Background and Challenges
Asbestos was once widely used in construction materials for its excellent heat resistance, fire resistance, and durability. However, its use is currently prohibited as inhaling its dust poses health hazards. Meanwhile, many buildings containing asbestos still remain, and the risk of asbestos dispersion during demolition and removal work remains a social issue. In Japan in particular, the demolition of asbestos-containing buildings is expected to peak around 2030, and the number of monitoring activities aimed at preventing asbestos dispersion at demolition sites is expected to further increase.
The Japanese Ministry of the Environment’s Asbestos Monitoring Manual describes a method for confirming asbestos dispersion by collecting dust sampled around demolition sites on a filter and visually counting fibrous materials using a phase-contrast microscope (PCM). However, this analysis is reported to take approximately 25 to 100 minutes per filter, and is highly dependent on the analyst’s concentration and experience, which can potentially increase workload and lead to variability in judgment.
Services and Technology
Mekolas® is a support system that uses AI-based analysis to analyze PCM images of dust collected at demolition sites to efficiently and accurately confirm the presence or absence of asbestos dispersion.
AI-based analysis of dust images complements the fiber detection and fiber counting work traditionally performed by analysts, and simultaneously reduces labor in monitoring and improves judgment accuracy.
While the use of AI for asbestos counting is under international consideration, this is the first product in Japan for which accuracy has been validated using actual environmental samples.
* Mekolas® was developed through joint research by JANUS, the National Institute for Environmental Studies in Japan, and Environmental Control Center, Co., Ltd.
Key Features
- AI model that replicates the judgments of expert analysts
Equipped with an AI trained on the fiber counting results of expert analysts as training data, the system demonstrates high versatility even for actual environmental samples.- Fiber detection: recall 0.95
- Fiber counting: recall 0.85
(Verification results based on comparison with analysis results from five analytical laboratories)
- Practical analysis time
With a processing time of approximately one second per image, the system achieves an analysis speed equivalent to that of an expert analyst. The system supports early detection of asbestos dispersion and timely corrective decision-making. - Field-oriented user interface
- Enhanced visibility by displaying markers over AI-detected fibrous materials
- Supports batch processing of multiple images
- Displays analysis results lists categorized by image and individual fiber
- Supports CSV output and image output (including highlighted images)
- Judgment based on current manuals
Based on the Ministry of the Environment’s Asbestos Monitoring Manual, only fibrous materials that meet the standards for fiber length, fiber width, and aspect ratio are automatically evaluated, enabling seamless integration into existing measurement workflows.
Mekolas® User Interface & Features
* At present, Mekolas® is available only in the Japanese version. The interfaces shown in this document have been rendered in English for explanatory purposes.
Key applications
- Asbestos dispersion monitoring at demolition and removal work sites
- Labor saving and efficiency improvement in analysis operations
- Reduction of workload for expert analysts
- Support tool for education and human resource development
Mekolas® is designed not to replace analysts but to support them.
Future Developments
- Around the summer of 2026
Official launch of Mekolas® for static images (PCM images) is planned in Japan. - FY 2027
Market launch of a next generation version with real-time detection capability is planned in Japan.
In the future, we plan to explore meeting operational needs, including application to scanning electron microscope (SEM) images and polarized light microscope images, as well as integration with existing automated imaging software.
References
- Yamamoto, T., Iwasaki, K., Iida, Y., Yuki, K., Nakaji, F., Yamashiro, H., Toyoguchi, T. and Terazono, A. Rapid fiber-detection technique by artificial intelligence in phase-contrast microscope images of simulated atmospheric samples. Ann. Work Expos. Health (2024) 68, 420-426.
- Iida, Y., Yamamoto, T., Iwasaki, K., Yuki, K.-I., Kiri, K., Yamashiro, H., Toyoguchi, T. and Terazono, A. Development of a rapid fiber-detection system using artificial intelligence in phase-contrast microscope images of actual atmospheric samples. Front. Anal. Sci. (25 June 2025).