Opening a New World of Non-Destructive Testing with Hyperspectral Technology
HSI (Hyperspectral Imaging) is a non-destructive testing method that utilizes hyperspectral technology to precisely diagnose minute defects, material distribution, and surface conditions of metal equipment. In particular, by analyzing the condition of critical components, such as fuel cladding tubes, using multi-spectral data, it maximizes the safety and performance of equipment. This, in turn, extends the lifespan of the equipment, prevents unexpected accidents, and enables efficient equipment management.
Vision
We open a New Era of Non-Destructive Testing with Hyperspectral Technology
We are expanding the possibilities of non-destructive testing through Hyperspectral Imaging (HSI) technology, setting a new standard for equipment diagnostics and maintenance.By leveraging AI modeling and hyperspectral imaging techniques, we precisely analyze diverse wavelength data to maximize equipment safety and performance.Moving forward, we will apply hyperspectral technology to various fields such as smart maintenance, autonomous inspection systems, and environmental and material monitoring, aiming to become a global leader in non-destructive testing technology.
R&D
AI-Based Automatic Defect Detection
By leveraging AI algorithms, hyperspectral data is used to detect microcracks, which are then automatically classified by AI.
Key Research Areas
1. Development of a Deep Learning-Based Defect Classification Model
• Defect Detection by Identifying Specific Spectral Signals in Defective Areas
• Automation and Real-Time Inspection System
• Classification of Various Defects, Including Cracks, Corrosion, and Surface Anomalies
2. Multi-Band Spectrum Integrated Analysis
• Comprehensive analysis of defect characteristics across multiple wavelength bands based on data collected from various ranges, including VNIR, NIR, and SWIR.
• Compensation for variations caused by external environmental conditions, such as light source and distance.
3. Development of Automated and Real-Time Inspection Systems
• Design of High-Speed Processing Algorithms for Real-Time On-Site Analysis.
Analysis of Deterioration and Deformation in Coating Materials
Using hyperspectral data, damage to coating materials is analyzed, and changes in the physical properties of the material are detected through an AI model.
Key Research Areas
1. Deterioration and Deformation Detection
• Detecting early signs of material changes by identifying deterioration and deformation signals after spectrum-based characteristic analysis.
• Improvement of data quality using preprocessing techniques such as SNV, MSC, and Smoothing.
2. Generation of Customized Predictive Reports for Each Facility
• Automatic classification of deterioration and deformation states using CNN and Auto-Encoder models.
• Detection of new deterioration patterns in unstructured data using clustering techniques.
Achievements
Selected for Leading Nuclear Enterprise Program
We have been selected for a leading nuclear energy enterprise project with our technology that leverages AI algorithms to accurately detect and classify defects in hyperspectral images. By successfully completing a Proof of Concept (PoC) for defect detection in fuel cladding tubes, we are contributing to the development of key technologies in the nuclear energy industry.
Acquired a Total of Three Certifications
We are researching technologies to process and analyze large-scale inspection data in real-time, enhancing inspection speed.