1. Industry background
Petrochemical production and storage enterprises belong to high-temperature, high-pressure, flammable, explosive, and toxic hazardous industries. Their production equipment is large-scale, dense, complex production processes, and closely coupled production processes. With the development of Internet technology, using advanced AI, Internet of Things, big data, cloud computing and other technical means to continuously improve the safety supervision level of the petrochemical industry has become the development trend of the industry. Based on AI intelligent detection technology, it can monitor and automatically warn other dangerous/violative behaviors such as off-duty/leaving duty, smoking, wearing protective equipment, open flames, liquid/gas leakage, personnel intrusion, etc. in the petrochemical supervision scene. Using intelligent video technology to help enterprises to produce safely, discover safety hazards in time, improve management level, and use artificial intelligence and human coordination to improve supervision efficiency and liberate manpower.
2. Program overview
TSINGSEE Qingxi video intelligent analysis system takes AI video intelligent identification and analysis capabilities as the core. Through intelligent identification and analysis of surveillance video, it can provide identification, capture, comparison, Alarm and other services, conduct accurate research and judgment on abnormalities and violations in petrochemical supervision scenarios, and assist in decision-making, etc., to meet the needs of enterprises based on video services for data perception, intelligent detection, intelligent analysis, and intelligent alarms.
In the scene supervision of the petrochemical industry, the TSINGSEE Qingxi video AI intelligent algorithm platform can carry out intelligent detection and analysis of AI scenes in key monitoring areas. In scenarios such as irregular positions, people in explosion hazard areas answering and calling, people smoking in non-smoking areas, and hoisting operation violations, etc., implement intelligent analysis and early warning, reduce on-site risks, and implement hierarchical management and control measures, and finally achieve the purpose of preventing problems before they happen through AI means .
3. AI video intelligent analysis technology
1. Universal intelligent detection
1) Wearing detection of protective equipment: Detect whether the protective equipment worn by workers in the hazardous chemical factory is compliant, including: helmet detection, work clothes detection, seat belt detection, etc. If abnormal wearing is detected, an alarm will be triggered immediately and record.
2) Vehicle detection: Supports the capture, detection and identification of dangerous chemical vehicles in the video screen, including: vehicle entry, vehicle departure, parking position detection, etc.
3) Personnel behavior detection: identify personnel behavior, including: detection of personnel falling to the ground, smoking, making phone calls, playing mobile phones, on-duty/off-duty detection, etc.
4) Fire safety inspection:
-
Smoke and fire identification: intelligently supervise the fire safety scenes of hazardous chemical enterprises, and detect whether there are smoke and fire hazards in the monitoring screen of fixed areas. When smoke is recognized, the system will automatically send out an alarm and record it.
-
Fire-fighting equipment detection: including fire-fighting equipment placement, fire-fighting equipment removal/missing detection, etc.
5) Intrusion detection in the warning area: detect whether there is any intrusion in the designated area of the video.
2. Professional intelligent detection
1) Safety inspection of the loading and unloading process: conduct safety supervision over the entire process of the filling operation of hazardous chemical vehicles, including before, during, and after loading and unloading, and conduct real-time monitoring and early warning of violations and behaviors during the operation process, including vehicle Front license plate detection, valve inspection action detection before loading and unloading, anti-slip car skid placement detection, etc.
2) Leak detection: real-time detection of liquid leakage (visible light), abnormal temperature liquid leakage (infrared thermal imaging).
4. Video Fusion Technology
Security monitoring/video aggregation/video monitoring EasyCVR video fusion platform can quickly and conveniently connect cameras deployed in various areas of the chemical plant area, realize unified aggregation and centralized management of scattered video monitoring resources, and use device tree for grouping and hierarchical management , to achieve the purpose of unified, centralized and efficient supervision of the monitoring center.
Video aggregation/video cloud storage platform EasyCVR can support multi-protocols (national standard GB/T28181, RTMP, RTSP/Onvif protocol, Hikvision Ehome, Hikvision SDK, Dahua SDK, Uniview SDK, Huawei SDK, EZVIZ SDK, Lecheng SDK), multi-device access, realize lightweight access to massive resources, unified device management, configuration control and analysis applications. Using the surveillance video deployed in each area, the entire process of filling, loading and unloading, berthing, storage, and parking of petrochemical plants can be supervised. The video platform supports single-screen and multi-screen display, and any one or multiple channels of video can be selected for viewing. The number of video windows is 1, 4, 9, and 16 optional.
After the data collected by the equipment is detected and analyzed by AI intelligence, it is unified into the EasyCVR video fusion platform, and the visualization results of data analysis and statistics are displayed, and a petrochemical enterprise safety production risk warning platform based on AI video intelligent recognition technology is constructed.
TSINGSEE Qingxi video intelligent analysis AI algorithm system is based on AI intelligent recognition technology, which can automatically and real-time monitor the safety production process of hazardous chemical enterprises (general security scene, dock management, warehouse management, SOP whole process, loading and unloading process, hot work, People, vehicles, objects, behaviors, etc., can accurately detect and identify abnormal events and give timely warnings, and provide comprehensive intelligent perception, active warnings, and decision-making assistance for the safety production process.