Work safety is an important topic for multi-segment companies. In the Utilities sector, the subject is even more sensitive, as activities are often carried out in high danger conditions, such as work on powered electric lines.
To ensure the protection of the workers carrying out activities that expose them to some kind of risk, the correct use of Personal Protective Equipment (PPE) and Collective Protective Equipment (CPE) is essential. Thus, companies invest both in the acquisition of protective equipment (PE) and in training employees for the correct use of equipment and the management of these assets.
Despite all these precautions, accidents can invariably occur, due to the lack or incorrect use of a PE — which can put the safety, health and well-being of employees at risk and, eventually, even affect third parties. Accidents are harmful to companies in many aspects, ranging from employees’ medical leaves due to accidents at work to damage to the image and reputation of companies.
The technology can and has been used in equipment and processes to ensure work safety, creating solutions that help companies and workers to automatically check whether the PE required for a given task is being used properly.
RFID PE identification
One of the solutions frequently adopted for the identification of each PE are RFID tags (Radio Frequency IDentification). RFID tags are stuck to PE pieces and their presence can be detected by readers, which read the unique identifier of the tags by radio frequency.
Readers can, for example, be installed in support teams’ vehicles, allowing workers to verify that all of the PE necessary for a given task is present. Thus, even before the car leaves for a job (or a set of jobs), it is possible to verify whether all of the necessary pieces of PE have been placed in the vehicle.
Each worker can also be equipped with a reader that will identify the owner of the PE, so that it is possible to monitor which piece of protective equipment has been used by whom throughout the operation and maintenance procedure. If there are any problems, such as a forgotten piece of PE, the operator can be warned while they are still in the field, through a portable terminal or smartphone, thus avoiding a potential accident at work.
In addition to its application in the field, RFID can also generate usage data that, after being collected by RFID readers, is stored in centralized systems. In these systems, it is possible to remotely monitor teams in real time and verify the compliance of each of them with the company’s safety standards.
Non-conformities can be reported immediately to the responsible managers. In addition, it is possible to generate PE data analysis reports. These reports, in turn, can help companies improve their safety policies and standards, as well as guide actions related to worker training and awareness regarding the correct use of protective equipment.
RFID is a well-known, low-cost technology that can be quite effective in the process of detecting the PE used by operation and maintenance teams. Another strong point of the technology is that it allows the exact identification of a piece of equipment, since a tag with a unique identifier is placed on each piece of equipment. Thus, workers are able to avoid using an incorrect but similar piece of PE (such as different types of gloves, which are intended for distinct uses) or even using a piece of PE that is out of date.
Despite the advantages mentioned above, PE detection solutions with RFID also have some problems and limitations. One of these limitation is the need for workers to carry an additional piece of equipment, the RFID reader, among the already numerous pieces of equipment that they needs to use during a job.
Another point to consider is that a job can have multiple steps, each requiring a different set of protective equipment. RFID tags will only allow the identification of which piece of PE is being used by operators, but it will not be possible to detect the step of the operation.
For instance, in an operation on a power grid, operators may need to put on a security harness only when they are working next to the wiring. Other contextual information — such as whether the grid is powered or not at a particular time in the operation — can also determine which pieces of PE are required at that particular time.
Computer vision for PE detection
Computer vision techniques, based on Artificial Intelligence (AI) — a field that studies ways to reproduce human intelligence patterns to solve problems and perform tasks such as sound and image processing —, can be used as a solution for non-intrusive PE detection, in addition to being a way to detect the context of the operation and maintenance procedure.
These computer vision techniques use cameras (which can, for example, be installed in the vehicle of the operation team) that monitor the entire job. The images of the cameras are processed by AI algorithms to detect the correct use of PE at each step of the operation.
In these solutions, AI algorithms are trained to identify objects in the images, be it pieces of PE (such as helmets, gloves, goggles etc.) or contextual elements (such as the workers themselves, poles, wiring, transformers etc.) that are relevant and help to give context to each step of the operation.
Thus, by analyzing the images captured by the cameras, AI will be able to identify the necessary PE in each of the steps according to the context — when the worker is on land or when they are performing an operation near wiring, for example.
That means that the system operates as if there is an inspector visually following the entire job and pointing out any deviations by the team. The solution may also include the use of thermal imaging cameras that can identify hot surfaces — equipment in operation, for example — that may bring additional risk to the team.
Despite its advantages over RFID technology, computer vision-based solutions also have some disadvantages. One is that, unlike RFID, it will not always be possible to identify a specific piece of PE through computer vision. Thus, the technology cannot, for example, be used to ensure that the piece of PE used is from the correct batch or not out of date.
An alternative to mitigate these limitations is to adopt a solution that combines multiple technologies, that is, that uses both RFID and computer vision, providing the best of both worlds. This solution would have both RFID accuracy and the resources to detect contextual information from computer vision, further increasing work safety.
This article presented two technologies that can be used for automatic verification of the use of personal and collective protective equipment. RFID technology is well known and can accurately identify each specific piece of equipment used by operators. Computer vision, in turn, allows the detection not only of protective equipment but also of its use at each stage of an operation. Solutions that combine the two technologies can bring even more safety to procedures, identifying protective equipment in a precise and contextualized way.
Solutions like these can help companies have greater control and insight into the real adherence to the use of protective equipment by field teams. In addition, they can assist organizations in defining their security processes and procedures and in planning team awareness and training actions. The solutions can help companies reduce the number of accidents at work that could be avoided by the simple use of protective equipment, cutting down costs, as well as ensuring the safety, health and well-being of employees.
Venturus is a private Science and Technology Institute that develops innovative projects, applying technology to create solutions that bring value to our customers and partners. To learn more about this or other topics please contact us or visit our website.