Situation Analysis

There are a large number of photovoltaic, wind power, hydropower and thermal power stations in our country. Among them, small hydropower stations are multi-faceted and have problems such as being small, scattered, weak and chaotic, and having a weak safety foundation. Currently, many power stations have adopted the "industrial video + manual watch" management method to assist manual workers in safe production management. The system has basic intelligent behavior recognition functions, but the recognition rate is low, the false alarm rate is high, and the application experience is poor.

The safety of power stations is related to the national economy and people's livelihood. In the security and anti-terrorism construction of power stations, traditional methods such as video surveillance and electronic fences are usually used, which can often only be traced back afterwards and cannot link multiple systems to provide early warning and active eviction.

In some specific scenarios of power stations, such as dams, slopes, photovoltaics, wind power, etc., traditional monitoring methods are used, mainly relying on manual inspections and inspections, which are inefficient and difficult to manage. For example, photovoltaic power stations are often built on hillsides. The area is vast, personnel operation efficiency is low, and safety risks are high.

solution

The AI-based safety production solution is divided into three parts, namely "next generation industrial video subsystem", "security anti-terrorism subsystem" and "dam safety comprehensive management subsystem".

The solution is based on the company's "Algorithm Factory" and "Algorithm Supermarket" basic platforms, which can provide a unified AI common service standard interface, can be linked to multiple systems, supports algorithms to quickly and accurately match power scenario applications, and has open API interfaces for easy integration and algorithm modules. distributed to facilitate flexible deployment.

The program composition is as shown below:

The next generation industrial video subsystem, as shown below:

System advantages

1、Meet national, industry and group regulatory requirements

2、Multi-system linkage can provide early warning, broadcast eviction and automatic drone inspection based on AI recognition results.

3、Compatible with mainstream surveillance brands, recyclable and protect investment

4、Rich types of algorithms and strong adaptability to scenarios

5、The algorithm model is lightweight and can be deployed at the edge

Security and anti-terrorism subsystem, as shown below:

System advantages

Achieve all-round monitoring of airspace and factory perimeter, can compare faces with the three public security stations, proactively identify dangerous personnel at long distances, intelligently link the system, proactively warn, and realize autonomous drone detection and eviction, eliminating potential hazards through intelligent means Security Risk.

Dam safety subsystem, as shown below:

 System advantages

The system meets the requirements of the group's dam center. Based on Beidou 3's millimeter-level dam safety monitoring, it can grasp the small displacement changes of dams, slopes, and structures, detect potential risks in a timely manner, and realize multi-system linkage based on artificial intelligence visual analysis, such as opening When the water is discharged from the gate and someone is found downstream, the horn is automatically driven away. The upstream intelligently identifies foreign object incidents and drones take active aerial inspections.