CN108174147B - Automatic optimization allocation method for video source and tracking server in video monitoring system - Google Patents

Abstract

The invention provides an automatic optimization allocation method of video sources and tracking servers in a video monitoring system, which comprises the following steps: 1. establishing a server cluster with a tree logic structure; 2. determining the range of a camera to be analyzed; 3. acquiring an IP address of a camera; 4. distributing a corresponding tracking server for the camera; 5. determining a time slice; 6. and (6) tracking. The invention has the beneficial effects that: after an accident occurs at a certain camera, the camera shoots a video source with accident specific target information, the video source camera a is determined as a starting point, n cameras in a coverage area with the reachable path length r as the radius are determined as query areas, the tracking server analyzes the video source in a time slice according to the specific target information, whether the target appears in the video source shot by the camera distributed to the tracking server is judged, the escaping path and the position of an escaping suspected person are determined, and the efficiency of tracking the escaping suspected person by using a video monitoring system is improved.

Description

Automatic optimal allocation method of video source and tracking server in video surveillance system

technical field

The invention relates to the technical field of urban video surveillance security, in particular to an automatic optimal distribution method of a video source and a tracking server in a video surveillance system.

Background technique

At present, the pursuit of escaped suspects after the accident still relies on manpower to check the surveillance camera images of the accident site and its vicinity. The method of human inspection and monitoring has helped our country to solve a large number of cases. The places where the cases occurred cover all parts of the country, and the number of solved cases is large. Taking Laixi City, Qingdao, Shandong as an example, from September 8, 2015 to September 8, 2016, the city directly cracked more than 770 criminal cases of various types through the video surveillance system. Using manpower to check and monitor the method of chasing escape suspects is the most common method of chasing escape suspects today.

In a long historical period before, manpower viewing surveillance images to hunt down escaped suspects played an important role in catching suspects. However, with the continuation of the time after the incident, the escape range of the suspect has gradually expanded, and it will become more and more difficult to search for the suspect in the actual pursuit and investigation, and more and more manpower will be required. In the later period, although additional manpower was dispatched to speed up the inspection and monitoring, due to the limitation of the total manpower, the method of chasing escaped suspects through manpower was not efficient. Obviously, efficiency is limited.

At present, the reasons for the inefficiency of chasing escaped suspects are as follows:

①. The efficiency of human watching surveillance video is low

In the process of watching surveillance videos to find suspects, manpower needs to watch a large number of video materials and find escaped suspects in a large number of video materials. Because the vast majority of the video data are pictures of escape suspects that do not appear, fatigue after long-term viewing will increase the probability of missing escape suspects.

②.Insufficient manpower to hunt down and long hunting time

Due to the limitation of the sum of manpower, there is a maximum manpower to hunt down escaped suspects. However, in the process of manpower watching surveillance videos to find escape suspects, the lack of manpower leads to slow viewing of surveillance videos. The longer it takes to hunt down the suspect, the greater the escape range of the suspect and the need to watch more surveillance videos. Such a vicious circle . Suspects are more likely to escape.

③. The geographic information search of surveillance cameras is unscientific

The process of watching surveillance cameras manually requires manual correspondence between cameras and geographic information, which is inefficient and cannot be guaranteed to be correct.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide an automatic optimal allocation method for video sources and tracking servers in a video surveillance system, so as to solve the technical problem that the current video surveillance system has low efficiency in tracking escaped suspects.

The present invention provides an automatic optimal allocation method for a video source and a tracking server in a video surveillance system, comprising the following steps:

Step 1. Establish a server cluster with a tree-like logical structure

The server cluster includes a video tracking information scheduling server, multiple domain name resolution servers, a query server and multiple tracking servers. The video tracking information scheduling server processes information from other servers and is the link between different types of servers. , the domain name resolution server adopts a tree-like logic structure to resolve the domain name into IP, the query server dynamically allocates a tracking server for each IP, and the tracking server is used for video analysis in the camera;

Step 2. Determine the range of the camera to be analyzed

Take the video source camera a as the starting point, and take the n cameras in the coverage area with the reachable path length r as the radius as the cameras to be analyzed, use the map to know the path length, and assign the priority to each camera in the order of increasing path length, The camera with the shortest path length has the highest priority;

Step 3. Obtain the IP address of the camera

In the Internet, each camera is assigned an IP address as a globally unique identifier. The video tracking information scheduling server will take the video source camera a as the starting point, and the n cameras in the coverage area with the reachable path length r as the radius will send the domain name resolution request to the domain name resolution server according to the priority level, and the domain name resolution server uses a tree It is a logical structure that divides the management tasks of the entire domain name space into multiple parts, which are managed by the domain name resolution server on each sub-node. Finally, the video source camera a is the starting point, and the reachable path length r is the coverage area of the radius. The IP addresses of the n cameras are sent to the video tracking information scheduling server;

Step 4. Assign the corresponding tracking server to the camera

The video tracking information dispatch server sends the resolved IP address to the query server, and the query server allocates an initialized empty queue space for each tracking server according to the principle of dynamic optimization and allocation, which is used to store the corresponding cameras that the tracking server needs to process. IP address; assign a tracking server to each IP according to the order in which the IP enters the query server; if all current tracking servers are occupied, the newly entered IP will be assigned to the tracking server with the shortest queue length; when the server with the shortest queue length When the number is greater than 1, it will be randomly selected for allocation; after the video is parsed, its corresponding IP address will be deleted from the queue of the tracking server;

Step 5. Determine the time slice

Using the map to know the distance between the camera in the area and the video source camera a, the start time and end time of the intercepted video can be expressed as:

in,

V _c is the fastest escape speed of the escaping suspect;

V _x is the slowest escape velocity of the escape suspect;

s _i is the distance from the ith camera to a, i=1,2,...,n;

Then the time slice of the video transmitted by the i-th camera can be expressed as where t ₀ is the time when the specific target information in a appears;

Step 6. Tracking

The video tracking information scheduling server transmits the specific target information and the video sources within the time slice of n cameras in the coverage area with the video source camera a as the starting point and the reachable path length r as the radius to the tracking server assigned by the cameras. , the corresponding tracking server analyzes the video source according to the specific target information, and determines whether the target appears in the video source.

Further, in step 3, obtaining the IP addresses of n cameras in the coverage area with the video source camera a as the starting point and the reachable path length r as the radius specifically includes the following steps:

Step 31, the video tracking information scheduling server first sends the domain name resolution request to the local domain name resolution server;

Step 32: After the local domain name resolution server receives the request, it first queries the local cache. If the local cache has a record item corresponding to the domain name and the IP address, the local domain name resolution server directly returns the query result to the video tracking information scheduling server. ;

Step 33: If the local cache does not have a record item corresponding to the domain name and IP address, the local domain name resolution server sends the domain name resolution request to the domain name resolution server on the root node, and then the domain name resolution server on the root node returns to the local domain name. The address of the domain name resolution server on the child node under a root node of the resolution server;

Step 34, the local domain name resolution server sends a request to the domain name resolution server on the child node in step 33, and then the domain name resolution server that accepts the request queries its own cache, if there is no record item corresponding to the domain name and IP address, then returns to the next The address of the domain name resolution server at the highest level;

Step 35, repeat step 34, until the correct domain name resolution server address is found, and the resolved IP address is sent to the local domain name resolution server;

Step 36: The local domain name resolution server sends the IP address of the camera to the video tracking information scheduling server.

Further, in step 3, the query method from the video tracking information scheduling server to the local domain name resolution server is recursive query.

Further, in step 3, the query mode between the domain name resolution servers is an iterative query.

Further, in step 4, allocating the corresponding tracking server to the camera specifically includes the following steps:

Step 41. Query whether there is a tracking server with an empty queue, and assign an IP to the server if there is;

Step 42: When there is no tracking server with an empty queue, query the queue lengths of all tracking servers;

Step 43: Select the tracking server with the shortest queue length, and add the IP corresponding to the camera to the queue of the tracking server. When the number of tracking servers with the shortest queue length is greater than 1, it will be randomly allocated among the alternative tracking servers;

Step 44. Repeat steps 41-43 if there are multiple IPs;

Step 45: After the tracking server completes the video parsing of the current IP, it deletes the current IP from the queue.

Compared with the prior art, the automatic optimal distribution method of the video source and the tracking server in the video surveillance system of the present invention has the following characteristics and advantages:

According to the automatic optimal distribution method of video source and tracking server in the video surveillance system of the present invention, after an accident occurs at a certain camera, the camera shoots the video source with the specific target information of the accident (such as the facial features of the escaped suspect), and through this The method determines that with the video source camera a as the starting point and n cameras in the coverage area with the reachable path length r as the radius as the query area, the tracking server analyzes the video source in the time slice according to the specific target information, and determines whether the target appears in the In the video source captured by the camera assigned to the tracking server, the escape path and location of the escaped suspect are determined, and the efficiency of tracking the escaped suspect by using the video surveillance system is improved.

The features and advantages of the present invention will become more apparent after reading the detailed description of the present invention in conjunction with the accompanying drawings.

Description of drawings

In order to illustrate the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are For some embodiments of the present invention, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without creative efforts.

1 is a flowchart of an automatic optimal allocation method for a video source and a tracking server in a video surveillance system according to an embodiment of the present invention;

2 is a schematic diagram of a server cluster in an automatic optimal allocation method for a video source and a tracking server in a video surveillance system according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a server cluster tree logical structure in an automatic optimal allocation method for video sources and tracking servers in a video surveillance system according to an embodiment of the present invention.

Detailed ways

Referring to FIGS. 1 to 3 , the method for automatically optimizing the distribution of video sources and tracking servers in the video surveillance system of the present embodiment includes the following steps:

Step 1. Establish a server cluster with a tree-like logical structure

The server cluster includes a video tracking information scheduling server, multiple domain name resolution servers, a query server and multiple tracking servers. The video tracking information scheduling server processes information from other servers and is the link between different types of servers. , the domain name resolution server adopts a tree-like logic structure to resolve the domain name into IP, the query server dynamically allocates a tracking server for each IP, and the tracking server is used for video analysis in the camera;

Step 2. Determine the range of the camera to be analyzed

Take the video source camera a as the starting point, and take the n cameras in the coverage area with the reachable path length r as the radius as the cameras to be analyzed, use the map to know the path length, and assign the priority to each camera in the order of increasing path length, The camera with the shortest path length has the highest priority. For example, if the video source camera a is located at the entrance of a supermarket, take the supermarket entrance as the center, and take the n cameras in the coverage area with the reachable path length r as the radius as the cameras to be analyzed, and the distance from the video source camera a in this area is The closest camera is used as the first parsed camera.

Step 3. Obtain the IP address of the camera

In the Internet, each camera is assigned an IP address as a globally unique identifier. The video tracking information scheduling server sends the domain name resolution requests of n cameras within the coverage area with the video source camera a as the starting point and the reachable path length r as the radius to the domain name resolution server. The management tasks of the entire domain name space are divided into multiple parts, which are managed by the domain name resolution server on each sub-node. Finally, the video source camera a is the starting point, and the coverage area with the reachable path length r is the radius. The IP address is sent to the video tracking information scheduling server.

Above, determining the IP addresses of n cameras in the coverage area with the video source camera a as the starting point and the reachable path length r as the radius specifically includes the following steps:

Step 31, the video tracking information scheduling server first sends the domain name resolution request to the local domain name resolution server;

Step 32: After the local domain name resolution server receives the request, it first queries the local cache. If the local cache has a record item corresponding to the domain name and the IP address, the local domain name resolution server directly returns the query result to the video tracking information scheduling server. ;

Step 33: If the local cache does not have a record item corresponding to the domain name and IP address, the local domain name resolution server sends the domain name resolution request to the domain name resolution server on the root node, and then the domain name resolution server on the root node returns to the local domain name. The address of the domain name resolution server on the child node under a root node of the resolution server;

Step 34, the local domain name resolution server sends a request to the domain name resolution server on the child node in step 33, and then the domain name resolution server that accepts the request queries its own cache, if there is no record item corresponding to the domain name and IP address, then returns to the next The address of the domain name resolution server at the highest level;

Step 35, repeat step 34, until the correct domain name resolution server address is found, and the resolved IP address is sent to the local domain name resolution server;

Step 36: The local domain name resolution server sends the IP address of the camera to the video tracking information scheduling server.

In step 3, the query method from the video tracking information scheduling server to the local domain name resolution server is recursive query. The query process of recursive query is as follows: if the video tracking information scheduling server asks for an address that the local DNS server does not know, then the local DNS server will continue to send a query request to its root DNS server, instead of letting the video tracking The information scheduling server performs the next query by itself; when the local domain name resolution server replaces the video tracking information scheduling server to query other domain name resolution servers, the video tracking information scheduling server is completely in a waiting state, and there are only two results returned: the query succeeds or the query fails.

In step 3, the query mode between the domain name resolution servers is an iterative query. The query process of the iterative query is as follows: when the root DNS server receives the query request from the local DNS server, it either gives the address to be queried, or tells the local DNS server "Which DNS server should you go to next? Query", and then let the local domain name resolution server perform subsequent queries; the root domain name resolution server informs the local domain name resolution server of the address of the next-level domain name resolution server that it knows, and asks the local domain name resolution server to query the next-level domain name resolution server. , there are only two kinds of results returned: the best query point or the IP address.

Step 4. Assign the corresponding tracking server to the camera

The video tracking information dispatch server sends the resolved IP address to the query server, and the query server allocates an initialized empty queue space for each tracking server according to the principle of dynamic optimization and allocation, which is used to store the corresponding cameras that the tracking server needs to process. IP address; assign a tracking server to each IP according to the order in which the IP enters the query server; if all current tracking servers are occupied, the newly entered IP will be assigned to the tracking server with the shortest queue length; when the server with the shortest queue length When the number is greater than 1, it will be randomly selected for allocation; after the video is parsed, its corresponding IP address will be deleted from the queue of the tracking server.

Above, allocating the corresponding tracking server to the camera specifically includes the following steps:

Step 41. Query whether there is a tracking server with an empty queue, and assign an IP to the server if there is;

Step 42: When there is no tracking server with an empty queue, query the queue lengths of all tracking servers;

Step 43: Select the tracking server with the shortest queue length, and add the IP corresponding to the camera to the queue of the tracking server. When the number of tracking servers with the shortest queue length is greater than 1, it will be randomly allocated among the alternative tracking servers;

Step 44. Repeat steps 41-43 if there are multiple IPs;

Step 45. After the tracking server completes the video analysis of the current IP, it deletes the current IP from the queue

Step 5. Determine the time slice

Using the map to know the distance between the camera in the area and the video source camera a, the start time and end time of the intercepted video can be expressed as:

in,

V _c is the fastest escape speed of the escape suspect, such as the average speed of the vehicle driven by the escape suspect;

V _x is the slowest escape velocity of the escaping suspect, such as the average walking speed of the escaping suspect;

s _i is the distance from the ith camera to a, i=1,2,...,n;

其中t₀为a中特定目标信息出现的时间；Then the time slice of the video transmitted by the i-th camera can be expressed as

where t ₀ is the time when the specific target information in a appears;

Step 6. Tracking

The video tracking information scheduling server transmits the specific target information and the video sources within the time slice of n cameras in the coverage area with the video source camera a as the starting point and the reachable path length r as the radius to the tracking server assigned by the cameras. , the corresponding tracking server analyzes the video source according to the specific target information, and determines whether the target appears in the video source.

In the method for automatically optimizing the distribution of video sources and tracking servers in the video surveillance system of this embodiment, after an accident occurs at a certain camera, the camera shoots a video source with specific target information of the accident (such as the facial features of the escaped suspect), This method determines that the video source camera a is the starting point, and n cameras in the coverage area with the reachable path length r as the radius are the query area, and the tracking server analyzes the video source in the time slice according to the specific target information, and judges whether the target appears. From the video source shot by the camera assigned to the tracking server, the escape path and location of the escaped suspect are determined, and the efficiency of tracking the escaped suspect by using the video surveillance system is improved.

Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by those skilled in the art within the essential scope of the present invention should also belong to the present invention. the scope of protection of the invention.

Claims (5)

1. An automatic optimization allocation method for video sources and tracking servers in a video monitoring system is characterized by comprising the following steps:

step one, establishing a server cluster with a tree logic structure

The server cluster comprises a video tracking information scheduling server, a plurality of domain name resolution servers, an inquiry server and a plurality of tracking servers, wherein the video tracking information scheduling server processes information from other servers and is a link for communication among different types of servers;

step two, determining the range of the camera to be analyzed

The video source camera a is used as a starting point, n cameras in a coverage area with the reachable path length r as the radius are used as cameras to be analyzed, the path length can be known by using a map, the priorities are distributed to the cameras in the sequence that the path lengths are sequentially increased, and the camera with the shortest path length has the highest priority;

step three, acquiring IP address of camera

In the internet, each camera is assigned an IP address as a global unique identifier, a video tracking information scheduling server sequentially sends domain name resolution requests to a domain name resolution server according to the priority level by taking a video source camera a as a starting point and n cameras in a coverage area with the reachable path length r as the radius, the domain name resolution server adopts a tree-shaped logical structure, divides the management task of the whole domain name space into a plurality of parts, is respectively managed by the domain name resolution server on each sub-node, and finally sends the IP addresses of the n cameras in the coverage area with the reachable path length r as the radius to the video tracking information scheduling server by taking the video source camera a as the starting point;

step four, distributing corresponding tracking servers for the cameras

The video tracking information scheduling server sends the analyzed IP address to the query server, and the query server allocates an initialized empty queue space for each tracking server according to a dynamic optimization allocation principle to store the IP address corresponding to the camera which needs to be processed by the tracking server; allocating a tracking server for each IP address according to the sequence of the IP addresses entering the query server; if all the tracking servers are occupied currently, allocating the newly entered IP address to the tracking server with the shortest queue length; when the number of the servers with the shortest queue length is greater than 1, randomly selecting the servers from the servers to distribute; after the video is analyzed, deleting the corresponding IP address from the queue of the tracking server;

step five, determining time slices

The distance between the camera in the area and the video source camera a can be known by using a map, and then the starting time and the ending time of the intercepted video can be expressed by a formula as follows:

wherein,

V_cthe highest escape speed is achieved for the escaping suspects;

V_xthe slowest escape speed of the escaping suspects;

s_ithe distance from the ith camera to a, i is 1, 2.

The time slice for the ith camera to transmit video can be expressed asWherein t is₀Is the time when the specific target information in a appears;

step six, tracing

The video tracking information scheduling server transmits specific target information and video sources in time slices of n cameras in a coverage area with a video source camera a as a starting point and with an reachable path length r as a radius to the tracking servers distributed by the cameras, and the corresponding tracking server analyzes the video sources according to the specific target information and judges whether the target appears in the video sources.

2. The method according to claim 1, wherein the step three of obtaining IP addresses of n cameras in a coverage area with a starting point of a video source camera a and a radius of an achievable path length r specifically comprises the steps of:

step 31, the video tracking information scheduling server sends the domain name resolution request to a local domain name resolution server;

step 32, after receiving the request, the local domain name resolution server queries a local cache, and if the local cache has a record item corresponding to the domain name and the IP address, the local domain name resolution server directly returns a query result to the video tracking information scheduling server;

step 33, if the local cache does not have the record item corresponding to the domain name and the IP address, the local domain name resolution server sends the domain name resolution request to the domain name resolution server on the root node, and then the domain name resolution server on the root node returns the address of the domain name resolution server on a child node under the root node of the local domain name resolution server;

step 34, the local domain name resolution server sends a request to the domain name resolution server on the child node in step 33, then the domain name resolution server receiving the request queries its own cache, and if there is no record item corresponding to the domain name and the IP address, the address of the domain name resolution server of the next level is returned;

step 35, repeating step 34 until a correct domain name resolution server address is found, and sending the resolved IP address to a local domain name resolution server;

and step 36, the local domain name resolution server sends the IP address of the camera to the video tracking information scheduling server.

3. The method according to claim 2, wherein in step three, the query from the video tracking information scheduling server to the local domain name resolution server is a recursive query.

4. The method according to claim 2, wherein in step three, the query mode between the domain name resolution servers is iterative query.

5. The method for automatically and optimally allocating video sources and tracking servers in a video monitoring system according to claim 1, wherein in the fourth step, the step of dynamically and optimally allocating tracking servers to cameras specifically comprises the following steps:

step 41, inquiring whether a tracking server with an empty queue exists or not, and if so, allocating an IP address to the server;

step 42, when no tracking server with empty queue exists, inquiring the queue length of all tracking servers;

step 43, selecting the tracking server with the shortest queue length, adding the IP address corresponding to the camera to the queue of the tracking server, and randomly allocating the tracking servers in the alternative tracking servers when the number of the tracking servers with the shortest queue length is greater than 1;

step 44, repeating steps 41-43 if there are multiple IP addresses;

and step 45, after the tracking server completes video analysis of the current IP address, deleting the current IP address from the queue.

Images