Electric vehicle identification and elevator protection method based on machine vision

The invention provides an electric vehicle identification and elevator protection method based on machine vision, which comprises the following steps of: continuously acquiring, identifying and reminding, carrying a trained electric vehicle part identification model in a background server, extracting image features through a feature extraction neural network, and then carrying out nonlinear classification through a full connection layer. If it is judged that someone tries to push in the electric vehicle, reminding of forbidding pushing in is carried out; and if the user does not listen to the persuasion, the elevator stops working. According to the method, recognition of the whole electric vehicle is replaced with recognition of local parts of the electric vehicle, the data processing amount can be reduced, the recognition speed and precision are improved, and the elevator protection effect is correspondingly improved. 本发明提供了基于机器视觉的电动车识别及电梯防护方法，持续采集--识别--提醒，在后台服务器中搭载训练好的电动车部件识别模型，通过特征提取神经网络进行图像特征的提取，然后通过全连接层进