“Back to Basics” is a weekly feature that highlights important but potentially overlooked information that EHS professionals should know. This week we’ll be looking into: Use computer vision to identify and mitigate the causes of workplace accidents.
It’s difficult to discuss environment, health, and safety (EHS) technology without mentioning artificial intelligence (AI). One of the promising aspects of AI is computer vision. Computer vision combines cameras with AI and machine learning tools to identify and communicate the root cause of workplace incidents.
As part of its Work to Zero initiative, the National Safety Council (NSC) published a white paper in 2022, “Using Computer Vision as a Risk Mitigation Tool,” which explored the role of this technology in safety. This study investigated four computer vision systems and evaluated their ability to identify risk, personal protective equipment, and workplace violence.
“This report reveals that computer vision technology, combined with advanced risk prediction algorithms, can accurately and consistently automatically monitor worker personal protective equipment (PPE), risk identification, and workplace violence and weapons detection,” according to the executive summary. “Additionally, this technology can be used to monitor fatigue, material impairment, and other impairment conditions during driving. Comprehensive software is available and can be easily deployed in a variety of industrial environments, from construction to warehouses to manufacturing, capturing existing closed-circuit television (CCTV) feeds and providing an intuitive and actionable dashboard for safety leaders.”
Barriers to the adoption of computer vision technology include pricing and privacy concerns, but the report notes that an increase in the number of systems and packages on the market could lead to lower prices. Additionally, many systems allow users to anonymize employee information and likenesses to enhance privacy.
What is it?
Computer vision is used in security and workplace safety applications and utilizes manually operated CCTV systems. This requires someone to sift through hours of images to search for incidents and decide how to prevent them in the future.
With AI, computer vision systems can be trained to identify and communicate the root cause of workplace incidents. Identify critical objects that can lead to accidents, such as working at heights, moving machinery, and unstable items on shelves, as well as objects that can help prevent injuries, such as hard hats, high-visibility vests, and other personal protective equipment. The system can monitor multiple workers and assess safety threats.
According to the NSC, computer vision systems can learn worker habits and understand what situations lead to accidents, allowing for understanding, training, and observation to prevent future accidents. Computer vision is ideal for industries that regularly move materials or use heavy machinery, such as manufacturing, logistics, construction sites, and industrial warehouses.
Some systems can be trained to recognize best practices and use them as a reference to assist in employee training and highlight deviations from best practices. It can also be useful in emergency situations by detecting when a person remains in a location for an unusual amount of time and alerting operators.
Restrictions
Video quality can be an issue with many CCTV networks, providing grainy, blurry video feeds that can make it difficult for AI software to identify things like PPE or track objects. However, the image quality has improved.
Some systems may not be able to distinguish between situations such as conversations between employees that lead to hugs or arguments that lead to physical fights, the report said.
Additionally, because the camera has a limited field of view, computer vision systems may not be able to obtain a complete view of the workspace.
ethical concerns
To address ethical concerns about the use of AI, the American Society of Safety Professionals (ASSP) endorsed a set of guiding principles for AI in 2024.
- Trust: AI enhances the skills of occupational safety and health (OSH) professionals, not replaces human judgment and decision-making. Occupational health and safety professionals must oversee AI-driven occupational safety and health solutions and hazard remediation to ensure decisions are made that consider context, ethics, and exposure.
- Transparency: Workers, managers, and leaders need to be informed about the capabilities and limitations of AI technologies used in the work environment.
- Equity: Occupational health and safety professionals must ensure that AI technologies do not exacerbate existing disparities or introduce new forms of discrimination in workplace safety and health-related practices.
- Privacy: Organizations must put safeguards in place to prevent unauthorized access, misuse, and abuse of sensitive information collected by AI systems.
According to ASSP President Pamela Walasky, CSP, FASSP, Defining the Role of AI in Safety, “Thoughtful and responsible adoption of AI can harness its potential to improve workplace safety, protect workers, and create healthier working environments.” “However, we also know that the use of AI comes with challenges, such as privacy risks and the need to address new hazards.
