According to the invention, a preset number of continuous images are acquired in real time through an image acquisition unit, deformation processing is carried out on the images through an image processing unit, then feature extraction is carried out on the images through a convolutional neural network, the images are matched with various vehicle models obtained through training of the convolutional neural network to judge whether vehicles exist in the images, and if the vehicle exists in the image, the vehicle in the image is marked, and the driving direction of the vehicle, the distance relative to the garage exit, the moving speed and the prediction time for reaching the garage exit are calculated. If the prediction time for the vehicle to arrive at the garage exit is longer than the preset safety time, a prompting unit executes the garage exit safety prompting operation, and a microprocessor prompts pedestrians at the garage exit in one or more modes of adjusting the brightness, the flicker frequency, the color, the on-off state or the voice broadcast of the prompting lamp through the driving unit. Therefore, when a vehicle drives out of the garage, safety prompts are given topedestrians and vehicles on the two sides of the exit of the garage in time.
本发明通过影像获取单元实时获取预设数量的连续的图像,通过图像处理单元首先对图像进行变形处理,然后,通过卷积神经网络对图像进行特征提取并与卷积神经网络训练得到的各种车辆模型进行匹配判断图像中是否存在车辆,若图像中存在车辆,对图像中的车辆进行标记并计算车辆的行驶方向、相对于车库出口的距离、移动速度和到达车库出口的预测时间。若车辆到达车库出口的预测时间大于预设的安全时间,提示单元执行车库出车安全提示操作,微处理器通过驱动单元调整提示灯的亮度、闪烁频率、颜色、开关状态或者语音播报中的一种或者多种方式对车库出口位置的行人进行提示。以此,有车辆从车库中驶出时,及时对车库出口两侧的行人以及车辆进行安全提示。
Garage departure safety prompting system based on convolutional neural network
基于卷积神经网络的车库出车安全提示系统
2021-03-16
Patent
Electronic Resource
Chinese
Garage vehicle outgoing pedestrian safety prompting system based on convolutional neural network
European Patent Office | 2021
|