Intelligent Rock Fracture Identification Based on Image Semantic Segmentation: Methodology and Application
D. D. Pan,Y. H. Li,C. J. Lin,X. T. Wang,Z. H. Xu
DOI: https://doi.org/10.1007/s12665-022-10705-1
IF: 2.8
2023-01-01
Environmental Earth Sciences
Abstract:In order to realize low-cost, fast, and accurate fracture identification in lining quality inspection and underground engineering stability evaluation, we propose an intelligent fracture identification method, which can achieve depth extraction of fracture information from rock images. Firstly, the fracture identification model combined with the residual network was constructed to obtain more fracture features. Secondly, dilated convolution layers were added to improve the resolution of the fracture feature points. Meanwhile, the perceptual field is expanded to achieve the effective capture of characteristics of tiny fractures. Lastly, the Rectified Linear Unit (ReLU) function is introduced to realize the nonlinear transformation of the model. And the predicted results are optimized by Dice loss function to accelerate the convergence of the model. The model is applied to identify concrete fractures and rock fractures. We also conduct reliability assessment and then compare the performance of our identification model to LinkNet, U-net, CNN + ASPP models. The maximum accuracy (98.93%), recall (83.78%) values of the method far exceed that CNN + ASPP model using the same dataset. High image similarity and smaller fracture features degrade the performance of the image identification method, while the method can significantly overcome these challenges and thus greatly improve the accuracy of fracture identification. And the identification method is more accurate and robust than the comprised method. Besides the proposed method with good generalization and can be used for the rapid identification of rock mass fractures in highways, constructions, slopes, water conservancies and other projects.