Intelligent analysis method for train fault image based on lightweight deep learning technology

The invention discloses a train vehicle fault image intelligent analysis method based on a lightweight deep learning technology, and the method comprises the steps: obtaining train image data, carrying out the marking of a to-be-detected train fault part in an image, and dividing the overall data into a training set and a test set; carrying out data enhancement operation, configuring hyper-parameters required by an algorithm, and inputting a training sample into a MobileDetectNet neural network model formed after improvement to carry out feature learning; and carrying out fault identification on the feature model obtained by inputting the new passing vehicle image into the learned feature model, and finally framing out a fault region and outputting corresponding alarm information. The MobileDetectNet network intelligent identification model is few in parameters, small in occupied memory and high in identification rate, the identification rate of motor train unit fault detection under the original equipment CP