This article describes a new type of intelligent video surveillance technology based on high-definition satellite maps in the world, also known as target-based intelligent video surveillance. It uses high-definition satellite maps as a GIS interface, and integrates the intelligent video analysis results of the system-associated cameras into the video surveillance interface based on high-definition satellite maps. It can present target information simply and intuitively, thereby greatly improving the user experience.
GIS video monitoring interface
At present, the mainstream video surveillance system adopts the “Tic Tac Toe†method, which is to display video images on a video wall or a display. The advantage of this method is that it is easy to implement, allowing users to directly see the most original video information. However, there is no lack of global information in these video information, and the GIS video surveillance interface is a useful supplement to the traditional “Tic Tac Grid†video surveillance system. The advantages of GIS-based video surveillance interface are mainly reflected in:
1. It is convenient for users to quickly understand the global security situation, determine the location of the touch alarm, and quickly formulate a response plan. The use of the “Tic-To-Zoom†method to display video images is like watching a panther. Although many camera images are displayed, due to the large number of camera points and observation angles, it is difficult for users to quickly understand or judge the overall situation of the surveillance area. It can be imagined that when there is a touch alarm event, the monitoring personnel need to constantly identify the real position of the target according to the image, which not only increases the monitoring fatigue level, but also reduces the monitoring personnel experience satisfaction. The response is very passive. In fact, this is common in the video surveillance industry. problem.
In the GIS framework, local information can be effectively organized, which is not only a simple superposition of local information, it can help users refine and understand the global situation, such as the current trend of the monitored target movement, etc., in order to make more timely and effective decision-making.
2. It can naturally integrate with other non-video information to form a unified operation platform. In addition to video images, the GIS interface can easily integrate non-video sensor information such as GPS, access control, RFID, and firefighting, making it the most natural interface choice for multi-sensor information fusion.
3. New functions that can be implemented on the GIS interface:
New business functions: monitoring and alarming the vehicle from the specified movement trajectory;
The new usage management method: Quickly access and manage the monitoring points in the local area through the GIS interface. For example, a mouse frame or a click area at a key part automatically pops up the camera image of this area and the adjacent location. Quickly formulate alarm policies on the GIS interface without having to set them one by one in each camera.
High-definition satellite map GIS monitoring interface The current mainstream video surveillance GIS interface uses an electronic map similar to Google Maps. As a carrier of video image information, using video high-definition satellite maps is more natural, and it can also produce more advanced applications.
At present, the combination of electronic maps and intelligent video surveillance is mainly reflected in the fact that when the intelligent video surveillance detects anomalies, the corresponding surveillance areas can be represented by flickering and other forms. However, this kind of trigger is not precise enough, and it is a kind of “switch†expression. It is impossible to express information such as the trajectory of the target in the monitoring area, and naturally cannot express advanced functions such as automatic tracking.
Electronic maps can be understood as a simplified satellite map. However, these simplified information can be used to automatically establish a matching relationship between the satellite image and the monitoring image by a mathematical method so that the target information found in the camera can be accurately marked on the satellite image. come out. For example, the discovered target trajectory can be converted into the trajectory of the target on the satellite map, so that the moving target information can be uniformly presented on the satellite map.
Using the HD satellite image as a background can enhance the realism of the system and make it easier for the monitoring personnel to understand and accept. At the same time, satellite maps are not the icing on the cake, but they are necessary because of the satellite maps needed to establish the matching relationship between the camera and the GIS.
The standard for measuring the merits of the monitoring system
The highest security realm summarized by the United States security community is “Situational Awareness,†which is situational awareness. This provides "actionable", actionable decision information. The value of a video surveillance system is mainly reflected in whether it can provide users with timely, effective and operable information for a period of time after the triggering of the incident.
Reasonable camera points, clear images, spectacular video walls, and low false positives in intelligence analysis are all common requirements for video surveillance systems. However, if these requirements cannot be translated into actionable decision-making and emergency measures, a system that spends heavily may be an image project. Therefore, the "operability" of information is the fundamental criterion for measuring the merits of video surveillance systems.
Current video surveillance bottlenecks
At present, the main application mode of the video surveillance system is still “after investigation and evidence collectionâ€, and it is a matter of remedying the problem rather than preventing it. In the face of massive information, people hope that intelligent video analysis can effectively filter out the vast majority of useless information and push useful information to users. Intelligent video analysis is a very promising development direction, but there are still some problems. This article focuses on how to effectively present the results of intelligent video analytics to users. This is precisely the current problem that the monitoring industry considers less.
The monitoring information is independent. The existing mainstream intelligent video surveillance intelligently analyzes each image and pushes the analysis results to the user. The camera resources have not been coordinated and coordinated. The same moving target passes through multiple camera views and is reported as multiple independent video clips. Due to the different trigger conditions of each camera, when an object triggers an alarm within the field of view of a camera and moves to the other camera's field of view, it is likely to be considered as a normal moving object. Therefore, the mainstream intelligent video surveillance system or only a certain camera triggers an alarm. Due to the short time, the monitoring staff may not be able to react and lose the target.
A system with multi-camera coordination can maintain continuous tracking across touch targets. From multiplying the monitoring time of the target. Greatly improve the accuracy and effectiveness of decision-making. If the tracking time of a touch alert target is used as a criterion for assessing the value of a video surveillance system, the value of a system with multi-camera cooperative functions is many times that of an ordinary intelligent video surveillance system.
Target-based intelligent video surveillance
Target-based intelligent video surveillance technology is a new concept proposed by the international surveillance industry. It is relative to the current mainstream camera-based intelligent video surveillance technology, is its upgrade and extension. Mainstream camera-based technologies do not share and correlate information between cameras. Information sources are independent. Target-based technology emphasizes the breakthrough of the camera-based technology for information sharing, and collaboratively allocates monitoring resources within the system. Automatically correlate information from the same target between cameras. The ultimate goal is to allow users to “forget about†a camera without knowing the camera’s position and orientation. It only needs to be concerned with the business-related goals. It is therefore very straightforward and easy to understand the current security situation and develop it in a timely and effective manner. Response to countermeasures.
Making the expression of monitoring information simpler and more direct is based on the core idea of ​​intelligent video surveillance technology.
The essence of intelligent video surveillance is a massive information extraction and filtering process.
The mainstream intelligent video surveillance technology enables the extraction of actionable information elements. The goal-based intelligent video surveillance technology is to organize these elements effectively, further sort and filter the information, and present it to the user, so that the user feels simple and direct.
The two supporting technologies of the target-based intelligent video surveillance technology are: First, the coordination of multiple cameras (monitoring points); second, the presentation of integrated information based on satellite maps. The following figure describes the relationship between the two support technologies and the target-based monitoring.
Key technologies for achieving multi-camera collaboration
Camera target "handshake", collaborative technology
GIS (High Definition Satellite Map) is the ideal platform for multi-camera collaboration technology. If the "Ticium Grid" form of monitoring is used, the correlation between cameras is very difficult to establish. The same target between adjacent cameras due to the perspective, zoom, imaging sensitivity, color differences, etc., will cause greater difficulty in the target "handshaking" (Re-identification) between different cameras. By using GIS as a unified platform, the targets between different cameras can be projected onto the GIS so that the position constraint can be used to effectively "handshake" and the effectiveness of multi-camera collaboration is greatly improved.
Considering multiple cameras at the system level can naturally introduce the concept of resource allocation. It is the pre-judgment of the direction in which the sports goal is moving and the monitoring resources are scheduled in advance. For example, let the PTZ camera “greet†a moving object at a certain preset position, and multiple PTZ cameras can be set up at the forked port for preparation, thereby improving the success rate of “handshaking†and covering a large area of ​​monitoring area with a small number of cameras. the goal of.
Dynamic platform tracking technology
Target tracking in a fixed camera is a mature technology. In order to cover a large area of ​​surveillance with a small number of cameras, it is usually necessary to use a dynamic platform tracking technology, that is, a PTZ target tracking technology.
Compared to a fixed camera, it is difficult to keep track of the target on the PTZ camera because there is no support for the background model. Especially when the camera is required to zoom in to zoom in on the target, and the target turns to cause a large change in the shape of the target image, the target is easily lost. Reliable PTZ tracking is currently a very advanced technology and is also an active international research direction.
At present, there are a small number of products with PTZ tracking function on the market, but manufacturers do not focus on promoting this feature. Therefore, when selecting such products, we must pay attention to the actual results. Separate PTZ tracking products have limited monitoring efficiency for larger monitoring systems (dozens of or more probes). PTZ tracking products can be deployed and coordinated within the system, so as to truly play a multi-camera collaboration efficacy.
Pattern recognition technology
One of the characteristics of intelligent video analysis is that it can identify patterns of moving objects such as people, cars, big trucks, bicycles (motorcycles), and crowds. These moving objects can use icons to represent their attributes and trajectories on the GIS interface. Person icon, car icon. For the user does not care about the sports object can not be displayed, so that the massive image information is effectively filtered, the amount of information controlled within the physiologically acceptable range of people.
The pattern recognition technology based on human-vehicle classification and license plate recognition and face recognition belong to a class of technologies. However, due to relatively more use environments, different sizes of people and vehicles, and more changing angles, maturity is relatively poor. However, such technologies are rapidly being put to practical use, and it is expected that the accuracy of face recognition will be approached or achieved in the near future.
Camera Information Satellite Image Integration Technology
Surveillance camera images and satellite maps are essentially images.
By matching the common features between them, the mapping relationship between the surveillance camera image and the satellite image can be achieved. In other words, there is a one-to-one mapping relationship between the pixels in the camera and the position of the satellite map. The trajectory of the moving object found in the camera can also be expressed in the satellite map.
Each camera “projects†the target information in the corresponding area on the satellite map, forming a complete, real-time global situation of the moving target.
The integration of camera information satellite imagery is essentially the transformation of different camera angles through mathematical transformation into a unified satellite perspective from a high altitude overlooking the confusion caused by the different perspectives on the judgment of the overall situation. It can be simply and directly. Express target information.
Conclusion
This article describes a current state-of-the-art intelligent video surveillance model based on goals. The core idea is to allow users to "forget about" the camera, "forget about" the traditional monitoring methods in the form of tessellation, and use the monitoring system from the perspective of the user's real concern. Its core technology is multi-camera collaboration and global information integration based on high-definition satellite maps. By associating the video analysis results between different cameras and projecting them onto an integrated satellite map platform, the monitoring personnel can help overcome the confusion caused by the lack of global concepts, numerous information sources, and different perspectives when using traditional monitoring technologies, and can be easily and timely controlled. The current security situation quickly forms a response plan.
GIS video monitoring interface
At present, the mainstream video surveillance system adopts the “Tic Tac Toe†method, which is to display video images on a video wall or a display. The advantage of this method is that it is easy to implement, allowing users to directly see the most original video information. However, there is no lack of global information in these video information, and the GIS video surveillance interface is a useful supplement to the traditional “Tic Tac Grid†video surveillance system. The advantages of GIS-based video surveillance interface are mainly reflected in:
1. It is convenient for users to quickly understand the global security situation, determine the location of the touch alarm, and quickly formulate a response plan. The use of the “Tic-To-Zoom†method to display video images is like watching a panther. Although many camera images are displayed, due to the large number of camera points and observation angles, it is difficult for users to quickly understand or judge the overall situation of the surveillance area. It can be imagined that when there is a touch alarm event, the monitoring personnel need to constantly identify the real position of the target according to the image, which not only increases the monitoring fatigue level, but also reduces the monitoring personnel experience satisfaction. The response is very passive. In fact, this is common in the video surveillance industry. problem.
In the GIS framework, local information can be effectively organized, which is not only a simple superposition of local information, it can help users refine and understand the global situation, such as the current trend of the monitored target movement, etc., in order to make more timely and effective decision-making.
2. It can naturally integrate with other non-video information to form a unified operation platform. In addition to video images, the GIS interface can easily integrate non-video sensor information such as GPS, access control, RFID, and firefighting, making it the most natural interface choice for multi-sensor information fusion.
3. New functions that can be implemented on the GIS interface:
New business functions: monitoring and alarming the vehicle from the specified movement trajectory;
The new usage management method: Quickly access and manage the monitoring points in the local area through the GIS interface. For example, a mouse frame or a click area at a key part automatically pops up the camera image of this area and the adjacent location. Quickly formulate alarm policies on the GIS interface without having to set them one by one in each camera.
High-definition satellite map GIS monitoring interface The current mainstream video surveillance GIS interface uses an electronic map similar to Google Maps. As a carrier of video image information, using video high-definition satellite maps is more natural, and it can also produce more advanced applications.
At present, the combination of electronic maps and intelligent video surveillance is mainly reflected in the fact that when the intelligent video surveillance detects anomalies, the corresponding surveillance areas can be represented by flickering and other forms. However, this kind of trigger is not precise enough, and it is a kind of “switch†expression. It is impossible to express information such as the trajectory of the target in the monitoring area, and naturally cannot express advanced functions such as automatic tracking.
Electronic maps can be understood as a simplified satellite map. However, these simplified information can be used to automatically establish a matching relationship between the satellite image and the monitoring image by a mathematical method so that the target information found in the camera can be accurately marked on the satellite image. come out. For example, the discovered target trajectory can be converted into the trajectory of the target on the satellite map, so that the moving target information can be uniformly presented on the satellite map.
Using the HD satellite image as a background can enhance the realism of the system and make it easier for the monitoring personnel to understand and accept. At the same time, satellite maps are not the icing on the cake, but they are necessary because of the satellite maps needed to establish the matching relationship between the camera and the GIS.
The standard for measuring the merits of the monitoring system
The highest security realm summarized by the United States security community is “Situational Awareness,†which is situational awareness. This provides "actionable", actionable decision information. The value of a video surveillance system is mainly reflected in whether it can provide users with timely, effective and operable information for a period of time after the triggering of the incident.
Reasonable camera points, clear images, spectacular video walls, and low false positives in intelligence analysis are all common requirements for video surveillance systems. However, if these requirements cannot be translated into actionable decision-making and emergency measures, a system that spends heavily may be an image project. Therefore, the "operability" of information is the fundamental criterion for measuring the merits of video surveillance systems.
Current video surveillance bottlenecks
At present, the main application mode of the video surveillance system is still “after investigation and evidence collectionâ€, and it is a matter of remedying the problem rather than preventing it. In the face of massive information, people hope that intelligent video analysis can effectively filter out the vast majority of useless information and push useful information to users. Intelligent video analysis is a very promising development direction, but there are still some problems. This article focuses on how to effectively present the results of intelligent video analytics to users. This is precisely the current problem that the monitoring industry considers less.
The monitoring information is independent. The existing mainstream intelligent video surveillance intelligently analyzes each image and pushes the analysis results to the user. The camera resources have not been coordinated and coordinated. The same moving target passes through multiple camera views and is reported as multiple independent video clips. Due to the different trigger conditions of each camera, when an object triggers an alarm within the field of view of a camera and moves to the other camera's field of view, it is likely to be considered as a normal moving object. Therefore, the mainstream intelligent video surveillance system or only a certain camera triggers an alarm. Due to the short time, the monitoring staff may not be able to react and lose the target.
A system with multi-camera coordination can maintain continuous tracking across touch targets. From multiplying the monitoring time of the target. Greatly improve the accuracy and effectiveness of decision-making. If the tracking time of a touch alert target is used as a criterion for assessing the value of a video surveillance system, the value of a system with multi-camera cooperative functions is many times that of an ordinary intelligent video surveillance system.
Target-based intelligent video surveillance
Target-based intelligent video surveillance technology is a new concept proposed by the international surveillance industry. It is relative to the current mainstream camera-based intelligent video surveillance technology, is its upgrade and extension. Mainstream camera-based technologies do not share and correlate information between cameras. Information sources are independent. Target-based technology emphasizes the breakthrough of the camera-based technology for information sharing, and collaboratively allocates monitoring resources within the system. Automatically correlate information from the same target between cameras. The ultimate goal is to allow users to “forget about†a camera without knowing the camera’s position and orientation. It only needs to be concerned with the business-related goals. It is therefore very straightforward and easy to understand the current security situation and develop it in a timely and effective manner. Response to countermeasures.
Making the expression of monitoring information simpler and more direct is based on the core idea of ​​intelligent video surveillance technology.
The essence of intelligent video surveillance is a massive information extraction and filtering process.
The mainstream intelligent video surveillance technology enables the extraction of actionable information elements. The goal-based intelligent video surveillance technology is to organize these elements effectively, further sort and filter the information, and present it to the user, so that the user feels simple and direct.
The two supporting technologies of the target-based intelligent video surveillance technology are: First, the coordination of multiple cameras (monitoring points); second, the presentation of integrated information based on satellite maps. The following figure describes the relationship between the two support technologies and the target-based monitoring.
Key technologies for achieving multi-camera collaboration
Camera target "handshake", collaborative technology
GIS (High Definition Satellite Map) is the ideal platform for multi-camera collaboration technology. If the "Ticium Grid" form of monitoring is used, the correlation between cameras is very difficult to establish. The same target between adjacent cameras due to the perspective, zoom, imaging sensitivity, color differences, etc., will cause greater difficulty in the target "handshaking" (Re-identification) between different cameras. By using GIS as a unified platform, the targets between different cameras can be projected onto the GIS so that the position constraint can be used to effectively "handshake" and the effectiveness of multi-camera collaboration is greatly improved.
Considering multiple cameras at the system level can naturally introduce the concept of resource allocation. It is the pre-judgment of the direction in which the sports goal is moving and the monitoring resources are scheduled in advance. For example, let the PTZ camera “greet†a moving object at a certain preset position, and multiple PTZ cameras can be set up at the forked port for preparation, thereby improving the success rate of “handshaking†and covering a large area of ​​monitoring area with a small number of cameras. the goal of.
Dynamic platform tracking technology
Target tracking in a fixed camera is a mature technology. In order to cover a large area of ​​surveillance with a small number of cameras, it is usually necessary to use a dynamic platform tracking technology, that is, a PTZ target tracking technology.
Compared to a fixed camera, it is difficult to keep track of the target on the PTZ camera because there is no support for the background model. Especially when the camera is required to zoom in to zoom in on the target, and the target turns to cause a large change in the shape of the target image, the target is easily lost. Reliable PTZ tracking is currently a very advanced technology and is also an active international research direction.
At present, there are a small number of products with PTZ tracking function on the market, but manufacturers do not focus on promoting this feature. Therefore, when selecting such products, we must pay attention to the actual results. Separate PTZ tracking products have limited monitoring efficiency for larger monitoring systems (dozens of or more probes). PTZ tracking products can be deployed and coordinated within the system, so as to truly play a multi-camera collaboration efficacy.
Pattern recognition technology
One of the characteristics of intelligent video analysis is that it can identify patterns of moving objects such as people, cars, big trucks, bicycles (motorcycles), and crowds. These moving objects can use icons to represent their attributes and trajectories on the GIS interface. Person icon, car icon. For the user does not care about the sports object can not be displayed, so that the massive image information is effectively filtered, the amount of information controlled within the physiologically acceptable range of people.
The pattern recognition technology based on human-vehicle classification and license plate recognition and face recognition belong to a class of technologies. However, due to relatively more use environments, different sizes of people and vehicles, and more changing angles, maturity is relatively poor. However, such technologies are rapidly being put to practical use, and it is expected that the accuracy of face recognition will be approached or achieved in the near future.
Camera Information Satellite Image Integration Technology
Surveillance camera images and satellite maps are essentially images.
By matching the common features between them, the mapping relationship between the surveillance camera image and the satellite image can be achieved. In other words, there is a one-to-one mapping relationship between the pixels in the camera and the position of the satellite map. The trajectory of the moving object found in the camera can also be expressed in the satellite map.
Each camera “projects†the target information in the corresponding area on the satellite map, forming a complete, real-time global situation of the moving target.
The integration of camera information satellite imagery is essentially the transformation of different camera angles through mathematical transformation into a unified satellite perspective from a high altitude overlooking the confusion caused by the different perspectives on the judgment of the overall situation. It can be simply and directly. Express target information.
Conclusion
This article describes a current state-of-the-art intelligent video surveillance model based on goals. The core idea is to allow users to "forget about" the camera, "forget about" the traditional monitoring methods in the form of tessellation, and use the monitoring system from the perspective of the user's real concern. Its core technology is multi-camera collaboration and global information integration based on high-definition satellite maps. By associating the video analysis results between different cameras and projecting them onto an integrated satellite map platform, the monitoring personnel can help overcome the confusion caused by the lack of global concepts, numerous information sources, and different perspectives when using traditional monitoring technologies, and can be easily and timely controlled. The current security situation quickly forms a response plan.
Oyster Extract Powder,Oyster Extract,Plant Extract
Lower Blood Pressure Co., Ltd. , http://www.bjorganicfertilizer.com