Vehicle model identification method based on deep residual network

The invention discloses a vehicle model identification method based on a deep residual network and relates to the machine vision field. Under different backgrounds, a vehicle characteristic can be extracted and a vehicle model can be identified. The method comprises the following steps of carrying out parameter initialization on the deep residual network; loading training data into the deep residual network, and combining a data enhancement strategy to train and acquire the trained deep residual network; and loading a vehicle model image into the trained deep residual network, identifying thetrained deep residual network and outputting a type label from a network output terminal. 本发明公开了种基于深度残差网络的车型识别方法，涉及计算机视觉领域，能够在不同背景下提取车辆特征，进行车型识别。本发明包括：对深度残差网络进行参数初始化；将训练数据载入深度残差网络，结合数据增强策略，训练得到训练好的深度残差网络；将车型图像载入训练好的深度残差网络，训练好的深度残差网络进行识别，并从网络输出端输出类别标签。