WO2020114138A1

WO2020114138A1 - Information associated analysis method and apparatus, and storage medium and electronic device

Info

Publication number: WO2020114138A1
Application number: PCT/CN2019/112925
Authority: WO
Inventors: 刘若鹏; 栾琳; 季春霖
Original assignee: 西安光启未来技术研究院
Priority date: 2018-12-06
Filing date: 2019-10-24
Publication date: 2020-06-11
Also published as: CN111291589A

Abstract

Embodiments of the present invention provide an information associated analysis method and apparatus, and a storage medium and an electronic device. The method comprises: extracting related information of a target object from a surveillance video, wherein the related information of the target object comprises body feature information, behavior feature information, and equipment information; and performing multi-dimensional associated analysis on the extracted related information of the target object to obtain a character portrait and a behavior trace of the target object. The problems in the prior art that manual viewing of surveillance videos takes time and effort, and the accuracy caused by single-dimensional analysis is low.

Description

Information correlation analysis method and device, storage medium, and electronic device

Technical field

The invention relates to the technical field of intelligent monitoring, in particular, to an information correlation analysis method and device, a storage medium, and an electronic device.

Background technique

In recent years, with the expansion of the scale of video surveillance systems in various places, the public security business has continued to rely on video. The application of video surveillance image information has become an important means for public security organs to prevent public security, fight crime, and direct communications.

In the current video investigation work, criminal investigators are "staring" at the player, holding a notebook and pen in hand, watching and recording, even when there are few active targets in the surveillance video at night or in remote areas. It can only be browsed “completely”, and even the omission of “one second” cannot appear. Browsing video recordings for a long time is very likely to cause visual fatigue of criminal investigators and affect the quality of video browsing work, and even cause visual damage to investigators. It is completely time-consuming, laborious, and inefficient for the investigators to manually browse and find suspected targets, and this traditional method of watching surveillance video usually analyzes and judges from a single information dimension, such as a single analysis of the target face, or Posture, or gait, or head shape, etc., the accuracy of this kind of judgment from a single dimension is often not high.

technical problem

There is no reasonable solution to the problem in the prior art that it is time-consuming and laborious to manually browse and monitor video, and the research and judgment from a single dimension results in low accuracy.

Technical solution

Embodiments of the present invention provide an information correlation analysis method and device, a storage medium, and an electronic device, to at least solve the problem of time-consuming and laborious manual browsing of surveillance video in the related art, and research and judgment from a single dimension results in low accuracy.

According to an embodiment of the present invention, an information correlation analysis method is provided, which includes: extracting relevant information of a target object from a surveillance video, wherein the relevant information of the target object includes: body characteristic information, behavior characteristic information, Equipment information; multi-dimensional correlation analysis is performed on the extracted relevant information of the target object to obtain a character portrait and behavior track of the target object.

Preferably, before extracting the relevant information of the target object from the monitoring video, the method further includes: collecting a monitoring video containing the target object, wherein the monitoring video includes at least one of the following image elements: the target The subject himself, the vehicle associated with the target object, the iconic item associated with the target object.

Preferably, extracting the relevant information of the target object from the surveillance video includes: extracting the relevant information of the target object from the surveillance video using a portrait multi-dimensional model, wherein the portrait multi-dimensional model is a model built using feature vectors of multiple dimensions of the human body , The portrait multi-dimensional model uses a convolutional neural network to extract features from the image.

Preferably, the relevant information of the target object is extracted from the surveillance video, including one of the following: the body characteristic information of the target object is extracted from the surveillance video, wherein the body characteristic information includes at least one of the following: gender, height, hair accessories , Clothing, carry-on items, walking patterns, facial features, head shape, facial decorations; extract the behavioral feature information of the target object from the surveillance video, wherein the behavioral feature information includes at least one of the following: cross-border behavior, activity Area, wandering area, aggregation behavior; extract target equipment information from surveillance video, where the equipment information includes at least one of the following: vehicle type, license plate, vehicle color, model, brand, car sticker, car Accessories information, the handheld mobile terminal of the target object.

Preferably, before performing multi-dimensional correlation analysis on the extracted relevant information of the target object, the method includes: acquiring position information of the handheld mobile terminal of the target object according to signal positioning.

Preferably, performing multi-dimensional correlation analysis on the extracted relevant information of the target object includes: associating at least two types of information in the following relevant information of the target object: physical characteristic information, behavior characteristic information, equipment information, The location information of the handheld mobile terminal; comprehensive analysis of the relevant information after the association in time, space, region, trajectory and social relationship.

Preferably, after performing a comprehensive analysis of the relevant information after correlation in time, space, region, and trajectory, the method further includes: predicting the future behavior of the target object according to the result of the comprehensive analysis; and/or The result of the comprehensive analysis is to track the position of the target object.

Preferably, the method further comprises: displaying at least one of the following information on the display screen: character portrait generation information, space-time retrospective display information, path tracking information, behavior habit model, security situation map, and real-time portrait monitoring information.

According to another aspect of the embodiments of the present invention, there is also provided an information correlation analysis device, including: an extraction module for extracting related information of a target object from a surveillance video using a portrait multi-dimensional model, wherein the target object The related information includes: physical characteristic information, behavior characteristic information, and equipment information; an analysis module is used to perform multi-dimensional correlation analysis on the extracted related information of the target object to obtain a character portrait and a behavior track of the target object.

According to another aspect of the embodiments of the present invention, there is also provided a storage medium in which a computer program is stored, wherein the computer program is configured to execute the steps in any one of the above method embodiments during runtime .

According to another embodiment of the present invention, there is also provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute The steps in any of the above method embodiments.

Beneficial effect

Through the embodiments of the present invention, the multi-dimensional related information of the target object is extracted from the surveillance video, including body characteristic information, behavior characteristic information, equipment information, etc.; the multi-dimensional related information of the extracted target object is multi-dimensionally correlated Analyze to get the portrait and behavior trajectory of the target object. Solve the problem of manually browsing and monitoring video in the related technology, which is time-consuming and laborious, and the research and judgment from a single dimension leads to low accuracy. Through the fusion processing and research and judgment of each feature vector of the tracking object, the static data is built with "people" as the core The dynamic data formed with various intelligent perception systems are interlinked, integrated, correlated, comprehensively analyzed and judged to form an integrated combat system capable of intelligent discovery, intelligent research and judgment, intelligent triggering, and intelligent tracking. With this as the core support, it can truly form a transition from passive disposal to active discovery, prediction, prevention and early warning.

BRIEF DESCRIPTION

The drawings described herein are used to provide a further understanding of the present invention and form a part of the present application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an undue limitation on the present invention. In the drawings:

1 is a block diagram of a hardware structure of a mobile terminal of an information association analysis method according to an embodiment of the present invention;

2 is a flowchart of a method for analyzing association of information according to an embodiment of the present invention;

3 is a schematic diagram of characteristics of acquiring relevant information of a target object according to an embodiment of the present invention;

4 is a schematic diagram of multi-dimensional model construction and information research and judgment according to an embodiment of the present invention;

5 is a structural block diagram of an information association analysis device according to an embodiment of the present invention.

Embodiments of the invention

Hereinafter, the present invention will be described in detail with reference to the drawings and in conjunction with the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other if there is no conflict.

It should be noted that the terms “first” and “second” in the description and claims of the present invention and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence.

实施例Examples 11

The method embodiment provided in Embodiment 1 of the present application may be executed in a mobile terminal, a computer terminal, or a similar computing device. Taking an example of running on a mobile terminal, FIG. 1 is a block diagram of a hardware structure of a mobile terminal of an information association analysis method according to an embodiment of the present invention. As shown in FIG. 1, the mobile terminal 10 may include one or more (only one is shown in FIG. 1) processor 102 (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA, etc. ) And a memory 104 for storing data. Optionally, the above mobile terminal may further include a transmission device 106 for communication functions and an input/output device 108. A person of ordinary skill in the art may understand that the structure shown in FIG. 1 is merely an illustration, which does not limit the structure of the mobile terminal described above. For example, the mobile terminal 10 may also include more or fewer components than those shown in FIG. 1, or have a different configuration from that shown in FIG.

The memory 104 may be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for acquiring data information in the embodiment of the present invention, and the processor 102 runs the computer program stored in the memory 104, thereby Implementation of various functional applications and data processing, that is, to achieve the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memories remotely provided with respect to the processor 102, and these remote memories may be connected to the mobile terminal 10 through a network. Examples of the above network include but are not limited to the Internet, intranet, local area network, mobile communication network, and combinations thereof.

The transmission device 106 is used to receive or send data via a network. The specific example of the network described above may include a wireless network provided by a communication provider of the mobile terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller (abbreviated as NIC), which can be connected to other network devices through the base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (Radio Frequency, RF for short) module, which is used to communicate with the Internet in a wireless manner.

The embodiment of the invention provides a method for acquiring food material information. FIG. 2 is a flowchart of the information association analysis method according to an embodiment of the present invention. As shown in FIG. 2, the method includes:

In step S201, relevant information of the target object is extracted from the surveillance video, wherein the relevant information of the target object includes: body characteristic information, behavior characteristic information, equipment information;

Step S203: Multi-dimensional correlation analysis is performed on the extracted relevant information of the target object to obtain a character portrait and behavior track of the target object.

Through the above method, the relevant information of the target object is extracted from the surveillance video, wherein the relevant information of the target object includes: body characteristic information, behavior characteristic information, equipment information; multi-dimensional correlation analysis is performed on the extracted relevant information of the target object, Get the character portrait and behavior track of the target object. It solves the problem of manual browsing of surveillance videos in the related technology that is time-consuming and laborious, and the research and judgment from a single dimension results in low accuracy. Through the fusion processing and research and judgment of each feature vector of the tracking object, the static data is constructed with "people" as the core The dynamic data formed with various intelligent perception systems are interlinked, integrated, correlated, comprehensively analyzed and judged to form an integrated combat system capable of intelligent discovery, intelligent research and judgment, intelligent triggering, and intelligent tracking. With this as the core support, it can truly form a transition from passive disposal to active discovery, prediction, prevention and early warning.

According to a preferred implementation of the embodiment of the present invention, before the above step S201, the method further includes: collecting a surveillance video containing the target object, wherein the surveillance video includes at least one of the following image elements: the target object himself, and The vehicle associated with the target object, the iconic item associated with the target object.

According to a preferred implementation of the embodiment of the present invention, the above step S201 may be implemented by the following steps: extracting the relevant information of the target object from the surveillance video by using a portrait multi-dimensional model, wherein the portrait multi-dimensional model is a feature that uses multiple dimensions of the human body Vector-based model, portrait multi-dimensional model uses convolutional neural network to extract features from images.

According to a preferred implementation of the embodiment of the present invention, the above step S201 may be implemented in at least one of the following ways: extracting the body characteristic information of the target object from the surveillance video, wherein the body characteristic information includes at least one of the following: gender, Height, hair accessories, clothing, carry-on items, walking patterns, facial features, head shape, facial ornaments; extract the behavioral feature information of the target object from the surveillance video, where the behavioral feature information includes at least one of the following: transboundary behavior , Active area, wandering area, gathering behavior; extract the target’s equipment information from the surveillance video, where the equipment information includes at least one of the following: vehicle type, license plate, vehicle color, model, brand, car sticker, car Handheld mobile terminal for accessories information and target objects.

According to a preferred implementation of the embodiment of the present invention, before performing multi-dimensional correlation analysis on the extracted relevant information of the target object, the method further includes: acquiring position information of the handheld mobile terminal of the target object according to signal positioning.

According to a preferred implementation of the embodiment of the present invention, performing multi-dimensional correlation analysis on the extracted relevant information of the target object includes: correlating at least two types of information in the following relevant information of the target object: body characteristic information, Behavioral characteristic information, equipment information, and location information of handheld mobile terminals; comprehensive analysis of the relevant information after association in space-time, region, trajectory, and social relationships.

According to a preferred implementation of the embodiment of the present invention, after a comprehensive analysis of the associated relevant information in time, space, region, and trajectory, the method further includes: predicting the future behavior of the target object based on the result of the comprehensive analysis; and// Or, based on the results of comprehensive analysis, track the location of the target object.

According to a preferred implementation of the embodiment of the present invention, the method further includes: displaying at least one of the following information on the display screen: character portrait generation information, space-time retrospective display information, path tracking information, behavior habit model, security situation Figure, portrait real-time monitoring information.

In order to better understand the technical solutions in the embodiments of the present invention, specific description will be given below with reference to the drawings.

3 is a schematic diagram of characteristics of acquiring relevant information of a target object according to an embodiment of the present invention. As shown in Figure 3, the smart terminal camera collects video information, such as people, vehicles, behaviors, etc., including the precise positioning of the face of the person, facial feature extraction, facial feature comparison, gender, age range, approximate height, hair accessories of the person , A variety of structured description information such as clothing, item carrying, walking form, etc.; various vehicle description information such as vehicle type, license plate, color, model, brand, car stickers, car accessories information; cross-border, area, wandering, aggregation Various behavior description information, through the multi-dimensional vector analysis model, to mine the depth information of various types of data, combined with the existing police information database, to achieve portrait and path tracking. Comprehensively expand the analysis and prediction functions of video data, complete the time and space tracing and behavior prediction (crime location, behavior address, etc.) of suspected, suspicious, and suspected persons for multi-dimensional tracking along the entire time axis.

Through the conversion from single face recognition to multi-dimensional feature recognition, the multi-dimensional feature model is used for comprehensive analysis, research and characterization of the target face, posture, gait, upper and lower body clothing, head shape, whether to wear glasses, etc. From the isolated probe to the full-time domain-connected network probe, combined with the geographical location information of the network, the time-for-precision is achieved to describe the target trajectory. From a single information dimension to the fusion of multi-dimensional information such as people, cars, mobile phones, etc., with "people" as the core object, synchronously discover, identify, generate trajectories for "people", "mobile phones", "cars", etc. to achieve trajectory retrieval, Moreover, the trajectories of "people", "car" and "mobile phone" can be merged to form a comprehensive research and judgment, comprehensive analysis, investigation and screening, and characterize the target relationship.

Synthesizing the above data, while building cross-regional, cross-temporal and spatial recognition capabilities, using advanced cloud computing, artificial intelligence technology, combined with supercomputing and big data processing technology, to achieve key personnel trajectory analysis, key personnel trajectory judgment, key industry data Monitoring, risk monitoring in key industries, multi-dimensional stereoscopic portraits, intelligent serial parallel cases, spatiotemporal trajectory judgment, social relationship analysis, security awareness situation maps, etc., and sorting, archiving, analysis, prediction, and mining various types from complex data The inherent, inevitable causal relationship behind the data, find the hidden laws, and promote these data from quantitative to qualitative changes, in-depth application and comprehensive application of massive data, with people as the core target resume file, to achieve backward and forward track.

4 is a schematic diagram of multi-dimensional model construction and information research and judgment according to an embodiment of the present invention. The multi-dimensional model construction and information research and judgment are divided into four parts: model building, data mining, comprehensive analysis, and fusion research and visualization.

1) Create a model

The establishment of a model is the establishment of a multi-dimensional model of portraits, which comprehensively utilizes multiple dimensional feature vectors such as face, head, clothing, behavior, posture, gait, gender, etc. to establish a multi-dimensional model of portraits, superimposing the time dimension and spatial dimension to improve object recognition Accuracy. To build the model, the convolutional neural network is mainly used to extract the features of the image. The convolutional neural network abstracts the high-dimensional features of the image. The traditional image algorithm uses artificially set features such as Sift or SURF. The artificially set features are often local features, which do not reflect the global features of the image well, and the artificially set feature dimensions are relatively low, which cannot fully reflect the features of the image, and rely heavily on human subjective experience. Using convolutional neural networks to extract high-dimensional features can extract 512 or 128 or other dimensions of features as needed. Able to synthesize information such as human facial information and clothing. Strong expression skills. Convolutional neural networks include convolutional layers, pooling layers, and fully connected layers.

2) Data mining

From massive data, through statistics, online analytical processing, information retrieval, machine learning, expert systems and pattern recognition, to classify, estimate, predict, group or correlate correlation rules, clustering, complex data type mining , To realize the information and value hidden in the search.

3) Comprehensive analysis

The comprehensive analysis is divided into cross-temporal analysis, cross-regional analysis, trajectory blending analysis, cross-time, domain, and trace comprehensive analysis, and social relationship analysis.

Spatial-temporal analysis: Taking "spatial-temporal" as the coordinates, time is used for accuracy. Backtracking the historical spatiotemporal information, obtaining qualitative support, and gradually sampling the existing information, superimposing calculations, can continue to approach precision to achieve the purpose of predicting the future.

Cross-regional analysis: Taking "region" as the coordinate, analyze the concentration of regional crowds, so that the police can dispatch police force distribution in time, control the gathering events, avoid accidents, and take appropriate measures according to the frequency of gathering.

Trajectory fusion analysis: Take the map as the carrier, according to the time and location of the target person, synthesize the map's track route map, find the behavior rules, speculate the target person's behavior purpose, and investigate the clues, so that the police can track and monitor the target person in real time.

Comprehensive analysis across time, domains, and traces: build a full-time and space-time management and control capability with "people" as the core, and continuously analyze the information of portrait data in the multi-dimensional space of time, region, and trajectory, and continue to approach precision. Accurate control depends on the precise situation Analyze, research and judge to realize the identity, traceability and disposal anytime and anywhere.

Social relationship analysis: use the video surveillance platform to intelligently analyze the video content, extract the fragments of the target personnel, and conduct a unified analysis of the historical fragments of the people nearby to build a social network of the target personnel, so that the police can quickly lock Suspects, investigate individuals or gangs to commit crimes.

4) Visualization of fusion research

The fusion research and judgment visualization system is divided into character portrait generation, space-time retrospective display, path tracking, behavior habit model, security situation map, and real-time portrait monitoring.

In the super intelligent tracking system of the present invention, multiple sensing terminals are deployed and integrated in the sensing layer, which has the intelligent sensing ability of different self-learning and self-adaptation according to different scenarios and environments. The front-end sensing layer and the back-end are based on super-computing capabilities Supported artificial intelligence combat engines work together to monitor and track targets, rather than simple camera integration and data overlay. Based on this system, the object of the present invention is around an individual "person". When the "person" enters the monitoring area, it can be recognized and identified, and a spatiotemporal file can be established to realize the precise location of the node at a certain time. Monitor their ongoing behavior; by backtracking forward, you can call the online and offline comprehensive behaviors in the past time period to make a pre-judgment on their next behavior; tracking backward, you can continue to observe Its action trajectory provides timely warning and intervention to risks.

The super intelligent tracking system of the present invention achieves accurate strikes through intelligent tracking of multi-dimensional and multi-domain full time and space. Once the suspects and vehicles involved in the crime are locked, multi-dimensional and multi-domain intelligent tracking can be started immediately. For example, importing suspect information can obtain accurate behavior trajectory and location information for easy capture; enter the suspect vehicle to obtain accurate parking information Or moving track. Realize that each video probe, RFID, Wi-Fi probe becomes a 24 hours sleepless policeman, record the phenomenon, and identify the risk of the problem in a timely manner, proactively monitor, track, locate, integrate and analyze the information and data, and comprehensively evaluate , Actively trigger early warning, give decision-making auxiliary suggestions, and the command and dispatch system will promptly issue intervention and disposal instructions to the civilian police. At the same time, around the key places and areas, it is also possible to intelligently identify and identify the key people, monitor and track their behavior, and carry out early warning and prevention.

In summary, the intelligent tracking system of the whole time and space domain of the present invention is essentially a combat system based on supercomputing-based artificial intelligence + big data engine, based on the time axis as a node, with "human" as the core to build a static The data and the dynamic data formed by various intelligent perception systems are interlinked, integrated, correlated, comprehensively analyzed and judged to form a comprehensive combat system that can be intelligently discovered, intelligently researched and judged, intelligently triggered, and intelligently tracked. With this as the core support, it can truly form a transition from passive disposal to active discovery, prediction, prevention and early warning.

实施例Examples 22

In this embodiment, an information correlation analysis device is also provided, which is used to perform the steps in any of the above method embodiments, and the content that has been described will not be repeated here. 5 is a structural block diagram of an information correlation analysis device according to an embodiment of the present invention. As shown in FIG. 5, the device includes: an extraction module 50 for extracting related information of a target object from a surveillance video using a portrait multi-dimensional model, Among them, the relevant information of the target object includes: body characteristic information, behavior characteristic information, equipment information; the analysis module 52 is used to perform multi-dimensional correlation analysis on the extracted relevant information of the target object to obtain the character portrait and behavior track of the target object .

Through the above-mentioned device, the extraction module 50 uses the portrait multi-dimensional model to extract the relevant information of the target object from the surveillance video, where the relevant information of the target object includes: body characteristic information, behavior characteristic information, equipment information; the analysis module 52 analyzes the extracted target Multi-dimensional correlation analysis is performed on the relevant information of the object to obtain the character portrait and behavior track of the target object. It solves the problem that the manual browsing of surveillance video in the prior art is time-consuming and laborious, and the research and judgment from a single dimension leads to low accuracy. Through the fusion processing and research and judgment of each feature vector of the tracking object, a static is built with "people" as the core The data and the dynamic data formed by various intelligent perception systems are interlinked, integrated, correlated, comprehensively analyzed and judged to form a comprehensive combat system that can be intelligently discovered, intelligently researched and judged, intelligently triggered, and intelligently tracked. With this as the core support, it can truly form a transition from passive disposal to active discovery, prediction, prevention and early warning.

According to a preferred implementation of the embodiment of the present invention, the device further includes: an acquisition module for acquiring a surveillance video containing the target object, wherein the surveillance video includes at least one of the following image elements: the target object himself, and The vehicle associated with the target object, the iconic item associated with the target object.

The extraction module 50 is also used to: use the portrait multi-dimensional model to extract the relevant information of the target object from the surveillance video, wherein the portrait multi-dimensional model is a model built using feature vectors of multiple dimensions of the human body, and the portrait multi-dimensional model uses volume The product neural network performs feature extraction on the image.

The extraction module 50 is also used to: extract the body characteristic information of the target object from the surveillance video, wherein the body characteristic information includes at least one of the following: gender, height, hair accessories, clothing, carry-on items, walking form, facial characteristics, Head shape and facial decorations; extract the behavior characteristic information of the target object from the surveillance video, where the behavior characteristic information includes at least one of the following: cross-border behavior, active area, wandering area, aggregation behavior; extract the target object's behavior from the surveillance video Equipment information, wherein the equipment information includes at least one of the following: vehicle type, license plate, vehicle color, model, brand, car sticker, car accessories information, and target object's handheld mobile terminal.

According to a preferred implementation of the embodiment of the present invention, the device further includes: a positioning module, configured to locate the acquisition module according to the signal, and used to acquire the position information of the handheld mobile terminal of the target object.

The analysis module 52 includes: an association unit for correlating at least two types of information in the following related information of the target object: physical characteristic information, behavior characteristic information, equipment information, position information of the handheld mobile terminal; analysis unit, used In order to conduct a comprehensive analysis of the relevant information after the association in time, space, region, trajectory and social relationship.

The device further includes: a prediction module for predicting the future behavior of the target object based on the result of the comprehensive analysis; and a tracking module for tracking the location of the target object based on the result of the comprehensive analysis.

The device further includes a display module for displaying at least one of the following information: character portrait generation information, space-time retrospective display information, path tracking information, behavior habit model, safety situation map, and portrait real-time monitoring information.

实施例Examples 33

An embodiment of the present invention further provides a storage medium in which a computer program is stored, wherein the computer program is set to execute any of the steps in the above method embodiments during runtime.

Optionally, in this embodiment, the above storage medium may be set to store a computer program for performing the following steps:

S1. Extract relevant information of the target object from the surveillance video, where the relevant information of the target object includes: physical characteristic information, behavior characteristic information, and equipment information;

S2. Perform multi-dimensional correlation analysis on the extracted relevant information of the target object to obtain a character portrait and behavior track of the target object.

Optionally, the storage medium is also configured to store a computer program for performing the following steps:

Collect a surveillance video containing the target object, wherein the surveillance video includes at least one of the following image elements: the target object himself, the vehicle associated with the target object, and the target object Iconic items.

A portrait multi-dimensional model is used to extract relevant information of the target object from the surveillance video, wherein the portrait multi-dimensional model is a model built using feature vectors of multiple dimensions of the human body, and the portrait multi-dimensional model uses a convolutional neural network to characterize the image extract.

Optionally, in this embodiment, the foregoing storage medium may include, but is not limited to: a U disk, a read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory (referred to as RAM), mobile hard disk, magnetic disk or optical disk and other media that can store computer programs.

An embodiment of the present invention further provides an electronic device, including a memory and a processor, where the computer program is stored in the memory, and the processor is configured to run the computer program to perform any of the steps in the above method embodiments.

Optionally, the electronic device may further include a transmission device and an input-output device, wherein the transmission device is connected to the processor, and the input-output device is connected to the processor.

Optionally, in this embodiment, the foregoing processor may be configured to perform the following steps through a computer program:

Optionally, the processor is also configured to store a computer program for performing the following steps:

Use the portrait multi-dimensional model to extract the relevant information of the target object from the surveillance video, wherein the portrait multi-dimensional model is a model built using feature vectors of multiple dimensions of the human body, and the portrait multi-dimensional model uses a convolutional neural network to characterize the image extract.

For specific examples in this embodiment, reference may be made to the examples described in the foregoing embodiments and optional implementation manners, and details are not repeated in this embodiment.

Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the present invention can be implemented by a general-purpose computing device, they can be concentrated on a single computing device, or distributed in a network composed of multiple computing devices Above, optionally, they can be implemented with program code executable by the computing device, so that they can be stored in the storage device to be executed by the computing device, and in some cases, can be in a different order than here The steps shown or described are performed, or they are made into individual integrated circuit modules respectively, or multiple modules or steps among them are made into a single integrated circuit module for implementation. In this way, the present invention is not limited to any specific combination of hardware and software.

Industrial applicability

The above is only the preferred embodiments of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.

Claims

An information association analysis method, which is characterized by including:

Extract the relevant information of the target object from the surveillance video, wherein the relevant information of the target object includes: body characteristic information, behavior characteristic information, equipment information;

Multi-dimensional correlation analysis is performed on the extracted relevant information of the target object to obtain a character portrait and behavior track of the target object.
The method according to claim 1, wherein before extracting the relevant information of the target object from the surveillance video, the method further comprises:

Collect a surveillance video containing the target object, wherein the surveillance video includes at least one of the following image elements: the target object himself, the vehicle associated with the target object, and the target object Iconic items.
The method according to claim 1, wherein extracting the relevant information of the target object from the surveillance video includes:

Use the portrait multi-dimensional model to extract the relevant information of the target object from the surveillance video, wherein the portrait multi-dimensional model is a model built using feature vectors of multiple dimensions of the human body, and the portrait multi-dimensional model uses a convolutional neural network to characterize the image extract.
The method according to claim 1, wherein the relevant information of the target object extracted from the surveillance video includes:

Extract the body characteristic information of the target object from the surveillance video, wherein the body characteristic information includes at least one of the following: gender, height, hair accessories, clothing, carry-on items, walking shape, facial characteristics, head shape, Facial decorations

Extracting the behavior characteristic information of the target object from the surveillance video, wherein the behavior characteristic information includes at least one of the following: cross-border behavior, active area, wandering area, aggregation behavior;

Extract the equipment information of the target object from the surveillance video, where the equipment information includes at least one of the following: vehicle type, license plate, vehicle color, model, brand, car sticker, car accessories information, the target Target's handheld mobile terminal.
The method according to claim 4, wherein before performing multi-dimensional correlation analysis on the extracted relevant information of the target object, the method comprises:

Acquire the position information of the handheld mobile terminal of the target object according to the signal positioning.
The method according to claim 5, wherein performing multi-dimensional correlation analysis on the extracted relevant information of the target object includes:

Associate at least two types of information in the following related information of the target object: the physical characteristic information, the behavior characteristic information, the equipment information, and the position information of the handheld mobile terminal;

Comprehensive analysis of the relevant information after the association in time, space, region, trajectory and social relationship.
The method according to claim 6, characterized in that, after a comprehensive analysis of the related information after the association in time, space, region and trajectory, the method further comprises:

Predict the future behavior of the target object based on the results of the comprehensive analysis; and/or

According to the result of the comprehensive analysis, track the position of the target object.
The method according to claim 7, wherein the method further comprises:

At least one of the following information is displayed on the display screen: character portrait generation information, space-time retrospective display information, path tracking information, behavior habit model, security situation map, and real-time portrait monitoring information.
An information correlation analysis device, characterized in that it includes:

The extraction module is used to extract the relevant information of the target object from the surveillance video by using the multi-dimensional portrait model, wherein the relevant information of the target object includes: body characteristic information, behavior characteristic information, equipment information;

The analysis module is configured to perform multi-dimensional correlation analysis on the extracted relevant information of the target object to obtain a character portrait and behavior track of the target object.
A storage medium, characterized in that a computer program is stored in the storage medium, wherein the computer program is configured to execute the method described in any one of claims 1 to 8 during runtime.
An electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute any one of claims 1 to 8. Described method.