国产麻豆

If your safety culture needs improvement, proactive video analytics can be an effective mechanism for change鈥攁nd protect people too.
By Barton Karto

Recycling and solid waste work is considered one of the most dangerous occupations in the U.S, being listed as high as the fourth most dangerous work environments, as measured in accidents and fatalities per worker. Naturally, organizations are eager to invest in methods that can improve these outcomes. Dedicated safety staff, policies, training, and PPE top the list for most companies. But what about technological investments that enhance safety?

Most organizations have video surveillance solutions installed on waste handling premises. A great majority of these are designed using the assumption that there is not enough manpower available to actively monitor them. Video surveillance has, thus, generally been used only after-the-fact, with safety staff reviewing footage forensically. Was that worker looking at his phone? Who parked that front end loader in the wrong place? Which vehicle was responsible for the collision? Knowing more about an incident that happened can improve future safety performance. In addition, recent innovations and cost improvements in video analytics can now help your facility avoid accidents before they happen.

Today鈥檚 video technology allows the video platform to shift from after-the-fact to proactive prevention of problems before they occur. Moreover, getting this right requires only modest capital outlays if your existing cameras are adequate. This article explores the value proposition of extending your investments in video surveillance platforms to include the new proactive safety features. We will cover some of the use cases currently available and consider the value and challenges to their deployment.

Technical Innovations
Two categories of technical advances are enabling a shift to proactive safety. Surveillance cameras are increasingly built with more powerful co-processors that allow them to do valuable work right inside the camera, rather than on the server. This change substantially reduces the cost and complexity of what we describe here.

Secondly, the rapid advances in the entire AI field have enabled video systems to be much more trainable. A system can be fed a series of images and later recognize similar objects and provide alerts on them. A common use case might be to feed a learning system images of an object type that commonly causes problems in the waste stream鈥攕uch as a propane cylinder鈥攁nd placing sensors near the entry points of the offending object into the stream. These systems are just emerging in the past few years at price points and ease-of-use levels that make more financial and operational sense.

Seeing Problems as they Happen
With these technical and cost improvements, what are some important things your surveillance system might do onsite?

PPE Detection
Cameras can very rapidly scan a person for whether they are properly wearing a helmet, ear protection, safety glasses, and so on. These cameras can be placed around the facility for continuous evaluation, and the system can send an alert when someone is out of compliance. Better still, a camera placed at the facility鈥檚 pedestrian entry door can lock the door until the worker meets facility safety criteria.

Flame Detection
Combusting li-ion batteries at a waste handling facility are an unavoidable risk. Using a combination of infrared video sensors, cameras can now quite reliably detect these hot spots. They measure the rate of temperature rise, the gases emitted by combustion, and can smartly differentiate between flames and the hot exhaust of a front-end loader.

Vehicle Safety Detection
Tipping floors are notoriously noisy and distracting environments, and new drivers, in particular, can put people at risk. Cameras can detect when vehicles move outside of the boundaries that you set or exceed your facility speed limit. They also can detect when drivers exit their vehicles in locations where they should remain safely in the truck.

Exclusion Zones
Cameras can sense when other objects are present in specific areas. This is useful for handling spills and maintains the operational flow of the waste handling facility.

Alerting
When an unsafe condition is met, people need to know. The safety and security system can be configured to send a great variety of alerts: beacon lights, messages on monitors, SMS and e-mail alerts, audible notifications, paging to handheld radios, and more.

Safety Investments
Most waste handling facilities have some sort of video surveillance system. The idea we present here is that the additional capital for proactive safety may be small compared to the overall capital and ownership costs. In other words, they are an incrementally small part of the cost of system deployment.

Figure 1 illustrates the overall capex for a large facility. Adding analytics software for these safety feature set鈥攖he bottom row鈥攊s only 16 percent of the total system cost. We suggest that that this additional capital outlay may produce some of the most valuable benefits these systems have to offer, by protecting the people who work at your facility.

Barriers to Progress
If the incremental cost for deployment is reasonable, why are these systems not being deployed more frequently? The dizzying pace of recent AI development means that commercialization of the full range of waste-specific applications is somewhat new. And they were expensive, especially when offsite third-party monitoring was included. Relatedly, older video systems would often indicate a false positive. Combined with scattered shortfalls in system configuration and tuning, the systems seemed to some to be too unreliable. Yet there are still challenges unrelated to cost and technology when deploying these additional features.

Proactive safety system design is significantly more challenging than a simple CCTV install. This usually requires close coordination with vendors who have expertise, and your internal decision makers to sort out which features to deploy. Key design criteria need thoughtful consideration. For example, device placement may be driven by which feature or features to deploy on a particular camera. And there are often limitations to how many features each device can manage, so it is not as simple as 鈥渢urn everything on for all the cameras.鈥� There is also a tradeoff to consider during initial system configuration and ongoing tuning. How many false positives are acceptable to avoid missing a dangerous incident? If the expectations are set too high, the result can be perceived as unsuccessful.
Lastly, cross-department collaboration is needed to determine which features make sense鈥攚ho will operate the system, and who will be responsible for responding to the alert information the system generates? Since there are so many alerting options, your team needs to spend time developing and maintaining SOPs and reviewing the period reports of safety violations.

Coordinating buy-in from and training for three shifts of onsite safety managers can also be a lift, and staffing turnover complicates the effort to adapt the entire team to the new tools. A strong safety culture eases the changes to the human element by framing the improvements as positive instead of punitive. Over time, your staff will more easily think of how the cameras can help.

An Effective Mechanism for Change
Good things do not come easy. Expect to spend more time personally sorting through the details and options, as well as working across the company to get agreement on what is important and how to make the project successful. Expect the project to take longer than a similar one with fewer defined-outcome KPIs. If your safety culture needs improvement, proactive video analytics can be an effective mechanism for change鈥攁nd protect people too. | WA

Video Analytics and Learning Systems

The term 鈥渧ideo analytics,鈥� as it applies here, means the real-time attribute analysis and pattern recognition performed on a video stream. For instance, is there any motion detected? Is that a person walking in front of the camera, or a deer? Is the person wearing a backpack? What color is the backpack? Video surveillance systems can also perform searches for footage with specific attributes. Video analytics systems also can learn through training. If you upload into the system photos that are known to show a certain make of truck, the system can now recognize that pattern and prompt an action.

Barton Kartoz is a Vertical Market Director for Minuteman Security & Life Safety. He has more than 25 years of experience developing and designing control, security, and safety systems for industrial markets. Prior to his time in industry, he was a commissioned officer in the U.S. Navy, where he served in various shipboard engineering roles. He may be reached at (609) 610鈥�0490 or [email protected].

References
MRF and collection worker deaths climbed in 2023: https://resource-
recycling.com/recycling/2025/01/07/mrf-and-collection-worker-deaths-climbed-in-2023
MRF fires increase substantially in 2024: https://resource-recycling.com/
plastics/2025/03/24/mrf-fires-increase-substantially-in-2024
NWRA and RRS Release Report on Threat of Lithium Batteries to Waste and Recycling Infrastructure 鈥� National Waste & Recycling Association: https://wasterecycling.org/press_releases/nwra-and-rrs-release-report-on-threat-of-lithium-batteries-to-waste-and-recycling-infrastructure
December Fire Report: This Year鈥檚 Trends: www.waste360.com/industry-
insights/december-fire-report-this-year-s-trends
NWRA and RRS Release Report on Threat of Lithium Batteries to Waste and Reyccling: Infrastructure 鈥� National Waste & Recycling Association: https://wasterecycling.org/press_releases/nwra-and-rrs-release-report-on-threat-of-lithium-batteries-to-waste-and-recycling-infrastructure