CN112348856A - Multi-camera collaborative analysis method and system based on network system topological structure - Google Patents
Multi-camera collaborative analysis method and system based on network system topological structure Download PDFInfo
- Publication number
- CN112348856A CN112348856A CN201910735108.8A CN201910735108A CN112348856A CN 112348856 A CN112348856 A CN 112348856A CN 201910735108 A CN201910735108 A CN 201910735108A CN 112348856 A CN112348856 A CN 112348856A
- Authority
- CN
- China
- Prior art keywords
- tracking
- monitoring
- target
- camera
- map
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 22
- 238000012544 monitoring process Methods 0.000 claims abstract description 116
- 238000000034 method Methods 0.000 claims abstract description 23
- 239000002245 particle Substances 0.000 claims abstract description 16
- 238000013507 mapping Methods 0.000 claims abstract description 10
- 230000033001 locomotion Effects 0.000 claims description 31
- 238000004891 communication Methods 0.000 claims description 19
- 230000000007 visual effect Effects 0.000 claims description 13
- 230000003993 interaction Effects 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 6
- 238000013178 mathematical model Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000000053 physical method Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/292—Multi-camera tracking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
- G06T7/248—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/277—Analysis of motion involving stochastic approaches, e.g. using Kalman filters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30232—Surveillance
Abstract
The invention discloses a multi-camera collaborative analysis method and a multi-camera collaborative analysis system based on a network system topological structure, which are used for acquiring position information of each camera and a monitoring subnet, and establishing a monitoring subnet and global scene map mapping; performing mathematical abstract modeling on the monitoring subnet and the global scene map mapping to obtain a corresponding topological node connection diagram; initializing tracking information, acquiring target information data, and determining the start of tracking; carrying out single target tracking by adopting a particle filter algorithm based on scale invariant features; adopting a layered cascade tracking strategy, and adopting a joint data association algorithm to cooperatively control each monitoring camera in the topological subnet; searching in the global topological map by adopting a space-time association algorithm; and judging the situations of target tracking termination, target tracking loss and the like, and stopping tracking. The method and the system can track the suspicious target across scenes, and the system adopts a distributed layered search method, thereby reducing the tracking cost and realizing the real-time tracking of the suspicious target.
Description
Technical Field
The invention relates to the technical field of visual monitoring, in particular to a camera auxiliary tracking method and system.
Background
With the rapid development of network technologies, video security monitoring technologies have been popularized and developed in a large scale, and video monitoring is deployed at more and more key parts of public activity areas. The monitoring cameras operate independently in a distributed mode, public safety of a city is guaranteed, and auxiliary monitoring information and event tracing evidence are provided. Therefore, the abnormal information of the video image is analyzed, the suspicious object is tracked in real time, the intelligence of network monitoring is improved, the image processing field is in the technical frontier, and the wide market demand is possessed.
In the current practical application range, most target intelligent tracking methods only aim at a single camera with a fixed position, the application of such a scene has great limitation, when an abnormal target is found, the gaze cannot be tracked in time, the target cannot be positioned at a favorable position in video monitoring, and the functions can be realized by a dynamic camera because the observation position and the observation angle of the dynamic camera are variable. Meanwhile, the situation of a single camera is easily affected by the visual field range and the shielding of a shielding object to generate identification errors, so that the visual field range can be enlarged through cooperation of multiple cameras, and the problems of shielding and the like can be solved from different visual angles and directions. Therefore, the application of multiple cameras will bring great help to the future dynamic target tracking system. For continuous tracking research of moving targets of multiple cameras in a large scene, the problem of cooperative tracking of multiple cameras or multiple intelligent devices is involved, which is also a key theory and technical problem to be solved urgently in a practical application system.
Aiming at the defects of an intelligent monitoring and tracking application system, the complexity of the background environment and the uncertainty of the target motion in the real environment, the technical difficulty of multi-camera relay tracking of moving targets across scenes and regions is as follows:
(1) the complexity and stability of the background, namely the scene in which the moving target is located, directly influence the target tracking effect. The interference factors in the background mainly include the change of the light brightness, the background change, the similar interference and the like.
(2) The target shielding problem is that shielding is a common problem and a difficult problem in target tracking, and tracking becomes unstable due to the loss of target information in the shielding process.
(3) The problem of multi-camera cooperative control is that the determination of which camera is used at each moment is needed, namely the problem of data association and information fusion among the multiple cameras.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a multi-camera collaborative analysis method and system based on a network system topological structure, which can accurately track suspicious targets in a fixed scene; the time complexity and the space complexity of target association are reduced; the reliability of cross-camera and cross-scene target tracking can be effectively ensured.
In order to solve the technical problem, the invention provides a multi-camera collaborative analysis method based on a network system topological structure, which comprises the following steps:
s1, constructing a layered real-time monitoring network hardware architecture, acquiring data information of each monitoring camera and each monitoring subnet, and establishing a monitoring subnet and global scene map mapping;
s2: performing mathematical abstract modeling on the monitoring subnet and the global scene map mapping to obtain a corresponding topological node connection topological graph to construct a monitoring subnet topological graph and a global scene topological graph;
s3: firstly, initializing video image information, acquiring a two-dimensional image, a position coordinate and a calibration time or connected data information of a tracking target, and determining the start of tracking;
s4: when the monitoring sub-network receives a tracking instruction, analyzing and processing a monitoring picture of a node where a tracking target is located, and performing single-target tracking by adopting a particle filtering algorithm based on scale invariant features in a single-node visual field range of a single monitoring camera to obtain associated real-time information to update a monitoring center to prepare for tracking handover;
s5: the process of tracking handover adopts a layered search, a cascade tracking and network cooperation method, monitoring subnets adopt a joint data association algorithm to cooperatively control each monitoring camera relay tracking target, and search in the global topological map by adopting a space-time association algorithm according to the tracking result of the topological subnets and depending on the interconnection information of each subnet of the global map topological map;
s6: and judging that the target tracking meets the condition of termination or loss, exiting the target follow-up process and stopping tracking.
Further, the hierarchical real-time monitoring network hardware architecture in step S1 includes: the high-rise monitoring center and the monitoring subnets which are connected with the high-rise monitoring center through the Ethernet and are positioned at the bottom layer establish global scene map mapping through the determined space geometric position of the camera and the space position information of the monitoring subnets.
Further, in step S2, the monitoring subnet topology abstracts the camera in the scene as a topology node, the global scene topology abstracts the position of the monitoring subnet in the map as a topology node, and the spatial direction of the adjacent node of each node relative to the local node is taken as the connection line direction to form a topology connection graph.
Further, in step S3, the data information is obtained by a physical method, and when finding a suspicious target, the finder sends the data information to the server monitoring the subnet through a short message or a network to perform distributed independent tracking.
Further, in step S4, determining the position of the target in the view field by using a scale invariant based image feature matching method, and then quickly tracking the moving single target by using particle filtering; and when the target is lost in the vision field of the node camera, judging the possible direction of the target movement by combining the position information of the camera, the visual angle information and the target movement track, and preparing for target handover.
Further, in step S5, the topological subnetworks of the monitoring subnetworks establish the handover probability of the adjacent nodes by using target motion trajectory distribution and using a joint probability data association algorithm, and then obtain the maximum weight matching to obtain the overall optimum, thereby implementing cross-scene target association and tracking handover;
further, the global topological map issues a matching search control command to each associated subnet, and circularly matches the real-time monitoring pictures of each monitoring node until a target is found or a space-time constraint condition is exceeded.
The other technical scheme of the invention is as follows: a multi-camera collaborative analysis system based on a network system topological structure comprises a high-level monitoring center and a plurality of monitoring sub-network two-layer structures; and the high-level monitoring center is connected with each monitoring subnet through Ethernet.
Further, the high-rise monitoring center includes: the system comprises a human-computer interaction module, a global topological map, a data association module, a target positioning module and a communication module; the system comprises a human-computer interaction module, a global topological map, a data association module, a target positioning module and a communication module, wherein the human-computer interaction module exchanges information with the system through a user interface, the global topological map abstracts a preset geographic information map into a mathematical topological model, the data association module provides a space-time association algorithm to find the most possible motion path of a target among monitoring subnets, the target positioning module traverses and searches the target according to the data association probability, and the communication module is connected with a high-level monitoring center and the monitoring subnets through Ethernet.
Furthermore, the monitoring subnet comprises a monocular tracking module, a tracking handover module, a local topological map and a communication module; the monocular tracking module adopts a particle filter tracking algorithm based on scale invariant features to perform fixed scene tracking on a target, the tracking handover module adopts a joint data association algorithm to perform probability estimation on motion information of the target, the local topological map provides a mathematical model for the tracking handover module, and the communication module performs information transmission with the high-rise monitoring center.
The invention has the technical effects that: 1. the invention relates to a multi-camera cooperation method and a multi-camera cooperation system based on a network system topological structure, which adopt a cascade multi-camera relay tracking system and are composed of a high-rise monitoring center and a plurality of monitoring subnets. The high-level monitoring center is connected with each monitoring subnet through the Ethernet, and has the functions of cooperative control, communication allocation, tracing inquiry and the like. 2. The monitoring subnet collects monitoring pictures in real time through multiple channels, provides 24-hour lossless high-quality signals, and has the functions of monocular tracking, cross-scene tracking and the like. 3. The sub-network and the global monitoring network are monitored, so that the target handover precision is effectively improved; the monocular tracking algorithm is carried out by adopting a particle filtering algorithm based on scale invariant features, and suspicious targets can be accurately tracked in a fixed scene. 4. The monitoring nodes are abstracted into a topological network, so that the time complexity and the space complexity of target association are reduced; by adopting algorithms such as joint probability data association, space-time association and the like, the reliability of cross-camera and cross-scene target tracking can be effectively ensured.
Drawings
FIG. 1 is a flow chart of a multi-camera collaborative analysis method based on a network system topology;
FIG. 2 is a micro control flow diagram of a multi-camera collaborative analysis method based on a network system topology;
FIG. 3 is a block diagram of a multi-camera collaborative analysis system based on a network system topology;
fig. 4 is a monitoring subnet and global scene directed topology connection graph.
Detailed Description
The present invention is further described with reference to the following drawings and specific examples so that those skilled in the art can better understand the present invention and can practice the present invention, but the examples are not intended to limit the present invention.
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Fig. 1-2 are flowcharts of a multi-camera collaborative analysis method based on a network system topology according to the present invention.
1. And constructing a layered real-time monitoring network hardware system, acquiring information such as position information, monitoring range, view angle and the like of each camera and each monitoring subnet, and establishing a monitoring subnet and global scene map mapping by utilizing a geographic information map. And a distributed hardware architecture with layered parallel and independent operation is adopted.
The monitoring subnet is positioned at the bottom position of the whole system, a storage server is arranged for locally storing multi-channel video acquisition information, a 24-hour lossless high-quality monitoring signal is provided, and regular backup is carried out by means of burning and the like. The high-rise monitoring center is the core of the whole intelligent video tracking system, is linked with each monitoring subnet through the Ethernet and has the functions of cooperative control, communication allocation, tracing inquiry and the like. For the selection of the camera for each node of the monitoring subnet, a static fixed camera which is unified and standardized in the industry is generally adopted to determine the space geometric position of the camera, so that stable map mapping of the monitoring subnet is obtained. And the spatial position information of the monitoring subnet is acquired by a geographic information system, such as an application of a Baidu map, and a global scene map is formed according to the acquired spatial position information.
2. In order to utilize the spatial position, the motion characteristics, and the like of the tracking target, the spatial direction of the adjacent node of each node relative to the local node is taken as the direction of the connecting line, and then the topology model becomes a directed topology connecting graph, as shown in fig. 4. The specific process comprises the following steps: and constructing a monitoring subnet topological graph and a global scene topological graph. For the monitoring sub-network, a camera in a scene is abstracted into topological graph nodes, and the spatial direction of the adjacent nodes of each node relative to the node is taken as the direction of a connecting line (without considering the condition of overlapping visual fields) by utilizing the spatial position, the motion characteristic and the like of a tracking target, so that the directed topological connecting graph of the monitoring sub-network is formed. For a global scene, abstracting the position of a monitoring subnet in a map into topological graph nodes, and forming a global scene directed topological connection graph by taking the spatial direction of adjacent nodes of each node relative to the node as a connection line direction by utilizing the spatial position, motion characteristics and the like of a tracking target.
3. After the hardware system is constructed and the physical model is abstracted to generate the mathematical model, the cross-scene target tracking problem is simplified into a single-target tracking problem and a multi-camera cooperative data association problem in a fixed scene. The user submits the picture of the suspicious target to the monitoring center through the mobile phone or the network, for example, the suspicious target Tl is found in the physical space of the node P3 in the monitoring subnet Z2, the system stores the image information and the positioning information of the Tl, updates the motion information (such as uniform speed, motion direction, motion time and the like) in the tracking process, and then transmits the target information and issues a tracking command to the server of the monitoring subnet Z2 through the communication module.
4. And after receiving the tracking command, the monitoring sub-network analyzes and processes the monitoring picture of the node where the target is located, and rapidly tracks the moving single target by adopting a particle filter algorithm based on scale invariant features. The core ideas of the algorithm are random sampling and importance resampling. Under the condition that the position of a target is unknown, particles are scattered in a scene randomly, the importance of each particle is calculated according to the feature similarity, then the weight of each particle is redistributed according to the probability, the maximum probability position of the target motion is found out, and the target position is marked; and finally, repeating the steps for each frame of the monitoring video. After the monitoring subnet Z2 tracks the target by adopting the particle filter algorithm, the information such as the real-time coordinate and the movement direction of the target can be obtained, so that the target information of the monitoring center is updated, and meanwhile, guarantee and support are provided for tracking handover.
Referring to fig. 2, when the monitoring subnet server performs tracking, after the tracking starts, the tracking object is manually selected, the position of the topology subnet where the tracking object is located is determined, the monitoring subnet node where the tracking object is located is determined, the position of the target in the visual field is determined based on the scale-invariant image feature matching method, and then the moving single-target camera tracking algorithm is rapidly tracked by adopting particle filtering. When the target is lost in the view field of the node camera, the possible direction of the target movement is judged by combining the position information of the camera, the visual angle information and the target movement track, and when the target leaves the view field of the current node camera or the monitoring range of the current monitoring subnet, the preparation for target handover is made.
When a suspicious target is found in the physical world, a pedestrian or a person found by a security guard or the like takes a target image and uploads the target image to a monitoring center through a short message or a network. The system automatically records the two-dimensional image information, the position coordinate information and the calibration time information of the target, and sends the related information to the server of the monitoring subnet, so that the server can perform distributed independent tracking, thereby reducing the tracking complexity and improving the tracking efficiency.
5. In the process of tracking handover, a joint data association algorithm is adopted in the topological subnet to cooperatively control each monitoring camera; and searching in the global topological map by adopting a space-time correlation algorithm. And setting the position of the adjacent node corresponding to the current node as a preset position by using a local subnet topological map, and then performing handover judgment on the relay camera required to be performed. The basic idea of the multi-camera handover association method is a combined data association algorithm, the core of the algorithm is to count the probability of a target going to adjacent nodes according to a target motion rule, and sequentially search the nodes according to the probability until the target is found to complete a handover task. Assuming that the motion track of the target Tl at the node P3 is obtained through particle filter tracking and the motion rule of the target Tl moves from south to north, judging that the target possibly goes to Pl and P4 according to a data association algorithm; judging that the possibility of the target going to Pl is 0.7 and the possibility of going to P4 is 0.3 through Tl conventional motion trajectories P8-P6-P3, and finally determining to find the target going to P1 through sequential matching retrieval of Pl and P4 pictures. The space-time association algorithm of the global topological map is similar to the above algorithm, except that the time range of the target reaching each monitoring subnet vision is estimated by the average speed of the target, so that each monitoring subnet is cooperatively scheduled.
And a tracking strategy of layered searching and cascading tracking is adopted, so that the time complexity and the space complexity of data association and video tracking are effectively reduced. Determining a subnet area where a target exists through manual calibration; and then utilize the underlying monitoring server. Adopting a probability association algorithm to relay and track suspicious targets among all cameras of the monitoring sub-network; when the target leaves the subnet area, returning to the global map retrieval, and sequentially matching adjacent areas according to a space-time association method; and once the target is determined to be present in a certain area, the monitoring subnet is entered again for relay tracking. Wherein the topology subnet tracking strategy is as follows: establishing the handover probability of adjacent nodes by using target motion trajectory distribution and adopting a joint probability data association algorithm; and then, solving the maximum weight matching to obtain the overall optimum, and realizing cross-scene target association and tracking handover. Global topological map tracking strategy: obtaining the speed and direction information of the target motion through the tracking result of the topological subnet; obtaining the associated probability relation of each subnet by depending on the interconnection information of each subnet of the global map topology map; and issuing a matching search control command to each associated subnet according to the time-space association probability, and circularly matching the real-time monitoring pictures of each monitoring node until a target is found or a time-space constraint condition is exceeded.
The monitoring subnet Z2 utilizes a bottom monitoring server to track the target Tl in a relay manner to obtain a motion track P8-P6-P3-P1; when the target leaves the Pl (Pl is a topological subnet edge node) visual field, the monitoring subnet Z2 judges that the target possibly leaves 0.8 of the visual field of the subnet, and then returns to the global topological map for retrieval; obtaining time constraint of target motion according to a space-time correlation method, matching all node pictures of the monitoring subnets Zl and Z3 in the range according to probability, and finally finding out the monitoring subnet Zl where the target arrives; and once the target is determined, immediately entering a monitoring subnet ZL to repeat the steps for relay tracking.
6. If the global topological map retrieval matching fails in the time constraint range, the time constraint range is expanded to carry out global retrieval in a larger area monitoring subnet Z3 and a TA area. And if the suspicious target is not found, judging that the task fails, quitting the target tracking process, and displaying word prompts such as tracking failure and the like on an interactive interface of the monitoring center.
Secondly, as shown in fig. 3, the invention relates to a multi-camera collaborative analysis system, a cascade multi-camera relay tracking system, which is composed of a high-level control center and a plurality of monitoring subnets. The high-rise monitoring center is connected with each monitoring subnet through the Ethernet, and has the functions of cooperative control, communication allocation, tracing inquiry and the like; the monitoring subnet collects monitoring pictures in real time through multiple channels, provides 24-hour lossless high-quality signals, and has the functions of monocular tracking, cross-scene tracking and the like.
The high-level monitoring center comprises a human-computer interaction module, a global topological map, a data association module, a target positioning module and a communication module. The system comprises a human-computer interaction module, a global topological map, a data association module, a target positioning module and a communication module, wherein the human-computer interaction module exchanges information with the system through a user interface, the global topological map abstracts a preset geographic information map into a mathematical topological model, the data association module provides a space-time association algorithm to find the most possible motion path of a target among monitoring subnets, the target positioning module traverses and searches the target according to the data association probability, and the communication module is connected with a high-level monitoring center and the monitoring subnets through Ethernet.
The monitoring subnet comprises a monocular tracking module, a tracking handover module, a local topological map and a communication module; the monocular tracking module adopts a particle filter tracking algorithm based on scale invariant features to perform fixed scene tracking on a target, the tracking handover module adopts a joint data association algorithm to perform probability estimation on motion information of the target, the local topological map provides a mathematical model for the tracking handover module, and the communication module performs information transmission with the high-rise monitoring center.
The above-mentioned embodiments are merely preferred embodiments for fully illustrating the present invention, and the scope of the present invention is not limited thereto. The equivalent substitution or change made by the technical personnel in the technical field on the basis of the invention is all within the protection scope of the invention. The protection scope of the invention is subject to the claims.
Claims (10)
1. A multi-camera collaborative analysis method based on a network system topological structure is characterized by comprising the following steps:
s1, constructing a layered real-time monitoring network hardware architecture, acquiring data information of each monitoring camera and each monitoring subnet, and establishing a monitoring subnet and global scene map mapping;
s2: performing mathematical abstract modeling on the monitoring subnet and the global scene map mapping to obtain a corresponding topological node connection topological graph to construct a monitoring subnet topological graph and a global scene topological graph;
s3: firstly, initializing video image information, acquiring a two-dimensional image, a position coordinate and a calibration time or connected data information of a tracking target, and determining the start of tracking;
s4: when the monitoring sub-network receives a tracking instruction, analyzing and processing a monitoring picture of a node where a tracking target is located, and performing single-target tracking by adopting a particle filtering algorithm based on scale invariant features in a single-node visual field range of a single monitoring camera to obtain associated real-time information to update a monitoring center to prepare for tracking handover;
s5: the process of tracking handover adopts a layered search, a cascade tracking and network cooperation method, monitoring subnets adopt a joint data association algorithm to cooperatively control each monitoring camera relay tracking target, and search in the global topological map by adopting a space-time association algorithm according to the tracking result of the topological subnets and depending on the interconnection information of each subnet of the global map topological map;
s6: and judging that the target tracking meets the condition of termination or loss, exiting the target follow-up process and stopping tracking.
2. The multi-camera assisted collaborative analysis method according to claim 1, wherein the hierarchical real-time monitoring network hardware architecture in the step S1 includes: the high-rise monitoring center and the monitoring subnets which are connected with the high-rise monitoring center through the Ethernet and are positioned at the bottom layer establish global scene map mapping through the determined space geometric position of the camera and the space position information of the monitoring subnets.
3. The multi-camera assisted collaborative analysis method according to claim 1, wherein in step S2, the topology map of the monitoring subnets abstracts the cameras in the scene into topology map nodes, the topology map of the global scene abstracts the positions of the monitoring subnets in the map into topology map nodes, and the spatial direction of the adjacent nodes of each node relative to the local node is taken as a connection line direction to form a topology connection map.
4. The multi-camera assisted collaborative analysis method according to claim 1, wherein in step S3, the data information is obtained by a physical method, and when a suspicious object is found, a finder sends the data information to a server of the monitoring subnet for distributed independent tracking through a short message or a network.
5. The multi-camera assisted collaborative analysis method according to claim 1, wherein in step S4, a scale invariant based image feature matching method is used to determine the position of the target in the view, and then particle filtering is used to perform fast tracking on the moving single target; and when the target is lost in the vision field of the node camera, judging the possible direction of the target movement by combining the position information of the camera, the visual angle information and the target movement track, and preparing for target handover.
6. The multi-camera assisted collaborative analysis method according to claim 1, wherein in step S5, the topology subnetworks of the monitoring subnetworks establish handover probabilities of adjacent nodes using a joint probability data-based association algorithm by using target motion trajectory distribution, and then obtain a maximum weight match to obtain an overall optimum, thereby achieving cross-scene target association and tracking handover.
7. The multi-camera aided collaborative analysis method according to any one of claims 1 or 6, wherein the global topology map issues matching search control commands to each associated subnet, and matches the real-time monitoring pictures of each monitoring node in a loop until a target is found or a spatio-temporal constraint is exceeded.
8. A multi-camera collaborative analysis system based on a network system topological structure is characterized by comprising a high-level monitoring center and a plurality of monitoring sub-network two-layer structures; and the high-level monitoring center is connected with each monitoring subnet through Ethernet.
9. The multi-camera collaborative analysis system of claim 8, wherein the high-level monitoring center comprises: the system comprises a human-computer interaction module, a global topological map, a data association module, a target positioning module and a communication module; the system comprises a human-computer interaction module, a global topological map, a data association module, a target positioning module and a communication module, wherein the human-computer interaction module exchanges information with the system through a user interface, the global topological map abstracts a preset geographic information map into a mathematical topological model, the data association module provides a space-time association algorithm to find the most possible motion path of a target among monitoring subnets, the target positioning module traverses and searches the target according to the data association probability, and the communication module is connected with a high-level monitoring center and the monitoring subnets through Ethernet.
10. The multi-camera collaborative analysis system of claim 8, wherein the monitoring subnet comprises a monocular tracking module, a tracking handover module, a local topology map, and a communication module; the monocular tracking module adopts a particle filter tracking algorithm based on scale invariant features to perform fixed scene tracking on a target, the tracking handover module adopts a joint data association algorithm to perform probability estimation on motion information of the target, the local topological map provides a mathematical model for the tracking handover module, and the communication module performs information transmission with the high-rise monitoring center.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910735108.8A CN112348856A (en) | 2019-08-09 | 2019-08-09 | Multi-camera collaborative analysis method and system based on network system topological structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910735108.8A CN112348856A (en) | 2019-08-09 | 2019-08-09 | Multi-camera collaborative analysis method and system based on network system topological structure |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112348856A true CN112348856A (en) | 2021-02-09 |
Family
ID=74367831
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910735108.8A Pending CN112348856A (en) | 2019-08-09 | 2019-08-09 | Multi-camera collaborative analysis method and system based on network system topological structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112348856A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113538584A (en) * | 2021-09-16 | 2021-10-22 | 北京创米智汇物联科技有限公司 | Camera auto-negotiation monitoring processing method and system and camera |
| CN115578694A (en) * | 2022-11-18 | 2023-01-06 | 合肥英特灵达信息技术有限公司 | Video analysis computing power scheduling method, system, electronic equipment and storage medium |
| CN115617532A (en) * | 2022-11-22 | 2023-01-17 | 浙江莲荷科技有限公司 | Target tracking processing method, system and related device |
| CN115731287A (en) * | 2022-09-07 | 2023-03-03 | 滁州学院 | Moving target retrieval method based on set and topological space |
| CN116866534A (en) * | 2023-09-05 | 2023-10-10 | 南京隆精微电子技术有限公司 | Processing method and device for digital video monitoring system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090324010A1 (en) * | 2008-06-26 | 2009-12-31 | Billy Hou | Neural network-controlled automatic tracking and recognizing system and method |
| CN102663743A (en) * | 2012-03-23 | 2012-09-12 | 西安电子科技大学 | Multi-camera cooperative character tracking method in complex scene |
| WO2012151777A1 (en) * | 2011-05-09 | 2012-11-15 | 上海芯启电子科技有限公司 | Multi-target tracking close-up shooting video monitoring system |
| CN104038729A (en) * | 2014-05-05 | 2014-09-10 | 重庆大学 | Cascade-type multi-camera relay tracing method and system |
| CN106096577A (en) * | 2016-06-24 | 2016-11-09 | 安徽工业大学 | Target tracking system in a kind of photographic head distribution map and method for tracing |
| CN106846374A (en) * | 2016-12-21 | 2017-06-13 | 大连海事大学 | The track calculating method of vehicle under multi-cam scene |
-
2019
- 2019-08-09 CN CN201910735108.8A patent/CN112348856A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090324010A1 (en) * | 2008-06-26 | 2009-12-31 | Billy Hou | Neural network-controlled automatic tracking and recognizing system and method |
| WO2012151777A1 (en) * | 2011-05-09 | 2012-11-15 | 上海芯启电子科技有限公司 | Multi-target tracking close-up shooting video monitoring system |
| CN102663743A (en) * | 2012-03-23 | 2012-09-12 | 西安电子科技大学 | Multi-camera cooperative character tracking method in complex scene |
| CN104038729A (en) * | 2014-05-05 | 2014-09-10 | 重庆大学 | Cascade-type multi-camera relay tracing method and system |
| CN106096577A (en) * | 2016-06-24 | 2016-11-09 | 安徽工业大学 | Target tracking system in a kind of photographic head distribution map and method for tracing |
| CN106846374A (en) * | 2016-12-21 | 2017-06-13 | 大连海事大学 | The track calculating method of vehicle under multi-cam scene |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113538584A (en) * | 2021-09-16 | 2021-10-22 | 北京创米智汇物联科技有限公司 | Camera auto-negotiation monitoring processing method and system and camera |
| CN113538584B (en) * | 2021-09-16 | 2021-11-26 | 北京创米智汇物联科技有限公司 | Camera auto-negotiation monitoring processing method and system and camera |
| CN115731287A (en) * | 2022-09-07 | 2023-03-03 | 滁州学院 | Moving target retrieval method based on set and topological space |
| CN115578694A (en) * | 2022-11-18 | 2023-01-06 | 合肥英特灵达信息技术有限公司 | Video analysis computing power scheduling method, system, electronic equipment and storage medium |
| CN115617532A (en) * | 2022-11-22 | 2023-01-17 | 浙江莲荷科技有限公司 | Target tracking processing method, system and related device |
| CN115617532B (en) * | 2022-11-22 | 2023-03-31 | 浙江莲荷科技有限公司 | Target tracking processing method, system and related device |
| CN116866534A (en) * | 2023-09-05 | 2023-10-10 | 南京隆精微电子技术有限公司 | Processing method and device for digital video monitoring system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112348856A (en) | Multi-camera collaborative analysis method and system based on network system topological structure | |
| CN104038729A (en) | Cascade-type multi-camera relay tracing method and system | |
| Collins et al. | A system for video surveillance and monitoring | |
| CN103004188B (en) | Equipment, system and method | |
| Valera et al. | Intelligent distributed surveillance systems: a review | |
| CN106791613B (en) | A kind of intelligent monitor system combined based on 3DGIS and video | |
| Micheloni et al. | A network of co-operative cameras for visual surveillance | |
| Castanedo et al. | Data fusion to improve trajectory tracking in a Cooperative Surveillance Multi-Agent Architecture | |
| Patricio et al. | Multi-agent framework in visual sensor networks | |
| US20120249802A1 (en) | Distributed target tracking using self localizing smart camera networks | |
| US20220137636A1 (en) | Systems and Methods for Simultaneous Localization and Mapping Using Asynchronous Multi-View Cameras | |
| CN110533700A (en) | Method for tracing object and device, storage medium and electronic device | |
| CN112053391B (en) | Monitoring and early warning method and system based on dynamic three-dimensional model and storage medium | |
| CN106331602A (en) | Home monitoring system based on infrared thermal imaging technology | |
| CN111526324B (en) | Monitoring system and method | |
| CN113935358A (en) | Pedestrian tracking method, equipment and storage medium | |
| CN114115289A (en) | Autonomous unmanned cluster reconnaissance system | |
| Casao et al. | Distributed multi-target tracking in camera networks | |
| Aguilar-Ponce et al. | A network of sensor-based framework for automated visual surveillance | |
| Hoffmann et al. | Spatial partitioning in self-organizing smart camera systems | |
| Stancil et al. | Active multicamera networks: From rendering to surveillance | |
| Merino et al. | Data fusion in ubiquitous networked robot systems for urban services | |
| Valera et al. | A review of the state-of-the-art in distributed surveillance systems | |
| Qureshi et al. | Towards intelligent camera networks: a virtual vision approach | |
| KR101957414B1 (en) | Apartment monitoring system based on CCTV view correction design, and method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210615 Address after: 224000 Building 1, intelligent terminal industry pioneer park, high tech Zone, Yandu District, Yancheng City, Jiangsu Province (d) Applicant after: Yancheng Jiyan Intelligent Technology Co.,Ltd. Address before: 224000 Building 1, intelligent terminal industry pioneer park, high tech Zone, Yandu District, Yancheng City, Jiangsu Province Applicant before: Co., Ltd of intelligent terminal industrial research institute of Yancheng Jinlin University |