N1: Research Roundup: Big Data 2

Joseph Dartt, CIH Moderator
US DOL / OSHA
St. Louis, MO 
United States of America
 
Dr David Mendels Author
xRapid Environment
Henderson, NV 
United States of America
 
Thomas Hawkinson, MS, CIH, CSP Author
Stantec Consulting Services
Golden Valley, MN 
United States of America
 
Wed, 5/25: 1:00 PM - 2:00 PM CDT
Research Roundup 
Music City Center 
Room: Meeting Room 102 A 

Content Level

Intermediate

Topics

Also part of the Virtual Program
Available as part of AIHce OnDemand
Big Data
Computer/Mobile Apps and Tools
Hazard Recognition/Exposure Assessment
Sampling and Analysis

Presentations

Mold Spores Counting and Identification by Means of Artificial Intelligence — A Case Study

At AIHCE 2021, we presented a course that introduced the use of artificial intelligence (AI) in image treatment and analysis. AI was particularly applied to counting microscopic fibers in air samples, as per NIOSH 7400 diagnostic. While fibers are inherently simple objects for training a detector, the approach can be extended to more complex objects. This presentation will show how we extended the object detector model to generate a much larger dataset in order to identify mold spores at the genus level. We will demonstrate how combining several genus species was used to obtain representative accuracy levels. The issue of accuracy in object detection and nonregression testing will be discussed. While mold is the primary objective of this case study, the approach and principles are generic enough to apply to any other object detector. 

Co-Authors

G. Atkinson, xRapid Environment
H. Thomas, xRapid Environment 

Author

Dr David Mendels, xRapid Environment Henderson, NV 
United States of America

Using Cloud-Based Systems to Monitor Particles During a Health Care Construction Project

Most health care construction controls for particulates rely on pressure differentials to contain particulates. In this situation, the institution wanted to actively measure particulate levels to assure patient safety and provide real-time intervention if particulate levels rose too high. A maximum acceptable level of 10 µg/cubic meter was identified in the surgical suite. TSI DustTrak monitors for real-time data and a networking service were used to monitor the output and send alerts to monitoring staff when threshold levels were exceeded. This allowed a timely review of the situation by construction managers to assure that containment levels were secure and construction related particulate was not being released to critical patient care areas.

This method can be extended to other particulate and nonparticulate real-time data to manage critical exposure situations. Data management via a cloud application allows for review of the data and reporting to institutions or businesses and regulatory agencies. A discussion of the structure of the monitoring system, examples of reporting and some of the challenges faced will be discussed. 

Author

Thomas Hawkinson, MS, CIH, CSP, Stantec Consulting Services Golden Valley, MN 
United States of America