How Can Voxel Help You, as a Safety Manager, Reduce Incident Frequency

If you’re a safety manager at a commercial or industrial facility such as a warehouse, distribution center, manufacturing plant or construction site, you may be confronted on a daily basis with safety protocol non-compliance, unsafe working environments and/or accidents that result in incident frequencies that are higher than what’s acceptable for you or your company.

Let’s face it — even a single incident is one too many, and with each one, workers’ compensation, liability, insurance and property claims increase, affecting the bottom line of your business and impacting your organizational standing and job security. By contrast, if you can reduce incident frequency, you can improve your business’s bottom line and gain favor with company higher-ups.

The Current State

What are the causes of safety and workplace incidents? Some of them include worker cellphone use, smoking on the job, improper lifting postures, failure to wear personal protective equipment (PPE), not following safety protocols, blocking aisles and egresses, aggressive vehicle driving and/or speeding, accidents and theft.

Many companies would like to reduce the number of safety incidents they experience, but employ primitive methods to do so. Some of these methods may be behavior-based, but questions are often asked about how they’re monitored.

Some companies rely on video security cameras to observe workers and workplaces, but again, the question of monitoring arises; in these cases, observation may be dedicated and scheduled, or it may be ad hoc or spontaneous. Cameras may record imagery continuously, sporadically or not at all.

If the number of video monitors for observation is less than the number of cameras in use, camera outputs will typically alternate between on and off states for several seconds or minutes at a time. This means that even in areas with active camera coverage, scenes or locations may not be viewable at all times.

Even in cases where the ratio of cameras to monitors is 1:1, only one set of eyes for observation generally means that not every output will be viewable simultaneously. Even if a camera image is being directly stared at, human fallibility ensures there will be many occasions when people, objects, vehicles and other details won’t be noticed, and/or potential hazards won’t be detected until it’s too late.

The above methods are manual, cumbersome, and reactive — rather than proactive — ways of dealing with safety incidents. Often, there’s no attempt made to measure cumulative events or perform subsequent statistical analysis.

Computer Vision

The solution to the above problems is to eliminate the fallibility of manual, human observation by using computer vision. Computer vision systems automatically and continuously scan video images in real-time to track discrete objects (such as doors, machines and tools), vehicles (such as forklifts and trucks) and people. These systems can also detect the presence or absence of secondary objects (such as boxes, pallets or manufactured products), take note of human movements (including postures thereof or the lack thereof), take note of whether safety gear is being worn and/or monitor safety-related conditions.

Using an artificial intelligence (AI)-based rules engine, risk hazards can be identified and analyzed, and safety violations, operational anomalies and security breaches can be recorded and reported. Consequential actions such as alerts and notifications can be triggered automatically. High-risk cases or events can be reported immediately to one or more people. Real-time analysis and insights can be viewed via a web-based user interface, and reports can be sent to email and mobile devices.

Enter Voxel

One computer vision solution as described above is made by San Francisco-based imaging startup Voxel.

Voxel’s solution employs AI technology to identify and analyze risks from video camera inputs. Voxel’s web-based solutions make use of any existing camera installations, including those that employ Internet Protocol (IP) cameras and real-time streaming protocol (RTSP) feeds. The signals from these are fed to Voxel via an application programming interface (API).

Voxel then passes the signals to a cloud server for real-time processing. Patterns are detected, and any anomalies in these patterns are scrutinized. Risks are scored, any events are recorded, and data gets graphed, so incident counts and trends can be observed. For customers, Voxel supplies out-of-the-box use-case solutions, and use cases can also be custom-defined.

For safety managers, Voxel computer vision solutions allow safety monitoring to be automated and intelligent. Managers no longer have to rely on manual, improvised or ad hoc solutions for safety monitoring and enforcement. Incident frequency can be reduced, and employees can be coached using captured footage as teaching examples to foster and cultivate an organizational culture of safety. Fewer incidents mean fewer workers’ compensation, liability, insurance and property claims, which means that safety managers are looked upon favorably.

If all of the above sounds good for your organization, schedule a demo to get a demonstration of Voxel’s computer vision solutions today