Subject | Engineering
Practical Training of Deep Learning
- Machine learning
- Spectral analysis
- Python
Lecture 01: Basic of Image Processing
Lecture 02: Pattern Recognition and Machine Learning
Lecture 03: Hyperspectral Imaging
Lecture 04: Python Tutorial; Grammar
Lecture 05: Python Tutorial: Numpy
Lecture 06: Python Tutorial; Matplotlib
Lecture 07: Python Tutorial; Pandas
Lecture 08: Python Tutorial: Scikit-learn
Lecture 09: Machine Learning Hands-on
Lecture 10: Gas and Dust Control for Improvement of Underground Work Environment
Lecture 11: Communication Systems in Underground Mines
Lecture 12: Environmentally Harmonious Mining System Utilizing Knowledge from Inactive and Abandoned Mine Management
Lecture 13: Student Presentations (1)
Lecture 14: Student Presentations (2)
Lecture 15: Natural Resources and International Relations
Content/学習内容
-
Basic of Image Processing
- Image Compression
This lecture covers the basics of image processing. It explains analog vs. digital, A/D conversion, resolution, bit depth, pixels, and color spaces. Next, it introduces image compression methods (lossless/lossy) and key formats (BMP, JPEG, PNG, TIFF). Finally, it covers pattern recognition and machine learning, discussing feature extraction, distance calculation, and classification, with examples in character recognition and image classification.
Videos
/学習動画
-
What are Digital and Analog?
This lecture covers the basics of image processing, including analog vs. digital, A/D conversion, resolution, bits & bytes, pixels, and color spaces (RGB, CMY), explaining the structure of digital images.
-
What is Digital Image?
This lecture covers image data size, resolution, color depth, compression methods (lossless/lossy), and key formats (BMP, JPEG, PNG, TIFF), emphasizing the importance of efficient digital image processing.
-
Introduction of Pattern Recognition and Machine Learning
This lecture covers pattern recognition and machine learning, explaining feature extraction, distance calculation, and classification with examples like letters and animals, highlighting the difference from simple search.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Pattern Recognition and Machine Learning
- Machine learning
- Neural network
This lecture covers pattern recognition and machine learning, explaining distance-based classification, supervised vs. unsupervised learning, kNN, neural networks, deep learning, CNNs, transformers, and applications in hyperspectral imaging.
Videos
/学習動画
-
Pattern Recognition
This lecture introduces pattern recognition basics, classification via distance calculation, and machine learning, highlighting feature extraction and different distance metrics.
-
Machine Learning (1)
This lecture covers nearest neighbor methods, supervised vs. unsupervised learning, classification vs. regression, and applications of machine learning.
-
Machine Learning (2)
This lecture explains neural networks, deep learning, CNNs, and transformers, highlighting advances in computing power and development environments that drove their widespread adoption.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Hyperspectral Imaging
- Feature Extraction
- Machine Learning
This lecture covers hyperspectral imaging fundamentals, data processing, and machine learning techniques. It explains differences from RGB, spectral absorption, PCA, K-means, and SVM, with practical Python exercises. Applications include ore grade estimation, mineral identification, and environmental monitoring. The course discusses hyperspectral imaging’s role in assessing and recovering mining-affected areas. Students apply learned techniques through assignments, enhancing practical skills.
Videos
/学習動画
-
Introduction of Hyperspectral Imaging
This lecture introduces hyperspectral imaging, explaining differences from RGB, spectral features, and applications in mineral exploration and environmental monitoring, emphasizing the importance of light-matter interactions like reflection and absorption in analysis.
-
Getting Ready for Python Programming and Machine Learning
This lecture introduces Python basics for hyperspectral data analysis, covering light absorption, preprocessing, feature extraction, PCA, K-means, and SVM, along with practical visualization and classification techniques.
-
Machine Learning for Hyperspectral Images
This lecture reviews hyperspectral image processing, introduces classification and ore grade estimation tasks, discusses applications in mining damage detection and environmental recovery, and explores the use of satellite data.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Python Tutorial; Grammar
- Programming
This class provides the fundamental grammatical knowledge of Python, a programming language widely used for deep learning, through exercises.
Videos
/学習動画
-
Python grammar; Data Type
The data type of variables in Python is explained.
-
Python Grammar; Statements, Functions
Several flow control statements, “if” statements and “for” statements, and functions are explained.
-
Excercise
Learn practical programming through exercises
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Python Tutorial: Numpy
- Numpy
This class provides instructions on how to use a Python library called “Numpy” to operate arrays.
Videos
/学習動画
-
Array Creation
Create an array with Numpy.
-
Array Operations
Operate an array with Numpy
-
Numpy Functions
Use Numpy functions for array calculations
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Python Tutorial; Matplotlib
- Matplotlib
This class provides data visualization techniques with Matplotlib.
Videos
/学習動画
-
Basic Graph Visualization
Visualize line graph, scatter plot, histogram, and heat map with Matplotlib.
-
Other Graphs
Visualize other types of graphs
-
Exercise
Learn Matplotlib through exercise.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Python Tutorial; Pandas
- Pandas
This class provides the use of Pandas to analyse table data.
Videos
/学習動画
-
Basic operation of Pandas
Operate table data with Pandas.
-
Data Analysis with Pandas
Visualiza the data by calculating statistical metrics with Pandas, and get insight from the data
-
Exercise
Learn Pandas through Exercise
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Python Tutorial: Scikit-learn
- Scikit-learn
- Machine Learning
This course offers the use of Scikit-learn to develop a machine learning model.
Videos
/学習動画
-
Machine learning model development workflow
Introduce a common workflow to develop machine learning model
-
Model development with Scikit-learn
Develop a machine learning model with Scikit-learn
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Machine Learning Hands-on
- Mining
This course provides the following content. Basic knowledge of machine learning, how to develop a machine learning model for the mining industry.
Videos
/学習動画
-
Machine Learning and the Mining Industry
Explain the relationship between Machine learning and the mining indstury.
-
Excercise
Develop a machine learning model using the minining related dataset.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
Gas and Dust Control for Improvement of Underground Work Environment
- Dust Collector
- Dust Concentration Measurement
This lecture explains the importance of dust control in underground work environments and introduces specific countermeasures. It covers the mechanisms and applications of various dust collectors, including gravity, cyclone, filter, wet scrubber, and electrostatic precipitators, and discusses selection methods based on cost and efficiency. The lecture also compares dust concentration measurement methods, such as the sampler method and light scattering method, and introduces personal samplers for monitoring workers’ dust exposure. A quiz is conducted at the end to assess understanding.
Videos
/学習動画
-
Introduction of Gas and Dust Control in Underground Mines
This lecture explains dust control in underground work environments. It covers the characteristics and applications of various dust collectors, including gravity, cyclone, filter, wet scrubber, and electrostatic precipitators. It also introduces dust concentration measurement methods, such as the sampler method and light scattering method. Finally, a quiz is conducted to reinforce students’ understanding.
Lecturers
/講師
-
Yuichi Sugai
Professor, Faculty of Engineering Department of Earth Resources Engineering, Kyushu University
-
Communication Systems in Underground Mines
- Wi-Fi Direct
- Power Line Communication (PLC)
The lecture provides an overview of communication systems in underground mines. The lecturer first introduces his background and shares insights from his visits to mines in South Africa and other locations. He explains the importance of communication in underground mining and highlights the challenges, such as signal attenuation and the constraints of narrow tunnels. Traditional wired communication and leaky feeder systems are discussed, along with their limitations. He then presents his research on a cost-effective communication system that integrates Wi-Fi Direct and power line communication. The lecture concludes with experimental results and an analysis of its cost-reduction benefits.
Videos
/学習動画
-
Introduction of Communication Systems in Underground Mines
The lecture introduces communication systems in underground mines. After presenting his background, the lecturer discusses mining site visits, emphasizing the importance and challenges of underground communication. He introduces his research on a cost-effective hybrid communication system using Wi-Fi Direct and power line communication.
Lecturers
/講師
-
Hajime Ikeda
Associate Professor, Department of Systems, Control and Information Engineering, Asahikawa National College of Technology
-
Environmentally Harmonious Mining System Utilizing Knowledge from Inactive and Abandoned Mine Management
- Digital Twin
- Environmentally Harmonious Mining
This lecture introduces “Smart Mining Plus,” an environmentally harmonious mining system using knowledge from Japan’s inactive mine management. AI, digital twins, and multimodal sensing are applied to minimize environmental impact and ensure sustainability. A case study focuses on Kazakhstan’s mines, addressing mine drainage and pollution control. The lecture also explores optimizing mining operations and human resource development to balance resource extraction and decarbonization goals for a sustainable future.
Videos
/学習動画
-
Introduction of SATREPS Project for Kazakhstan
This lecture introduces “Smart Mining Plus,” an environmentally harmonious mining system utilizing AI and digital twins to reduce environmental impact and ensure sustainable mining using knowledge from Japan’s inactive mine management.
Lecturers
/講師
-
Youhei Kawamura
Professor, Division of Engineering, Hokkaido University
-
Student Presentations (1)
Student presentations toward assignments
Videos
/学習動画
-
Application of Machine Learning on the Classification of Bushveld Geochemical Data of Chromite Layers in South Africa
Geochemical data of chromite layers in South Africa’s Bushveld Complex were classified using machine learning. Python and Random Forest were utilized, achieving 69% accuracy. Removing outliers and expanding data are key to improving accuracy.
-
Machine Learning for Geochemical Classification of Chromitite
Machine learning was applied for geochemical classification of chromitite. After data processing, SVM, decision trees, and Random Forest were compared. Random Forest achieved the highest accuracy (92%), suggesting its potential to enhance geological exploration efficiency.
-
Iron Ore Grade Estimation with Hyperspectral Imaging (HIS)
Hyperspectral imaging was used to evaluate iron ore grade. K-means and SVM were compared, with SVM achieving high accuracy (98.6%). This method is useful for ore classification and reducing environmental impact, improving exploration and beneficiation processes.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
-
Student Presentations (2)
Student presentations toward assignments
Videos
/学習動画
-
Assignment 1: Sample Classification, Assignment 2: Iron Ore Grade Estimation
Iron ore grade estimation and sample classification were conducted using hyperspectral imaging. K-means and SVM were compared, with SVM achieving high accuracy (99.6%). Selecting optimal parameters is crucial for improving accuracy, enhancing ore evaluation efficiency.
-
Using Machine Learning for PPV Estimation with Blasting Data
Machine learning predicts peak particle velocity (PPV) from blasting data. Stacking regressor performed best (R² = 72.3%), aiding environmental impact assessment and blast planning accuracy.
Lecturers
/講師
-
TORIYA Hisatoshi
Associate Professor, Graduate School of International Resource Sciences, Akita University
-
OWADA Narihiro
Graduate School of International Resource Sciences, Akita University
-
-
Natural Resources and International Relations
- Energy Security
- Belt and Road Initiative
- Renewable Energy Transition
- Resource Dependency Risk
This lecture, Natural Resources and International Relations, explores resource-related issues in international relations, focusing on geopolitics and energy security. The first session introduced fundamental geopolitical concepts, historical contexts, and theories by Mackinder and Spykman, analyzing the evolution of international relations concerning energy resources. The second session focused on Central Asia’s energy geopolitics, discussing China’s Belt and Road Initiative, resource investments, and regional energy security. Additionally, it examined the growing importance of mineral resources in the renewable energy transition and Japan’s risk of dependency on China for resource procurement.
Videos
/学習動画
-
Natural Resources and International Relations (1)
In the first lecture of “Natural Resources and International Relations,” the instructor introduces himself and his research areas, followed by an explanation of fundamental concepts in geopolitics and energy security. The lecture covers the historical background of geopolitics, theories by Mackinder and Spykman, and the evolution of international relations concerning energy resources. Additionally, it discusses Japan’s energy security challenges and the geopolitical significance of the Middle East.
-
Natural Resources and International Relations (2)
The second lecture focuses on energy geopolitics and security in Central Asia. It discusses regional order shifts after the Cold War, the impact of China’s Belt and Road Initiative, and China’s economic and energy investments in Central Asian countries. Additionally, the increasing importance of mineral resources for renewable energy transition and Japan’s risk of dependency on China are examined.
Lecturers
/講師
-
Fumiaki Inagaki
Professor, Graduate School of International Resource Sciences, AkitaUniversity
Staff/スタッフ
-
TORIYA HisatoshiAkita University Graduate School of International Resource SciencesAssociate Professor
-
OWADA NarihiroAkita University Graduate School of International Resource Sciences
-
Yuichi SugaiKyushu University Faculty of Engineering Department of Earth Resources EngineeringProfessor
-
Hajime IkedaAsahikawa National College of Technology Department of Systems, Control and Information EngineeringAssociate Professor
-
Youhei KawamuraHokkaido University Division of EngineeringProfessor
-
Fumiaki InagakiAkitaUniversity Graduate School of International Resource SciencesProfessor
Competency/コンピテンシー
Course Objectives
This course systematically covers the fundamental concepts and applications of deep learning. Students will understand the basics of image processing, pattern recognition, and machine learning while acquiring data processing and analysis skills using Python. Additionally, through case studies on deep learning applications in the mining industry, students will develop practical problem-solving skills. Ultimately, the course aims to enable students to apply deep learning, set their own challenges, select appropriate methods, and analyze and solve problems effectively.
Learning Outcomes
Students will gain an understanding of the fundamentals of image processing and machine learning while acquiring data analysis techniques using Python. They will also learn about hyperspectral imaging and application technologies related to the mining environment, developing the ability to tackle practical challenges. Through lectures and exercises, students will comprehend the process of problem-solving using deep learning and, ultimately, acquire the capability to apply these skills in their own research or professional work.
Contact/お問合せ先
Akita University, Graduate School of International Resource Sciences, International Strategy Division
kokusaisenryaku@jimu.akita-u.ac.jp






