

Product Advantages
With the world's leading AI video monitoring and recognition technology, the recognition accuracy rate is as high as 99%; it can be applied to multiple business scenarios and push lightweight detection and alarm information to customers in real time.
The Division AI video intelligent analysis system can directly connect with the NVR or the local area network where the camera is located for detection, and perform aggregation processing and intelligent analysis on the video at the terminal to generate a video stream of the detection result, reducing the network bandwidth cost of the user.
Division AI video intelligent analysis supports privatization deployment. It can connect inspection equipment to cloud management, check status, and assign inspection tasks. At the same time, it presets rich AI video monitoring and recognition skills. It has a lightweight AI model that can be quickly extracted. valuable data.
Construction site safety helmet inspection
Attendance at gate entrance and exit
Gas tank filling monitoring
Dust and smoke detection
Bright Kitchen Stove
Video Violation Review
Construction
In the construction site environment, the price comparison is complicated, the distance is relatively long, the characteristics of safety risk are relatively obvious, and the detection rate is relatively high. However, the texture features of the head are not very obvious, resulting in a low head detection rate. The target detection rate is over 80% under the zoom of 3-5 times, and difficult targets can be found within 5 seconds.
High-altitude tower crane safety helmet detection performance
Safety helmet detection accuracy: 0.909548
Head detection accuracy: 0.941176
Gate Exit
Gate entry and exit attendance: Count the number of employees entering and exiting the gates in construction sites, parks and other areas to identify whether they meet the safety standards for wearing helmets, masks, etc. to enter the park, so as to achieve intelligent attendance
Label Description
-I : entry gate
-O : Exit gate
-N : no helmet
-W : Carry a hard hat, but not wear it
Red frame: no helmet
Green frame: wear a helmet
AI picture detection server 1. Provide detection parameter configuration interface: configure filling station number, name, position information of each scale on the picture 2. Provide picture upload interface: parameters include picture, filling station number, time; return to test each scale With or without liquefied gas cylinder status
Plan Description
1. The platform scale reports the filling status information to the management platform every 15 minutes (to be determined). Delays or failures are regarded as violations.
2. The management platform obtains a screenshot of the video of the filling station every minute (the time can be configured), and sends it to the Al image detection server to analyze whether there are liquefied gas cylinders on the scale.
In construction sites, roads, factory workshops, catering kitchens, outdoor forest areas and other scenarios, monitor whether there is dust, smoke, etc., and timely alarm reminders to reduce pollution to the atmospheric environment and personal and property damage caused by fires.
Label Description
-S : smoke and dust
Based on artificial intelligence visual detection and analysis technology, it can intelligently monitor the sanitation status and operation specifications of the back kitchen, effectively guarantee the status of kitchen safety guards, reduce labor supervision costs, and improve the supervision efficiency of law enforcement departments.
Detection Parameters
Human body testing, no masks, no hats, no gloves, smoking, uncovered trash cans for a long time, dirty and messy
For videos broadcast on offline large screens such as squares, shopping malls, hospitals, and governments, intelligently identify whether they contain pornography, violence, terrorism, politics and other illegal content, and block the broadcast content in advance to avoid public broadcast accidents
Label Description
Terrorist Tags
-NV : normal
-S : special symbol
-W : war scene
-B : bloody scene
Yellow Label
-NY : normal
-O : porn
Related Products
Cooperation Cases
400-006-5266
400-006-5266
Headquarters Hotline (Monday to Friday 9:00 - 18:00)
Hello, The Division pre-sales team is always online.