Construction safety is a “pain” that has plagued traditional construction sites for a long time. Looking at the current safety management status of traditional construction sites, the loopholes visible to the naked eye are causing the whole society to think about safety construction—the construction site operators are complicated, and the cost of manpower management and control is high; people, objects, and vehicles are dense, and the site environment presents an unpredictable situation. Complexity; it is difficult to control the progress of the project, the rectification and implementation measures are not implemented in place; the building materials are easily stolen or damaged, etc.
On the other hand, in the field of security monitoring, computer vision technology enables surveillance cameras to have the ability to “see and understand” like humans. Based on the depth algorithm, the camera can replace the human eye to intelligently track and recognize the target, and through artificial neural network, key feature matching and other algorithms and intelligent statistical models, it can realize target recognition, target tracking, and motion trajectory analysis in the monitoring area. And feature classification and so on, this is artificial intelligence.
The artificial intelligence that integrates the Internet of Things, big data, Internet and other technologies is applied to the investigation and management of safety hazards on the construction site. By covering the construction site in an all-round way, it can effectively reduce the blind spot of vision, and realize 24-hour monitoring of the construction site and visual management, which is convenient for Real-time control of the construction progress of the construction site, enhance the ability to detect hidden safety hazards on the construction site, reduce the probability of accidents, and ensure the safety of life and property of construction workers. How exactly does it work?
Hard hat inspection/reflector inspection
Through real-time video monitoring and early warning whether on-the-job workers have taken safety precautions as required, such as whether they are wearing safety helmets correctly, whether they are wearing reflective clothing, protective clothing and other work clothes. If an abnormality is detected, a voice alarm will be issued.
Real-time detection of specific areas, such as detection and early warning of personnel entering dangerous areas such as power distribution boxes, timely capture of intruders and alarm, and also linked to on-site voice for prompts to facilitate timely stop. At the same time, it can be linked to the access control system. For example, construction turnstiles are installed at the entrance of the construction site. Workers must enter through face recognition (or fingerprint recognition, QR code scanning, etc.), effectively preventing outsiders from entering the construction site and protecting the safety of personnel and equipment.
It can accurately identify the smoke and flames in the monitoring area in real time. Once the fireworks are detected, an alarm will be issued immediately, and the alarm information will be pushed to the relevant management personnel for processing.
Fire escape occupancy/blockage identification
Monitor corridor space, safety exits, fire truck passages and other areas. Once occupancy, accumulation of debris, blockage, etc. are detected, an alarm will be triggered immediately to remind workers to deal with them in time, reducing fire hazards in internal production and management.
Wear a mask to identify
Detect whether workers wear masks in the monitoring area. Once a person who does not wear a mask or does not wear a mask correctly is detected, the voice player can be linked to give a voice reminder.
Sleep recognition during working hours
Identify the behavior of people sleeping and leaving their posts in the monitoring area to prevent some key positions from causing immeasurable losses due to negligence of duties. When an abnormal situation is detected, the system will automatically alarm, and capture and record the images when they leave or sleep.
Risky behavior detection
Detect whether there are dangerous behaviors in the operation area, such as climbing, falling, staying, suspicious wandering, etc., and warn in advance to prevent problems before they occur
With the rapid development of artificial intelligence in recent years, deep learning methods and performance have been improved, and technologies such as computer vision, image processing, video structuring and big data analysis have also been continuously improved, making security products gradually become intelligent. In terms of technological maturity, the technology for processing security images has been relatively complete, and the industry’s guiding policies have further accelerated the application of artificial intelligence technology.