Browsing by Author "Ohtomo, Y"
Now showing 1 - 2 of 2
- Results Per Page
- Sort Options
- item: Conference-Full-textMineralogical classification and concentration estimation in mining with app using hyper-spectral imaging and machine learning(Division of Sustainable Resources Engineering, Hokkaido University, Japan, 2024) Okada, N; Takizawa, K; Wakae, S,; Ohtomo, Y; Kawamura, Y; Iresha, H; Elakneswaran, Y; Dassanayake, A; Jayawardena, CThis study presents an innovative method for identifying minerals by combining the capabilities of hyperspectral imaging with machine learning. Although hyperspectral images are challenging to process due to their extensive dimensions and substantial size, our solution effectively tackles this complexity by providing a user-friendly machine learning tool specifically tailored for hyperspectral data. This self-developed tool simplifies the process of constructing datasets and enhances machine learning processes for identifying mineral species and estimating their concentrations. The interface is designed to be easy to use, allowing non-experts to effectively identify minerals without needing professional expertise. This is further enhanced by the integration of machine learning capabilities. Our instrument is positioned as an innovative solution that greatly enhances geological surveys in mining regions, leading to useful outcomes for mineral-related research and industrial applications.
- item: Conference-Full-textPrediction of overbreak phenomenon in tunnel blasting using ORF index(Division of Sustainable Resources Engineering, Hokkaido University, Japan, 2024) Sato, N; Im, H; Nakazawa, Y; Jang, H; Ohtomo, Y; Kawamura, Y; Iresha, H; Elakneswaran, Y; Dassanayake, A; Jayawardena, CIn the drill and blast method, one of the most critical issues is overbreak. Overbreak leads to decreased work efficiency and increased operational costs, and it is recognized as a problem that needs to be addressed. Although several factors contributing to overbreak have been proposed, the specific parameters with the most significant impact are still unclear. However, it is evident that the geological conditions of the rock mass have a significant influence. In this study, the Overbreak Resistance Factor (ORF) was adopted to predict the occurrence of overbreak and to create an indicator for it. As a result, we were able to predict the occurrence and grasp the trends of overbreak. Geological data were collected from a mountain tunnel in Japan, and overbreak data were gathered from the 3D CG model of the tunnel face, which was constructed using Structure from Motion (SfM). Using these datasets, an overbreak prediction model using an Artificial Neural Network (ANN) was developed, and sensitivity analysis was performed to create an overbreak chart based on the influence of different parameters