Product test report: Zeno Videopark’s ZN-IV1304N-P integrated video intelligent analyzer

Looking around in an “intelligent’ world, all walks of life are more or less using smart applications. Over the years, Zeno Videopark has kept up with this trend and has been committed to intelligent applications by continuously introducing intelligent products, such as, power plant intelligent analyzer, face recognition analyzer, full-target analyzer, ATM intelligent analyzer and behavior analyzer to different market segments. In this article, we provide an in-depth analysis on Zeno Videopark’s power plant intelligent analyzer.

Due to the new possibilities brought by IoT and AI technology, the power industry has ushered in an intelligent transformation. Power operation and maintenance have become more and more intelligent and refined, power plant safety also has attracted the attention of a society with growing safety concerns. Therefore intelligent power plant safety management has become a vital need that the power industry cannot ignore.

Based on this market trend, Zeno Videopark launched the ZN-IV1304N-P integrated video intelligent analyzer (hereinafter referred to as “the analyzer”), which uses self-developed deep learning algorithms to conduct AI analysis on on-site personnel and supervisors, ensuring that the behavior of personnel conforms to relevant power distribution room regulations, and that the working environment and safety of on-site personnel are up to standard. What is the intelligent analysis capability of this analyzer? asmag.com has conducted a product test and provides the results in this article.

Small size, embedded integrated design

The analyzer adopts an embedded integrated design, and features a black metal shell, that is sturdy, durable and small in size at 215mm (length) × 160mm (width) × 45mm (height), making it convenient for the analyzer to be embedded in an integrated system for industrial applications.

The system uses 12VDC 2A power supply, and power consumption is 10W (without having to use a hard disk). What’s more, the analyzer uses a triple watchdog circuit design making sure that it never crashes. With a working temperature range of -10-55°C and a working humidity range of 10-90 percent, including in harsh environments such as electrical rooms, where it can work effectively and smoothly.

Multiple intelligent functions

The analyzer uses Zeno Videopark’s self developed advanced human bone analysis, behavior analysis and sign/identification analysis technology. By working with a monocular camera to achieve AI video intelligent early warning, it can realize compliance detection including (helmet, uniform, face),tripwire, intrusion detection, people falling down detection, people counting,abnormal retention detection, object abandoned detection, flame detection, etc. 

Personnel compliance

Personnel compliance detection including helmet detection, uniform detection, face detection
With AI-based face recognition, the analyzer can confirm whether the visitor is registered within the system, face sign-in and face door opening functions can be realized.
 
An incorporated AI algorithm will analyze whether personnel and employees are wearing a safety helmet, and the correct clothing. If the worker is detected to be non-compliant, an alarm will be triggered. During testing, the alarm was triggered in a fast and responsive manner whenever it detected that a person is not wearing the correct safety equipment and clothing. 
 

Intrusion detection

Draw a specific area in the video; an alarm will be triggered when someone intrudes into the area. During testing, when an illegal intrusion is detected, the system triggers an alarm in real time in a quick and responsive manner. 
 

Falling down detection

The AI algorithm will trigger an alarm when someone falls to the ground. During the test, when the inspector falls to the ground, the system triggers an alarm in real time in a quick and responsive manner.
 

People counting

Select an area in the video and preset the number of people allowed in the area. If the number of people in the area does not match the preset number, an alarm will be triggered. In the test, the abnormal personnel threshold is set to 4 people. When there are fewer or more than 4 people, the system triggers an alarm in real time in a quick and responsive manner. 
 

Object abandoned detection

Select an area in the video by drawing, and the time a suspicious or leftover object remains in the picture is counted. When a foreign object appears in the video and remains there for a certain time, an alarm will be triggered immediately. In the test, the time an object should remain in the picture is set to 5 minutes, and when the item stays there for over five minutes, an alarm is triggered immediately in a quick and responsive manner.
   

Flame detection

The system uses AI algorithms to detect flames appearing in the video. An alarm is triggered when the flames are detected, and the relevant monitoring personnel will be informed immediately. 
 
In addition to the above-mentioned AI intelligent analysis functions; the analyzer can access to multiple sensors and allows sensor data superimposition, so that important sensor data can be displayed intuitively and clearly on the video. When a relevant event triggers an alarm, the on-site monitoring personnel will be informed immediately.

Good integration ability

The analyzer supports different interfaces and has: 2 100/1000 Base-TX RJ-45 ports, 2 USB2.0 ports, 5 switching inputs, 4 relay outputs NC/NO, RS-232 and 485 ports, as well as audio input/output to meet the on-site connection of various external equipment. 

The analyzer is capable of seamlessly integrating with the video surveillance systems. It is compatible with mains camera brands such as, Hikvision, Dahau, and Uniview or any third party camera brand via ONVIF, GB28181 and RTSP connections. SDKs are also provided so different platforms can have the analyzer integrated into them.

Simple and convenient web management

The analyzer meets the latest market demand, using the HiSilicon Hi3559A chip and the embedded Linux OS. The system simultaneously uses the advanced G.711 audio compression processing technology, which can recognize real-time voice intercom communication between the center and the equipment at the front end. It also supports image capturing from various sources with images stored locally or uploaded to the monitoring center.

The analyzer also comes with a Web system, and configuration and management functions such as real-time preview, function configuration, video playback and real-time alarm reception can be done without software or apps. It supports configuration of hard disk to storage alarm video . Upon an event, only the alarm information and alarm photos need to be uploaded in real time. The relevant video can be read and played back accordingly, reducing the need to set up high-bandwidth connectivity

Contributing to intelligent power industry

At present, the analyzer has been widely implemented in the power industry. In one instance, it integrates with front-end data collection and central control management to form a solution for IoT-based power distribution rooms.

In this solution, the environmental monitoring devices are responsible for collecting environmental data in the power distribution room; the analyzer is responsible for dangerous area intrusion detection, off-limits area entry detection, fall detection, face detection and capacity and flame detection; and central control management is responsible for data transmission (environmental and video data) and alarm processing. The three complement each other to provide an intelligent, comprehensive monitoring solution that integrates with various surveillance subsystems and gathers relevant power distribution room data, in the process smartening up the power plant. 

The electric power industry is only one of the applications of Zeno Videopark's integrated video intelligent analyzer. Through seamless integration with the video surveillance system, the analyzer allows video image analysis and smart alarms and can be widely extended to various sectors such as construction sites, healthcare, education, banks, and airports to ensure the safety of personnel in large places.

Product features:

1. The use of HiSilicon Hi3559A chip and deep learning algorithms based on skeletal target detection has greatly improved the accuracy of intelligent video analysis.
2. Supports compliance detection including (helmet, uniform, face),tripwire, intrusion detection, people falling down detection, people counting,abnormal retention detection, object abandoned detection, flame detection, etc.
3. Can be seamlessly integrated with main camera brands through ONVIF, GB28181-, and RTSP.
4. Supports the configuration of hard disk to store alarm video. During an alarm, only the alarm information and alarm pictures in real time can be uploaded. The relevant video can be read and played back accordingly, reducing the need to set up high-bandwidth connectivity.
5. Highly reliable hardware design, using a triple watchdog circuit design to ensure that the system never crashes.
6. The built-in Web system allows real-time preview, function configuration, video playback and real-time alarm reception functions all in once.

Expert review:

The analyzer employs skeletal analysis, behavior analysis and sign/identification analysis developed by Zeno Videopark to support compliance detection including (helmet, uniform, face),tripwire, intrusion detection, people falling down detection, people counting,abnormal retention detection, object abandoned detection, flame detection, making it ideal for intelligent early warning and abnormal behavior detection in power distribution rooms and switch rooms.

In addition to the power industry, through seamless integration with the video surveillance system, the analyzer enables multiple functions such as video image analysis and intelligent alarms and can be widely extended to many sectors such as construction sites, healthcare, education, banks, and airports to ensure the safety of personnel in large places.