Design of Rock Sample Classification and Recognition Technology Based on Deep Learning

Abstract

In oil and gas exploration and mineral resource exploration, rock sample identification plays an immeasurable role due to its strong operability. Besides the methods of identifying rock samples by gravity, magnetism, remote sensing, electromagnetics, etc. a new way is to establish an automatic identification and classification model of rock samples by using image deep learning methods. This paper uses the transfer learning method to build a deep convolutional neural network model to realize the automatic classification and recognition of rock samples. First, the original data is preprocessed to ensure that it can be better used for model training. Then, based on the deep learning tool framework TensorFlow, this article uses the transfer learning method to migrate the Inception-V3 deep convolutional neural network model. To be trained in the preprocessed target data set, and improve the learning effect of the target task, so that more reasonable image features can be extracted, until it is trained into a neural network model in the target field, and the rock sample can be better classified.