Video surveillance is an important technology and application field in video engineering. In less than a century of development, video technology supporting its development has undergone roughly three leapfrogging stages of development.
In the 1930s, video technology represented by television broadcasting came out of the laboratory, entered radio and television stations and millions of households, and achieved a leap from still image transmission to moving image transmission; at the end of the 1980s, conference television and video were used. The video technology represented by monitoring has come out of the laboratory, entered numerous video conference rooms and security systems, and realized the leap from analog video communications to digital video communications.
Today, the video technology represented by Intelligent Video Surveillance (IVS) is in the application phase of going out of the laboratory and entering all walks of life. It will realize the leap from “mechanical†video information processing to intelligent video information processing. At present, China is in a critical period of this leap. Intelligent video surveillance is just one part of intelligent video technology, a part that brings video technology into the smart age.
The development of IVS is like many human technologies. Video technology is also the result of bionics guidance. Video technology mimics the function of the human eye and "pulls" the eye to practically inaccessible time and space. Television broadcasting provided a large number of audio-visual entertainment programs that enriched people's cultural life. The advent of video surveillance not only provided life-like on-the-spot information, but also solved problems in security and production, making people deeply feel the application of video technology. The breadth of the scope and the important role played.
Video surveillance has developed for 30 to 40 years. The earliest analog video surveillance equipment installed in the bank's business hall used video cameras and video tape recorders. Later, it was developed as digital surveillance. It was actually digital video recording (DVR). Camera, acquisition card plus computer (mainly used for compression, storage); network video surveillance devices in recent years, mainly composed of embedded cameras, transmission networks and computers.
After nearly four or five years of development, intelligent video surveillance has evolved from simple video surveillance to video surveillance with artificial intelligence.
The origin of video surveillance is very simple. From the perspective of demand, people have always needed to expand the functions of the eyes and hope to “hang†an “eye†in the necessary places for continuous monitoring; from the perspective of technical feasibility: Sensor technology has evolved from single-point sensors to multi-points, from one-dimensional sensors to two-dimensional and three-dimensional sensors (images, video sensors), and rapid development has provided the possibility of implementing surveillance-based monitoring.
In fact, this process once again demonstrates the tremendous influence of bionics on technology development: Monitoring technology, that is, the further development of intelligent video surveillance technology, is still the result of bionics guidance. From the point of view of bionics, the appearance of various sensors mimics the results of various human senses, from tactile sensations to auditory senses, and sights, such as microphones that imitate human ears and cameras that imitate human eyes.
How to imitate human brain function on the basis of imitating vision Recently, how to imitate human brain function on the basis of imitating vision, make existing monitoring equipment have certain intelligence and become the focus of research. Intelligent import makes the content and meaning of sensing have a subjective "feeling" and can give analysis, judgments and conclusions. It does not only faithfully record the physical quantity of the sensing object.
Early sensors were often simple, sensing one-dimensional, unambiguous physical quantities, and were easy to analyze and facilitate conclusions, such as temperature sensors, when the obtained temperature exceeds a certain threshold, it indicates that an abnormal situation has occurred. The visual sensor represented by a camera is a multi-dimensional, multi-sensor sensor: It is difficult to determine the meaning of the content of the recorded image, such as difficult to determine what happened to the scene the camera is facing.
At the present stage, the solution to this problem is mainly accomplished manually: On-duty personnel use the eyes to screen “hundreds†of camera information, perform “retrieval†of stored mass information, and perform manual comparison and analysis of suspicious information.
However, people (especially people's eyes and brains) are easily fatigued, and there are data indicating that if a person on duty is staring at the screen for more than ten minutes, he or she will unconsciously lose most of the information. In the London Underground case, for example, security personnel often spend hundreds of hours looking at large amounts of storage tapes to find valuable information. Looking for eyes and brain that will not fatigue, analyze the monitored content, reduce the burden, and improve monitoring efficiency. This is the goal of intelligent video surveillance.
Monitoring intelligence can not only help us keep a close eye on the screen and improve efficiency, but also can undertake and accomplish more work through “smartâ€. The most typical "smart" currently focuses on the following four aspects:
(1) Camera protection. This is the top priority. If the camera is not protected, everything else is free. Therefore, the camera can be automatically transferred, alarmed, and notified to the maintenance personnel in case of being blocked, moved, or blurred.
(2) Intrusion detection and tracking of moving targets. The timely detection of suspicious targets entering the surveillance area abnormally can identify the movement of single or multiple targets (such as the direction of movement, speed of movement, etc.). After the suspicious object is detected, a control instruction is sent to cause the camera to automatically track the target. After the object exceeds the monitoring range of the camera, the adjacent camera is automatically notified to work cooperatively, and the tracking is continued and an alarm signal is issued.
(3) Retention and moving objects alarm. When objects (such as parcels, suitcases, etc.) in a scene (such as waiting rooms, conference rooms, etc.) stay in sensitive areas for too long, or objects existing in the original scene (such as laptops, valuable instruments) are not When it is moved, the system will send an alarm signal and at the same time automatically find the suspicious person who placed the retentate or removed the original item in the previous video screen.
(4) analysis of group behavior. Include analysis of normal behaviors and abnormal behaviors such as crowd and traffic flow. It can analyze the normal behavior of groups in the scene, such as counting the number of people or vehicles crossing the entrance or designated area, highway traffic flow, and identifying the overall movement characteristics of the crowd, including speed and direction. It can also analyze and judge the abnormal behaviors of the groups in the scene, such as detecting, classifying, tracking and recording past pedestrians, vehicles and other suspicious objects, judging whether there are illegal vehicles on the road, whether there is a faulty vehicle, and whether there are pedestrians and vehicles. In the restricted area, there are long-term deterrence, stay, retrograde, etc., and detect whether there are abnormal behaviors such as crowd gathering, running, fighting in public places.
Of course, smart video surveillance can do more than this. It can help us do more things in the future, but the premise is that we must have enough intelligence to design and implement such an intelligent video surveillance system.
In the 1930s, video technology represented by television broadcasting came out of the laboratory, entered radio and television stations and millions of households, and achieved a leap from still image transmission to moving image transmission; at the end of the 1980s, conference television and video were used. The video technology represented by monitoring has come out of the laboratory, entered numerous video conference rooms and security systems, and realized the leap from analog video communications to digital video communications.
Today, the video technology represented by Intelligent Video Surveillance (IVS) is in the application phase of going out of the laboratory and entering all walks of life. It will realize the leap from “mechanical†video information processing to intelligent video information processing. At present, China is in a critical period of this leap. Intelligent video surveillance is just one part of intelligent video technology, a part that brings video technology into the smart age.
The development of IVS is like many human technologies. Video technology is also the result of bionics guidance. Video technology mimics the function of the human eye and "pulls" the eye to practically inaccessible time and space. Television broadcasting provided a large number of audio-visual entertainment programs that enriched people's cultural life. The advent of video surveillance not only provided life-like on-the-spot information, but also solved problems in security and production, making people deeply feel the application of video technology. The breadth of the scope and the important role played.
Video surveillance has developed for 30 to 40 years. The earliest analog video surveillance equipment installed in the bank's business hall used video cameras and video tape recorders. Later, it was developed as digital surveillance. It was actually digital video recording (DVR). Camera, acquisition card plus computer (mainly used for compression, storage); network video surveillance devices in recent years, mainly composed of embedded cameras, transmission networks and computers.
After nearly four or five years of development, intelligent video surveillance has evolved from simple video surveillance to video surveillance with artificial intelligence.
The origin of video surveillance is very simple. From the perspective of demand, people have always needed to expand the functions of the eyes and hope to “hang†an “eye†in the necessary places for continuous monitoring; from the perspective of technical feasibility: Sensor technology has evolved from single-point sensors to multi-points, from one-dimensional sensors to two-dimensional and three-dimensional sensors (images, video sensors), and rapid development has provided the possibility of implementing surveillance-based monitoring.
In fact, this process once again demonstrates the tremendous influence of bionics on technology development: Monitoring technology, that is, the further development of intelligent video surveillance technology, is still the result of bionics guidance. From the point of view of bionics, the appearance of various sensors mimics the results of various human senses, from tactile sensations to auditory senses, and sights, such as microphones that imitate human ears and cameras that imitate human eyes.
How to imitate human brain function on the basis of imitating vision Recently, how to imitate human brain function on the basis of imitating vision, make existing monitoring equipment have certain intelligence and become the focus of research. Intelligent import makes the content and meaning of sensing have a subjective "feeling" and can give analysis, judgments and conclusions. It does not only faithfully record the physical quantity of the sensing object.
Early sensors were often simple, sensing one-dimensional, unambiguous physical quantities, and were easy to analyze and facilitate conclusions, such as temperature sensors, when the obtained temperature exceeds a certain threshold, it indicates that an abnormal situation has occurred. The visual sensor represented by a camera is a multi-dimensional, multi-sensor sensor: It is difficult to determine the meaning of the content of the recorded image, such as difficult to determine what happened to the scene the camera is facing.
At the present stage, the solution to this problem is mainly accomplished manually: On-duty personnel use the eyes to screen “hundreds†of camera information, perform “retrieval†of stored mass information, and perform manual comparison and analysis of suspicious information.
However, people (especially people's eyes and brains) are easily fatigued, and there are data indicating that if a person on duty is staring at the screen for more than ten minutes, he or she will unconsciously lose most of the information. In the London Underground case, for example, security personnel often spend hundreds of hours looking at large amounts of storage tapes to find valuable information. Looking for eyes and brain that will not fatigue, analyze the monitored content, reduce the burden, and improve monitoring efficiency. This is the goal of intelligent video surveillance.
Monitoring intelligence can not only help us keep a close eye on the screen and improve efficiency, but also can undertake and accomplish more work through “smartâ€. The most typical "smart" currently focuses on the following four aspects:
(1) Camera protection. This is the top priority. If the camera is not protected, everything else is free. Therefore, the camera can be automatically transferred, alarmed, and notified to the maintenance personnel in case of being blocked, moved, or blurred.
(2) Intrusion detection and tracking of moving targets. The timely detection of suspicious targets entering the surveillance area abnormally can identify the movement of single or multiple targets (such as the direction of movement, speed of movement, etc.). After the suspicious object is detected, a control instruction is sent to cause the camera to automatically track the target. After the object exceeds the monitoring range of the camera, the adjacent camera is automatically notified to work cooperatively, and the tracking is continued and an alarm signal is issued.
(3) Retention and moving objects alarm. When objects (such as parcels, suitcases, etc.) in a scene (such as waiting rooms, conference rooms, etc.) stay in sensitive areas for too long, or objects existing in the original scene (such as laptops, valuable instruments) are not When it is moved, the system will send an alarm signal and at the same time automatically find the suspicious person who placed the retentate or removed the original item in the previous video screen.
(4) analysis of group behavior. Include analysis of normal behaviors and abnormal behaviors such as crowd and traffic flow. It can analyze the normal behavior of groups in the scene, such as counting the number of people or vehicles crossing the entrance or designated area, highway traffic flow, and identifying the overall movement characteristics of the crowd, including speed and direction. It can also analyze and judge the abnormal behaviors of the groups in the scene, such as detecting, classifying, tracking and recording past pedestrians, vehicles and other suspicious objects, judging whether there are illegal vehicles on the road, whether there is a faulty vehicle, and whether there are pedestrians and vehicles. In the restricted area, there are long-term deterrence, stay, retrograde, etc., and detect whether there are abnormal behaviors such as crowd gathering, running, fighting in public places.
Of course, smart video surveillance can do more than this. It can help us do more things in the future, but the premise is that we must have enough intelligence to design and implement such an intelligent video surveillance system.
Steel Structure Building,Steel Structure Workshop Co., Ltd. , http://www.nssbuilding.com