The utilization of AI computer vision technology to prevent safe accidents
Hello, It’s Mobilio, a company that pursues zero safety accidents in industrial sites with AI-based CCTV solutions. In the last post, we briefly introduced the composition and key features of Mobilio AI Based CCTV. AI CCTV is designed to capture key events occurring in the workplace in real time, determine event content and risk using artificial intelligence (AI) technology, and take the right actions for the situations. Looking deeper, AI CCTV uses a technology called computer vision to detect and analyze events. Of course, this technology is also based on AI technology.
Today, we are going to summarize the widely used computer vision technology for safety accident and event detection at industrial sites. This technology is widely used in industrial sites as well as in public safety and security. Although it is still growing rapidly, AI computer vision technology is evolving faster as it is combined with the Internet of Things (IoT).
The Internet of Things (IoT) can be summarized as a huge network of devices that can send and receive information over the Internet. In particular, Industrial IoT (Internet of Things for Industries, IIoT) supports the diverse use of IoT devices in industries and manufacturing. For example, heating and cooling system (HVAC) manufacturers can install IoT sensors in their products, monitor their performance remotely, and identify potential failures. This scenario can be applied equally to a company’s business processes and workplaces. Through this real-time data acquisition and situation analysis, companies can make and execute accurate decisions faster. This system is a key infrastructure for preventing and reducing safety disasters.
AI Computer Vision to Industrial Internet of Things
AI computer vision can be used for various event detection and analysis operations in industrial sites. Companies such as manufacturing, logistics, construction and mining have so far relied heavily on human supervision or visual inspection to assess product quality or work progress and prevent safety accidents. In order to maintain the quality standards set by the company and protect the brand image of the company, thorough and repeated visual inspection by production stage is essential.
But there are a few problems lurking here. A person cannot perform a visual inspection for 24 hours continuously. Also, there are frequent situations where some defects cannot be detected due to the supervisor’s physical or psychological problems. And this method also has a time-consuming and costly problem. Under these conditions, AI computer vision technology accurately identifies abnormal conditions or quality issues 24 hours a day, contributing significantly to reducing industrial accidents. This requires learning for AI computer vision systems to distinguish between “normal” and “abnormal” and to detect product damage or errors. This is called machine learning. In addition, linking information collected from IoT sensors applied to tasks and products increases accuracy.
There are many ways to visually supervise and observe using IoT and AI computer vision technology in industrial sites. For example, General Electric, which produces large motors and industrial engines, operates a maintenance process that can identify and deal with the cause of the failure without disassembling the failed engine. This scenario is possible because it uses sensors built inside the engine and ultra-small detection robots that operate according to the situation of failure. In this scenario, IoT (Industrial Internet of Things) and AI computer vision technology on robots play an important role.
Scenarios for the use of AI computer vision technology
The above mentioned AI computer vision technology can be utilized in a wide variety of areas around the manufacturing industry. Let’s take a quick look at the available scenarios:
1) Detect safety accidents
Detect and respond to events such as falling objects in the workplace, falling of workers, and facility crashes that are being occurred frequently. If IIoT sensors are installed in the workplace and on workers safety gears, more accurate analysis of the situations (incident severity, location, etc.) are possible.
2) Product tracking
An increasing number of companies are attaching RFID, bar codes, QR codes or IoT sensors for delivery, history management, and maintenance of produced products. AI computer vision can recognize these IDs attached to products and perform tasks according to predetermined procedures and methods. This allows you to pinpoint the location, specifications, etc. of the products shipped or sold, and to perform the necessary tasks quickly. This scenario requires the development of machine learning models and learning processes using related data in advance.
3) Automatic equipment and robot work observation
In May, there were two unfortunate accidents in which workers died due to operating errors of production robots. Korea ranks first in the world in terms of robot ratio per 10,000 workplaces. Therefore, this kind of safety accident is likely to continue to occur in the future. Given this situation, companies with high automation rates should also consider using AI computer vision technology. AI computer vision technology is a good way to strengthen worker safety nets and manage whether automation facilities such as robots operate normally.
4) Worker protection
Workers at high risk working sites must wear protective equipment such as safety helmets, goggles, safety jackets and protective gloves. However, these principles are often not followed for various reasons and circumstances. This is due to rising temperatures in summer and reduced work efficiency due to wearing protective gear. AI computer vision technology can detect and take action against workers who do not comply with safety rules.
5) Fire Detection
AI computer vision can detect fire hazards and occurrences through analysis of the appearance, color, and flickering of objects captured in the video. Depending on the type and severity of the event identified, the control center and field supervisor will be informed the situation immediately and take the actions to prevent the accidents.
6) Face recognition
AI computer vision technology can recognize the face of an operator. And based on the results of the confirmation compared to the workers’ image database, it can perform tasks such as requiring only authorized personnel to enter the workplace. Facial recognition technology can be used to prevent accidents through unauthorized personnel access control, as well as to protect and secure major facilities.
As you’ve seen before, more and more companies are applying AI computer vision technology to business processes and worker protection because of the variety of applications and effects of AI computer vision technology. It is the same reason why Mobilio introduces the AI-based CCTV solution.
Mobilio Co., Ltd. specializes in safety accident prevention solutions.