Automatic dynamic target identification system based on intelligent camera

Through the automatic identification of dynamic target system of the intelligent camera, combined with the data module and the learning module, the prediction and alarm of dynamic targets are achieved, the problem of intelligent cameras being unable to be promptly warned is solved, and the timeliness and security of the monitoring system is improved.

CN119068406AInactive Publication Date: 2024-12-03SHENZHEN NINGHE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410939931.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-15
Publication Date
2024-12-03
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure CN119068406A_ABST
    Figure CN119068406A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of video monitoring, in particular to an automatic dynamic target recognition system based on an intelligent camera, which comprises a dynamic recognition module, a data module, a data processing module, a data processing module and a display module, wherein the dynamic recognition module is used for dynamically recognizing a target in a picture shot by a camera terminal, dynamically analyzing the target in the dynamic picture and predicting the dynamic activity of the target; and the storage module is used for storing the image picture shot by the camera terminal and carrying out dynamic data analysis on the shot dynamic picture content. According to the automatic dynamic target recognition system based on the intelligent camera, the motion recognition module is arranged and can predict the motion law according to the recognized dynamic target of a monitoring picture, the abnormal dynamic target with regularity occurring in monitoring is locked, and then the dynamic target is automatically recognized according to previous data matching. And the abnormal dynamic track of the dynamic target is confirmed, and alarm information is sent to remind the monitoring staff, so that the monitoring system can prevent the existing potential risk in advance, and the monitoring security effect of the monitoring system is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of video monitoring, and in particular to an automatic dynamic target recognition system based on an intelligent camera. Background Art

[0002] Smart cameras mainly refer to security cameras for video surveillance, which are suitable for a series of security activities such as security inspections and real-time monitoring. Surveillance cameras are all remotely operated and real-time monitoring, with high resolution, anti-vibration, anti-glare and night vision functions. The biggest advantage of smart cameras is that various software can be installed in the background to meet security needs, such as face recognition, automatic tracking, filtering systems and video storage compression. During the monitoring process, there will be target activities, forming dynamic monitoring content. Therefore, dynamic recognition refers to the rapid recognition and judgment of the target status through real-time monitoring and analysis of the target, in the fields of security monitoring, intelligent transportation and industrial production.

[0003] The current smart cameras already have a certain degree of dynamic recognition capability. When an active target appears in a static monitoring environment in the screen monitoring, the smart camera can capture the dynamic target in the first time and can realize dynamic tracking. However, the current smart cameras can only perform dynamic recognition and cannot perform dynamic prediction based on the specific movements of the target during dynamic recognition. Therefore, if a thief steals under surveillance, there will usually be regular movements that can be traced, such as looking around, raising hands to test, and abnormal movement. The smart camera can only take pictures of these pictures, and the subsequent inspection or someone staring at the surveillance screen can know it. If the thief is caught after checking the surveillance, the property may have been used and squandered, causing losses to the victim. Therefore, the current smart camera automatic recognition of dynamic targets system lacks the function of dynamic recognition and early warning, which makes the timeliness of monitoring poor and cannot achieve the first-time alarm processing. Summary of the invention

[0004] The purpose of the present invention is to provide an automatic dynamic target recognition system based on an intelligent camera to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: an automatic dynamic target recognition system based on an intelligent camera, the automatic dynamic target recognition system comprising a data module, a dynamic recognition module and a learning module;

[0006] The automatic dynamic target recognition system is connected to a camera terminal and a monitoring terminal by wire. The camera terminal is used for shooting and monitoring the screen image and is a smart camera. The monitoring terminal is used for controlling and detecting the camera terminal and is a monitoring computer device.

[0007] The dynamic recognition module is used to dynamically recognize the target in the picture taken by the camera terminal, and dynamically analyze the target in the dynamic picture to predict the dynamic activity of the target;

[0008] The data module is used to store the images captured by the camera terminal, and to perform dynamic data analysis on the captured dynamic images, store images with abnormal dynamic information of the target, and then classify and store the monitoring content of the analysis data;

[0009] The learning module is used to model and learn the stored dynamic images, analyze and learn the overlapping dynamic content of the target, and generate a dynamic recognition plan for early warning.

[0010] In a further embodiment, the dynamic recognition module includes an environment confirmation module, a multi-point recognition module, a motion recognition module and a motion prediction module;

[0011] The environment confirmation module is used to intelligently identify the images captured by the camera terminal, and record and determine some fixed environmental targets in the captured images, and distinguish the positions of dynamic targets and environmental fixed targets;

[0012] The multi-point identification module allows the user to identify each dynamic target at multiple points when there is more than one dynamic target in the environment, and to identify and monitor the multiple dynamic targets simultaneously.

[0013] In a further embodiment, the motion recognition module is used to dynamically recognize targets in the monitoring screen, confirm different dynamic targets, and classify and recognize different dynamic targets according to the appearance types of the dynamic targets;

[0014] The motion prediction module is used to perform dynamic prediction analysis on the target in the monitoring image and analyze the further dynamic trajectory of the target according to the motion law.

[0015] In a further embodiment, the motion prediction module includes a model matching module, a dynamic confirmation module, a dynamic feedback module and an alarm module;

[0016] The model matching module analyzes the motion state of the dynamic target and matches the type after dynamically identifying the target in the monitoring screen, and confirms that the dynamic target corresponds to the next dynamic trajectory;

[0017] The dynamic confirmation module confirms the matching degree between the corresponding target and the analysis model according to the next actual dynamic trajectory of the corresponding target in the monitoring screen when the dynamic prediction of the target successfully matches the model.

[0018] In a further embodiment, the dynamic feedback module, when the corresponding monitoring dynamic target and the analysis model match to each other, the dynamic feedback module confirms that the dynamic target is abnormal and immediately generates abnormal monitoring information;

[0019] The alarm module is used to transmit the corresponding abnormal data information to the alarm module when the dynamic feedback module generates abnormal monitoring information. After the alarm module receives the abnormal monitoring information, it immediately sends out the alarm information.

[0020] In a further embodiment, the calculation formula of the motion prediction module is as follows:

[0021]

[0022] Where α is the displacement matching percentage, β is the control parameter, (x2.y2) is the coordinate after displacement, and (x1.y1) is the starting coordinate of displacement.

[0023] In a further embodiment, the data module includes a data storage module, a data analysis module and a data classification module;

[0024] The data storage module is used to store the content of the picture taken by the camera terminal, and to separately store the monitoring pictures with dynamic targets, while the static monitoring pictures store information in a normal covering storage manner.

[0025] In a further embodiment, the data analysis module is used to analyze the monitoring images with dynamic targets stored separately in the data storage module, summarize the rules of dynamic targets, and mark and extract the content with abnormal dynamic tracks of targets;

[0026] The data classification module is used to sort and classify the target abnormal dynamic trajectory content analyzed by the data analysis module according to similarity, and classify and store the dynamic abnormal target images of the same type.

[0027] In a further embodiment, the learning module includes a model building module, a dynamic simulation module and a simulation result module;

[0028] The model building module builds a dynamic prediction model according to the abnormal dynamic pictures stored in the data module, and generates a prediction model with a unified dynamic trajectory by using a large number of similar abnormal dynamic pictures.

[0029] In a further embodiment, the dynamic simulation module is used to conduct deep autonomous learning of the content of the prediction model, analyze the dynamic performance of the prediction model, and make judgments and predictions on the dynamic trajectory;

[0030] The simulation result module is used to generate visual model data with comparison based on the training results of the dynamic simulation module, and feed the simulation results back to the automatic dynamic target recognition system for reference by the dynamic recognition module.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. The present invention, by setting a motion recognition module, can predict the motion law according to the dynamic target identified in the monitoring screen, lock the abnormal dynamic target with regularity appearing in the monitoring, and then confirm the abnormal dynamic trajectory of the dynamic target according to the previous data matching, and send out an alarm message to remind the monitoring staff, so that the monitoring system can take early precautions against the potential risks in the process of dynamic recognition, thereby improving the monitoring security effect and monitoring safety of the monitoring system;

[0033] 2. The present invention analyzes and studies the dynamic information of the monitoring data through the learning module, determines the law of abnormal dynamic information, and generates a reference abnormal dynamic prediction scheme by establishing a learning model. Then, the motion recognition module can confirm in real time whether there is an abnormality in the dynamic target of the monitoring content according to data comparison, thereby ensuring the efficiency and safety of monitoring;

[0034] 3. The present invention stores a large amount of monitoring data by setting up a data module, and classifies and stores the stored monitoring data. After that, the learning model can perform simulation and predictive learning of dynamic rules with the support of a large amount of data, thereby improving the security effect of the monitoring dynamic automatic recognition dynamic target system, ensuring that the monitoring process can achieve early alarm and reduce the problems caused by monitoring lag. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A schematic diagram of the framework of the automatic dynamic target recognition system of the present invention;

[0036] Figure 2 It is a schematic diagram of the framework of the data module of the present invention;

[0037] Figure 3 It is a schematic diagram of the framework of the dynamic identification module of the present invention;

[0038] Figure 4 It is a schematic diagram of the framework of the learning module of the present invention. DETAILED DESCRIPTION

[0039] The following will be described clearly and completely in conjunction with the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Embodiment 1

[0041] See also Figure 1 - Figure 4 ,This embodiment provides an automatic dynamic target recognition system based on a smart camera, and the automatic dynamic target recognition system includes a data module, a dynamic recognition module and a learning module;

[0042] The automatic dynamic target recognition system is connected to a camera terminal and a monitoring terminal through wires. The camera terminal is used for shooting and monitoring the screen image and is a smart camera. The monitoring terminal is used for controlling and detecting the camera terminal and is a monitoring computer device.

[0043] The dynamic recognition module is used to dynamically recognize the target in the picture taken by the camera terminal, and dynamically analyze the target in the dynamic picture to predict the dynamic activity of the target;

[0044] The data module is used to store the images captured by the camera terminal, and to perform dynamic data analysis on the captured dynamic images, store images with abnormal dynamic information of the target, and then classify and store the monitoring content of the analysis data;

[0045] The learning module is used to model and learn the stored dynamic images, analyze and learn the overlapping dynamic content of the target, and generate dynamic recognition solutions for early warnings. Through deep learning, the safety and efficiency of the monitoring system can be improved.

[0046] The dynamic recognition module includes an environment confirmation module, a multi-point recognition module, a motion recognition module and a motion prediction module;

[0047] The environment confirmation module is used to intelligently identify the images captured by the camera terminal, and record and confirm some fixed environmental targets in the captured images, and distinguish the positions of dynamic targets and fixed environmental targets;

[0048] The multi-point identification module allows users to identify multiple dynamic targets at multiple points when there is more than one dynamic target in the environment, and simultaneously identify and monitor multiple dynamic targets. It can identify and monitor multiple dynamic targets synchronously to ensure comprehensive monitoring.

[0049] The motion recognition module is used to dynamically identify the targets in the monitoring screen, confirm different dynamic targets, and classify and identify different dynamic targets according to their appearance types;

[0050] The motion prediction module is used to perform dynamic prediction and analysis on the targets in the monitoring screen. According to the motion law, it analyzes the further dynamic trajectory of the target. By predicting the state of the corresponding dynamic target in advance, it can promptly confirm whether there are dangers and abnormalities of the dynamic target, and then arrange staff to deal with it in time, thereby improving the timeliness of monitoring and providing safer protection for the monitored area.

[0051] The motion prediction module includes a model matching module, a dynamic confirmation module, a dynamic feedback module and an alarm module;

[0052] The model matching module analyzes the motion state of the dynamic target and matches the type after dynamic recognition of the target in the monitoring screen, confirms that the dynamic target corresponds to the next dynamic trajectory, and facilitates the confirmation of abnormal activity patterns of real-time dynamic targets;

[0053] The dynamic confirmation module, when the dynamic prediction of the target successfully matches the model, confirms the matching degree between the corresponding target and the analysis model according to the actual dynamic trajectory of the corresponding target in the monitoring screen, and confirms whether there is any unsafe abnormal movement of the real-time dynamic target, thereby ensuring the safety of the monitoring environment.

[0054] Dynamic feedback module: When the corresponding monitoring dynamic target matches the analysis model, the dynamic feedback module confirms that the dynamic target is abnormal and immediately generates abnormal monitoring information. After confirming the abnormal dynamic target, it promptly feedbacks the alarm to enhance the monitoring effect of the system.

[0055] The alarm module is used to transmit the corresponding abnormal data information to the alarm module when the dynamic feedback module generates abnormal monitoring information. After receiving the abnormal monitoring information, the alarm module immediately sends out an alarm message to facilitate early alarm, ensuring that the staff can view the monitoring content in time and make deployment arrangements in advance, reducing unnecessary losses caused by the lag of monitoring.

[0056] The calculation formula of the motion prediction module is as follows:

[0057]

[0058] Among them, α is the displacement matching percentage, β is the comparison parameter, (x2.y2) is the coordinate after displacement, (x1.y1) is the starting coordinate of displacement. The dynamic target moves dynamically within the monitoring area, and the horizontal and vertical coordinates can be set, and capture points can be set at the key positions of the dynamic target. According to the moving position of the point, the motion law of the dynamic target is confirmed. According to the motion trajectory, the motion law of the real-time dynamic target can be judged in advance. If the capture point of the dynamic target continues to move according to the comparison data, it can be judged that the dynamic target has abnormal activity, and then the alarm reminder is enhanced to ensure the efficiency of monitoring and security.

[0059] Example 2

[0060] See also Figure 2 , further improvements were made on the basis of Example 1:

[0061] The data module includes a data storage module, a data analysis module and a data classification module;

[0062] The data storage module is used to store the content of the pictures taken by the camera terminal, and to separately store the monitoring pictures with dynamic targets. The static monitoring pictures store information in a normal covering storage manner.

[0063] The data analysis module is used to analyze the monitoring images with dynamic targets stored separately in the data storage module, summarize the rules of dynamic targets, and mark and extract the contents with abnormal dynamic tracks of targets;

[0064] The data classification module is used to sort and classify the abnormal dynamic trajectory content of the target analyzed by the data analysis module according to similarity, and to classify and store the dynamic abnormal target images of the same type. Through classified storage, the previous monitoring data is judged and stored in a targeted manner, so as to facilitate the analysis and learning of the data and improve the utilization rate of the data.

[0065] Example 3

[0066] See also Figure 4 , further improvements were made on the basis of Example 1:

[0067] The learning modules include model building module, dynamic simulation module and simulation result module;

[0068] The model building module builds a dynamic prediction model based on the abnormal dynamic images stored in the data module, uses a large number of similar abnormal dynamic images to generate a prediction model with a unified dynamic trajectory, and uses a large amount of data to build a dynamic learning model. The monitoring can then perform deep learning simulation on the movement laws of abnormal dynamic targets, thereby providing effective comparison content for the monitoring system, making it easier for the monitoring system to more accurately predict whether the movement trajectory of the dynamic target is unsafe.

[0069] The dynamic simulation module is used to conduct in-depth autonomous learning of the content of the prediction model, analyze the dynamic performance of the prediction model, and judge and predict the dynamic trajectory. Based on a large amount of data, the overlapping data is combined, and different dynamic target action patterns are simulated to generate dynamic target prediction data that can be used as a reference. When the dynamic target moves the coordinates of the corresponding pixel points in the monitoring screen, it can quickly confirm whether the dynamic target has risk anomalies, ensuring that the monitoring system can confirm the risks and anomalies of the dynamic target more timely;

[0070] The simulation result module is used to generate comparative visual model data based on the training results of the dynamic simulation module, and feed the simulation results back to the automatic dynamic target identification system for reference by the dynamic identification module. The simulation results generated by the model can be used as a reference for motion prediction of dynamic targets, which plays a good role in early risk avoidance for security monitoring and ensures that monitoring can achieve real-time security effects.

[0071] The smart camera automatically identifies dynamic targets. During use, the camera terminal performs real-time shooting and monitoring of the monitoring area. When a dynamic target appears in the picture, the camera terminal locks the dynamic target, and then the dynamic recognition module locks and monitors the corresponding dynamic target. When the monitored dynamic target continues to move, the motion path of the dynamic target is recorded by the motion recognition module, and then the motion prediction module analyzes the recorded motion path, and then calls the dynamic data generated by the simulation result module in the learning module for matching. After the motion path of the dynamic target matches the simulation data, when the matching degree is consistent or the similarity is higher than 80%, the alarm module sends an alarm message, and the staff at the monitoring terminal can know the risk of the dynamic target at the first time.

[0072] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic dynamic target recognition system based on an intelligent camera, characterized in that: The automatic dynamic target recognition system includes a data module, a dynamic recognition module and a learning module; The automatic dynamic target recognition system is connected to a camera terminal and a monitoring terminal by wire. The camera terminal is used for shooting and monitoring the screen image and is a smart camera. The monitoring terminal is used for controlling and detecting the camera terminal and is a monitoring computer device. The dynamic recognition module is used to dynamically recognize the target in the picture taken by the camera terminal, and dynamically analyze the target in the dynamic picture to predict the dynamic activity of the target; The data module is used to store the images captured by the camera terminal, and to perform dynamic data analysis on the captured dynamic images, store images with abnormal dynamic information of the target, and then classify and store the monitoring content of the analysis data; The learning module is used to model and learn the stored dynamic images, analyze and learn the overlapping dynamic content of the target, and generate a dynamic recognition plan for early warning.

2. The automatic dynamic target recognition system based on a smart camera according to claim 1, characterized in that: The dynamic recognition module includes an environment confirmation module, a multi-point recognition module, a motion recognition module and a motion prediction module; The environment confirmation module is used to intelligently identify the images captured by the camera terminal, and record and determine some fixed environmental targets in the captured images, and distinguish the positions of dynamic targets and environmental fixed targets; The multi-point identification module allows the user to identify each dynamic target at multiple points when there is more than one dynamic target in the environment, and to identify and monitor the multiple dynamic targets simultaneously.

3. The automatic dynamic target recognition system based on a smart camera according to claim 2 is characterized in that: The motion recognition module is used to dynamically recognize targets in the monitoring screen, confirm different dynamic targets, and classify and recognize different dynamic targets according to their appearance types; The motion prediction module is used to perform dynamic prediction analysis on the target in the monitoring image and analyze the further dynamic trajectory of the target according to the motion law.

4. The automatic dynamic target recognition system based on a smart camera according to claim 2 is characterized in that: The motion prediction module includes a model matching module, a dynamic confirmation module, a dynamic feedback module and an alarm module; The model matching module analyzes the motion state of the dynamic target and matches the type after dynamically identifying the target in the monitoring screen, and confirms that the dynamic target corresponds to the next dynamic trajectory; The dynamic confirmation module confirms the matching degree between the corresponding target and the analysis model according to the next actual dynamic trajectory of the corresponding target in the monitoring screen when the dynamic prediction of the target successfully matches the model.

5. The automatic dynamic target recognition system based on a smart camera according to claim 4 is characterized in that: The dynamic feedback module, when the corresponding monitoring dynamic target and the analysis model match to each other, the dynamic feedback module confirms that the dynamic target is abnormal and immediately generates abnormal monitoring information; The alarm module is used to transmit the corresponding abnormal data information to the alarm module when the dynamic feedback module generates abnormal monitoring information. After the alarm module receives the abnormal monitoring information, it immediately sends out the alarm information.

6. The automatic dynamic target recognition system based on a smart camera according to claim 2, characterized in that: The calculation formula of the motion prediction module is as follows: Where α is the displacement matching percentage, β is the control parameter, (x2.y2) is the coordinate after displacement, and (x1.y1) is the starting coordinate of displacement.

7. The automatic dynamic target recognition system based on a smart camera according to claim 1, characterized in that: The data module includes a data storage module, a data analysis module and a data classification module; The data storage module is used to store the content of the picture taken by the camera terminal, and to separately store the monitoring pictures with dynamic targets, while the static monitoring pictures store information in a normal covering storage manner.

8. The automatic dynamic target recognition system based on a smart camera according to claim 7, characterized in that: The data analysis module is used to analyze the monitoring images with dynamic targets stored separately in the data storage module, summarize the laws of dynamic targets, and mark and extract the content with abnormal dynamic tracks of targets; The data classification module is used to sort and classify the target abnormal dynamic trajectory content analyzed by the data analysis module according to similarity, and classify and store the dynamic abnormal target images of the same type.

9. The automatic dynamic target recognition system based on a smart camera according to claim 1, characterized in that: The learning module includes a model building module, a dynamic simulation module and a simulation result module; The model building module builds a dynamic prediction model according to the abnormal dynamic pictures stored in the data module, and generates a prediction model with a unified dynamic trajectory by using a large number of similar abnormal dynamic pictures.

10. The automatic dynamic target recognition system based on a smart camera according to claim 9, characterized in that: The dynamic simulation module is used to conduct in-depth autonomous learning of the content of the prediction model, analyze the dynamic performance of the prediction model, and make judgments and predictions on the dynamic trajectory; The simulation result module is used to generate visual model data with comparison based on the training results of the dynamic simulation module, and feed the simulation results back to the automatic dynamic target recognition system for reference by the dynamic recognition module.