14-04-2023 дата публикации
Номер: CN115965598A
Принадлежит:
The invention discloses an aviation rivet classification and anomaly detection method based on deep learning, and the method comprises the following steps: 1, collecting aviation rivet image data, and carrying out the preprocessing of the image data; step 2, performing target frame category labeling and data enhancement processing on the aviation rivet image data; 3, training the aviation rivet classification and anomaly detection model by adopting a migration training and freezing training mode; and step 4, using a convolutional neural network taking a DarkNet-53 structure as a baseline as a trunk network to extract features to perform neural network reasoning, using a Focus layer, a cross-stage local network layer, a spatial pyramid pooling structure module, an up-sampling layer and a connection layer as a neck network to perform feature fusion, and finally using a decoupling detection head to generate a detection result. Compared with the prior art, the method has the positive effects ...
Подробнее