With the popularity of surveillance HD cameras and the maturity and improvement of intelligent video analysis technology, more and more industries are increasingly demanding intelligent video analysis. The application scenarios are more extensive and thinner than before. Turn. For some common functions of intelligent video analysis in monitoring, we carry out a simple analysis of some scenarios of intelligent analysis from the aspect of intelligent building.
    First, the category of intelligent analysis
VideoAnalytics involves image processing, tracking technology, pattern recognition, artificial intelligence, digital signal processing (DSP) and many other fields. The main intelligent analysis products are concentrated in the front-end and back-end categories. The front-end intelligence is to transplant some video analysis algorithms into the camera, and realize real-time video analysis and inspection in the camera. For example, the Jinsanli license plate recognition camera can identify the passing vehicles through motion detection and video triggering, and the vehicle capture rate is ≥99. %.
Back-end intelligence means that the front-end camera is only responsible for collecting video information and passing the code stream to the back-end server for analysis. Because intelligent analysis requires a lot of calculations, all the pressure from the back end is not conducive to real-time processing. Therefore, in recent years, the intelligent analysis mode of the video surveillance system has gradually moved forward from the smart image of the back-end smart forward camera. Currently, intelligent analysis mainly consists of the following categories:
    Diagnostic intelligence analysis
This kind of intelligent analysis is mainly for the accurate analysis, judgment and alarm of common camera faults, video signal interference and video quality degradation, such as snowflake, scrolling, blur, color cast, gain imbalance, pan/tilt control, picture freeze, etc.
   Identification class intelligence analysis
This kind of intelligent analysis mainly focuses on the analysis and processing of static scenes. Through the core technologies of image recognition, image comparison and pattern matching, it can extract and analyze the related feature information of people, cars and objects, such as license plate recognition and face recognition. , body color recognition, model identification, etc. License plate recognition is widely used in traffic flow detection, control, induction, and diversion; airport, port, community and other entrance and exit vehicle management; illegal, illegal red driving, illegal driving, traffic accidents and other illegal or illegal vehicle monitoring; highways, bridges, tunnels ETC, parking security, anti-theft management, and statistics on the travel time of various models.
   Behavioral intelligence analysis
This kind of intelligent analysis focuses on the analysis and processing of dynamic scenes, such as vehicle retrograde and related traffic violation detection, zone intrusion detection, wall over-the-counter detection, squall line crossing detection, item theft detection, item legacy, occupational operation detection, trajectory tracking, Population aggregation detection and passenger flow statistics.
Second, the application of intelligent analysis in the intelligent building industry
   Community security monitoring products
In order to better protect the property and the security of the community, according to the actual monitoring needs of the community users, cameras will be installed in key parts such as the main entrances and exits of the community, the community square, the elevator car, the surrounding, the gate, the house, the warehouse, the machine room and the parking lot. . The prevention products launched in different places have emerged in an endless stream. Because there are many monitoring points in the community, they are not described here. The following are mainly introduced for the frequent entry and exit of walls, entrances and exits, garages, etc.
Wall warning
Most of these products are mainly based on perimeter warning and intrusion detection. When there is a cross-line or accompanying line in the monitoring area, the camera will automatically report the environmental situation. For example, the perimeter intrusion alarm linkage video wall, the current intelligent analysis camera monitors someone to climb over the wall, the camera automatically uploads the alarm information to the video management platform, and at the same time, the monitoring center TV wall will be linked according to this action to display real-time video. The administrator can take appropriate measures based on the content of the image to determine the real-time situation.
    Entrance and exit non-motor vehicle access control
When someone passes the gate/access control, the face recognition server stores the face data recognized by the camera and uploads it to the platform for face comparison. When the platform confirms that it is the same person, the server linkage gate/access control is opened, such as Jinsanli entrance and exit non-motor vehicle lane face recognition automatic opening system.
   Garage entrance and exit monitoring
Using the license plate recognition technology, when a car is about to pass the gate, the license plate recognition camera recognizes the key information such as the license plate and the driver, and compares it with the original vehicle data of the management platform, so that the whitelist vehicle does not stop automatically opening the gate. The blacklist vehicle gate does not open the parallel alarm bell. According to the JT/T604-2011 standard test, the recognition rate of the Jinsanli license plate recognition camera reached 98.4%.
    Garage special condition alarm
Through the audio recognition technology, when someone in the local garage calls out for help, the pickup detects the audio and uploads it to the intelligent analysis camera. The camera analyzes the sound abnormally, and uploads the abnormal sound of the call to the integrated management platform. The platform performs the linkage TV. The wall and the preset position will zoom in and out of the live video of the rescuer on the TV wall, and the security personnel can quickly take action to gain time for the rescued person to be rescued.
    Property inspection
Â
It is well known that there is a certain distance between the real estate marketing center and the building. If the owner wants to visit the building, it is extremely inconvenient and the building under construction often has potential safety hazards. For such problems, landscape capture cameras can be installed on the floors with excellent views. Through remote transmission, information can be transmitted to the video wall of the marketing center for playback, which enables customers to view the floor landscape in the marketing center.
   Commercial center security monitoring products
The business center is generally located in the center of the city, with large traffic and complex environment, which poses a problem for safe and effective monitoring.
    Business center flow detection
The business center flow detection can be divided into pedestrian street congestion detection and marketing center/hotel/shop traffic statistics according to indoor and outdoor environment. In the pedestrian street congestion monitoring, the front-end intelligent camera analyzes the traffic of the pedestrian street. When the traffic exceeds the environment and can accommodate the upper limit of the flow, the management platform automatically triggers the alarm system to issue a congestion alarm, prompting the security personnel to take the current limiting measure; the indoor passenger flow detection is mainly The flow of people entering and leaving every day is counted, and the flow information is stored on the management platform to generate various reports such as daily reports, monthly reports, annual reports, etc., and the on-site reception personnel are not required to register one by one.
    Mall/hotel lobby face recognition
Through the face recognition function, the mall can distinguish between visiting people, such as distinguishing the types of consumer groups, distinguishing between new and old customers, and identifying theft personnel. When the system identifies large customers according to the set VIP customer characteristics, it can automatically display the name and title of the big customer to the hotel service staff, so that the staff can more conveniently receive the reception, and at the same time can create a feeling of homecoming for the customer; when the camera is based on the system The entered blacklist identifies the theft and can trigger the alarm function on its own.
    High altitude parabolic monitoring
This monitoring is involved in commercial centers and residential quarters with high-rise buildings. Under normal circumstances, we think that the building has four facades, and the residential quarters can be divided into facades, inner façades and side façades according to the construction situation. High-altitude parabolic monitoring combines the height of the building floor and the visual field of view to determine the camera. Point and quantity. It is usually installed at an angle that combines a top view and a bottom view, and each camera covers 10 to 15 layers. When an object is thrown at a high altitude, the camera video information is transmitted back to the management platform to link the alarm system to alarm.
    Special function product
Due to the differences in the monitoring environment, technicians will design intelligent analysis products with different functions according to the needs of the site.
    Custom monitoring mode
The customized monitoring mode is divided into a person monitoring mode, a core area monitoring mode, and a patrol monitoring mode. Due to the large number of monitoring points in the local industry, the invalid images account for nearly 95% of the total. If one screen is viewed, it will cause a lot of manpower and material waste. The configurable rotation mode is realized by the customized monitoring mode, and the security personnel's energy is used in the cutting edge to effectively improve the security efficiency of the real estate.
  Variable stream storage
The intelligent analysis camera can automatically identify whether there is someone on the front-end monitoring screen. When the monitoring area is in an unattended state, the front-end camera automatically performs frame-down storage, which can greatly save storage space. Taking Jin Sanli intelligent analysis camera as an example, the 1080p picture can be reduced from the original 4M storage bandwidth to 1.2M, and the storage device can save up to 70%.
   High code stream in sensitive area, low code stream in non-sensitive area
The area from 25% to 30% above the general video is the top image of the sky or building, which belongs to the lower area where the sensitive elements appear. This can be set to the low stream on the video surveillance system, and in the picture. Key areas such as the following people can be set to high code stream. Save storage capacity while saving real-time bandwidth.
Third, the conclusion
Although intelligent analysis products have entered the industrialization stage in the security industry, there are still many difficulties in intelligent analysis technology. The problems of false positives and false negatives are still the most concerned issues for customers. How to solve these problems is still important for the entire security industry technicians. Question. In the entire smart security field, as the market continues to be segmented and users' requirements for intelligent analytics products continue to increase, the market as a whole presents opportunities and challenges. With the innovation of intelligent analysis technology, we believe that intelligent analysis will develop in a more professional and more precise direction.
Sliding Gate Metal Wheels,SlidingGate Wheel With Zerk For GarageDoor.SlidingGate Wheel With Zerk For FactoryGate,SlidingGate Wheel With Zerk For HeavyDoor
Xiamen XMillion Co.,Ltd. , https://www.xmilliondoor.com