Intelligent image analysis method for anti-loose iron wire fracture faults of train vehicle

The invention discloses an intelligent image analysis method for anti-loose iron wire fracture faults of a train, and the method comprises the steps: 1, obtaining image data containing anti-loose iron wires on a train, marking the to-be-detected anti-loose iron wires in an image, and dividing the overall data into a training set, a verification set, and a test set; 2, performing data enhancement operation, configuring hyper-parameters such as a learning rate, a momentum gradient and weight attenuation required by an algorithm, and then inputting a training sample into an improved yolov7 model for feature learning; and step 3, inputting a new passing vehicle image into the learned feature model to carry out target identification, and finally framing out a fault area and outputting corresponding alarm information. According to the technical scheme, the performance of model training and reasoning can be balanced, the algorithm has the real-time performance and the stronger feature extraction capacity, the featur