Research on visualization method of lithology intelligent recognition based on deep learning mine tunnel image

This paper presents an image processing and deep learning methodology for identifying different types of rock pictures. Pre-processing work such as rock image acquisition, gray scaling, Gaussian blurring, and feature dimensionality reduction is performed to extract useful feature information and to recognize and classify the rock images using TensorFlow-based CNN and PyQt5. The rock picture dataset is established and separated into the workout, confirmation, and test sets. The framework is compiled and trained. The categorization approach is evaluated using image data from the validation and test datasets, with key metrics like accuracy, precision, and recall being analyzed. Finally, classification model conducts a probabilistic analysis of the measured data to determine the equivalent lithology type for each image. The experimental results show that a method combining deep learning and TensorFlow-based CNN and PyQt5 to recognize and classify rock images has an accuracy rate of up to 98.8%, which might be successfully utilized for rock image recognition. The system can be extended to the field of geological exploration, mine engineering, and other rock and mineral resources development so that it can recognize rock samples more efficiently and accurately and match them with the intelligent design system of roof-helm flexible full support to effectively improve the reliability, and economy of the support scheme. Simultaneously, the system serves as a reference for the support design of other mining and underground space projects.