4:15 PM - 4:30 PM
[HRE12-04] Collaboration between Information Technology and Applied Geology - Automation identification method for minerals as an example
Keywords:Digital Transformation, Applied Geology, interdisciplinary fusion, Artificial Intelligence
The digital transformation (DX) of civil engineering is rapidly advancing as a national policy, and there are many needs for research and development through the fusion of different fields of applied geology and information engineering, and it is expected that engineers with an understanding of both fields will be trained and active in these fields. However, machine learning (AI), which has made remarkable progress and is active in various fields, has not yet been put to practical use in applied geology.
In recent years, research and development has been conducted to apply hyperspectral data, which are areal representations of reflection intensities in the visible light and near-infrared (400 nm to 1,000 nm) regions, to the field of geology. In particular, there have been many reports of research and development in mineralogy, where the distribution of minerals in borehole cores is contrasted with scanning electron microscope (SEM) data and hyperspectral data, and mineral identification is performed using machine learning (AI). The authors believe that this new method is also effective in applied geology, and have formed a team of engineers specializing in applied geology and information engineering to conduct research and development through interdisciplinary fusion.
The data acquired by the Hyperspectral Camera (HSC) was analyzed using a statistical method (Euclidean distance) and a neural network, a type of machine learning (AI) method, to determine the presence of turbidite in indoor experiments. The results were validated for the determination of turbidite content in laboratory experiments. The results of this study were published in the April 2023 issue of the Civil Engineering Journal.
In this presentation, through this research, I will describe the difficulties and interesting aspects of research and development through the fusion of different fields of applied geology and information engineering, as well as future challenges, from the perspective of a young employee who has been with the company for four years.
In recent years, research and development has been conducted to apply hyperspectral data, which are areal representations of reflection intensities in the visible light and near-infrared (400 nm to 1,000 nm) regions, to the field of geology. In particular, there have been many reports of research and development in mineralogy, where the distribution of minerals in borehole cores is contrasted with scanning electron microscope (SEM) data and hyperspectral data, and mineral identification is performed using machine learning (AI). The authors believe that this new method is also effective in applied geology, and have formed a team of engineers specializing in applied geology and information engineering to conduct research and development through interdisciplinary fusion.
The data acquired by the Hyperspectral Camera (HSC) was analyzed using a statistical method (Euclidean distance) and a neural network, a type of machine learning (AI) method, to determine the presence of turbidite in indoor experiments. The results were validated for the determination of turbidite content in laboratory experiments. The results of this study were published in the April 2023 issue of the Civil Engineering Journal.
In this presentation, through this research, I will describe the difficulties and interesting aspects of research and development through the fusion of different fields of applied geology and information engineering, as well as future challenges, from the perspective of a young employee who has been with the company for four years.