Research on Application of Image Enhancement Technology in Automatic Recognition of Rock Thin Section
Yuxuan Xu, Zongyang Dai, Yixin Luo
Abstract
Abstract Artificial intelligence technology has rapidly emerged in various new industries due to its high efficiency and has been successfully used in many fields. However, it has been slow to start in the field of petroleum exploration, under the background of the need for more efficient exploration and development in the petroleum field. In this paper we used the ResNet-18 convolutional neural network to make an attempt to automatically identify rock thin section, and finds that this method can efficiently identify rock thin section and has a higher accuracy rate. In addition, we adopted appropriate image enhancement technology, which can significantly improve the recognition accuracy of the model. It proves that related machine learning technology has broad application prospects in the fields of petroleum exploration and petroleum geology.