1A2-D03 Development of a Polarized Microscpic Image Management System for Supporting Asbestos Qualitative Analysis utilizing Dispersion Staining

Kuniaki Kawabata, Hirokazu Yamazaki, Yugo Tsubota, Taketoshi Mishima, Kazuhiro Hotta, Hajime Asama, Hiroshi Mizoguchi, Haruhisa TakahashiThe Proceedings of JSME Annual Conference on Robotics and Mechatronics (Robomec) 2009 · DOI

Abstract

This paper describes the development of an automated polarized microscopic imaging system designed to support the qualitative analysis of asbestos. Dispersion staining, a visual observation method, requires experts to manually count thousands of particles and fibers across multiple samples. To alleviate this labor-intensive burden, the proposed system automates image acquisition using a motorized polarizer and stage, registering high-resolution images into a database via an online process. A prototype system was developed and evaluated, demonstrating significant improvements in efficiency and data management for asbestos analysis.

Introduction: The Challenge of Asbestos Analysis

Asbestos identification in building materials is a critical public health task. The standard qualitative analysis method, dispersion staining, involves immersing samples in liquids with specific refractive indices. Under a polarized microscope, asbestos fibers exhibit characteristic dispersion colors, allowing for identification.

However, the Japanese Industrial Standard (JIS) A 1481 mandates that experts count at least 1,000 particles per sample across three different immersion liquids, totaling over 3,000 particles. This manual process is not only time-consuming but also physically and cognitively demanding, leading to potential errors from fatigue. Furthermore, the lack of a permanent digital record makes secondary verification difficult.

Overview of the microscopic setup and dispersion staining visualization.

System Architecture and Methods

The developed system aims to automate the "search and capture" phase of analysis. It consists of three primary layers: automated acquisition, database management, and a graphical user interface (GUI).

The hardware setup utilizes an Olympus BX61 microscope with a motorized XYZ stage and a QImaging Retiga 4000R camera. The system automatically rotates the polarizer to four specific angles (0°, 45°, 90°, and 135°) to capture the necessary dispersion colors for each field of view.

System architecture for automated polarized image capturing and management.

A key technical innovation is the automatic cover glass detection. Because manual placement of the 22mm cover glass varies, the system employs image processing to detect the glass edges. By thresholding the red color component—which is prominent at the glass boundaries—the system automatically defines the scanning area, ensuring seamless image acquisition without manual boundary setting.

The management software, built on a MySQL database, monitors the acquisition in real-time. It automatically registers each image with its corresponding metadata, including sample ID, stage coordinates (x, y), and polarization angle. This creates a "digital twin" of the slide for retrospective review.

Key Results and Discussion

The prototype system was tested using a sample prepared according to JIS A 1481 standards. The system successfully captured a 15x15 grid (225 fields of view) at three polarization angles, totaling 675 images, in approximately 30 minutes.

The motorized stage demonstrated high precision with a positioning accuracy of ±1 μm. This accuracy is vital for the GUI, which provides a "bird's-eye view" by stitching adjacent fields into a 3x3 grid, allowing experts to browse the sample seamlessly. The successful synchronization of image files with the MySQL database proves the viability of a digital management system for specialized microscopic analysis.

Conclusion and Future Work

The current system serves as a powerful management and acquisition tool, acting as a "force multiplier" for asbestos experts. By handling the mechanical movement and data logging, it allows analysts to focus on the high-level task of identification.

Future developments will focus on:

  • Automated Counting: Integrating image processing algorithms to automatically detect and count asbestos fibers.
  • Clinical Validation: Assessing the reduction in physical and cognitive burden in real-world laboratory settings.
  • UI Refinement: Enhancing the GUI to better align with the specific decision-making workflows of asbestos analysts.

This research lays the groundwork for a fully automated diagnostic pipeline, improving both the efficiency and reliability of asbestos qualitative analysis.

Summary updated

Abstract

This paper describes to develop an automated polarized microscopic imaging system for supporting asbestos qualitative analysis. Dispersion staining is a visual observation method and it is done as that the experts count all particles in the view of the microscope and also the number of the fibrous asbestos fibres. For supporting this work, we are developing an automated system to ease the experts' burden for efficient observation. The system takes the images by the microscope with automated polarizer and registers them on the database via on-line process. A prototype system is developed and the experimental result is also discussed.
BibTeX
@inproceedings{kawabata2009d,
  author = {Kuniaki Kawabata and Hirokazu Yamazaki and Yugo Tsubota and Taketoshi Mishima and Kazuhiro Hotta and Hajime Asama and Hiroshi Mizoguchi and Haruhisa Takahashi},
  title = {{1A2-D03 Development of a Polarized Microscpic Image Management System for Supporting Asbestos Qualitative Analysis utilizing Dispersion Staining}},
  booktitle = {The Proceedings of JSME Annual Conference on Robotics and Mechatronics (Robomec)},
  year = {2009},
  doi = {10.1299/jsmermd.2009.\_1a2-d03\_1},
}
Figure