Research on Rock Image Recognition Based on Deep Learning

The classification and identification of rock samples is an important part geological analysis. In order to reduce the cost prevent human subjective judgment from affecting results, use deep learning establish intelligent a new way. Based on convolutional neural network model AlexNet structure, migration for image set analysis established, automatic recognition lithology realized by using method. traditional method artificial has big drawbacks. method, this paper establishes suitable cuttings core industrial cameras at logging site. Due small number samples, data enhanced first. Then, feature extraction performed based structure such as network. Finally, compare analyze experimental results each model, obtain best lithology. Thereby reducing improving efficiency work.