Artificial information checking and collecting and early warning method based on video data

Abstract

The invention relates to the technical field of video data analysis and processing, and discloses a video data-based artificial information checking, collecting and early warning method, which comprises the following steps of: S1, collecting video data from cameras mounted at different positions through an identification system; s2, preprocessing the collected video; s3, key features of the preprocessed video are extracted; a predetermined risk area and different types of predetermined transaction areas are divided based on a scene view, risk areas are dynamically divided according to historical data and a real-time monitoring result, key monitoring is performed on a high-risk area, the pertinence and effectiveness of risk management are improved, manual information check and collection are performed on behavior tracks with scores higher than a predetermined threshold, and the risk management efficiency is improved. And the result is fed back to the system in time, and subsequent analysis and early warning are adjusted and optimized, so that the intelligent level and accuracy of the early warning system are improved, the overall safety and reliability of the monitoring system are improved, and potential safety risks are effectively prevented and processed.

Description

Technical Field

[0001] The present invention relates to the technical field of video data analysis and processing, and in particular to an artificial information verification and early warning method based on video data. Background Art

[0002] With the development of the Internet and mobile communication technologies, the amount of video data has exploded. In application scenarios such as public safety monitoring and intelligent traffic management, how to quickly and accurately extract useful information from massive video data has become a challenge.

[0003] Traditional video surveillance systems mainly rely on manual monitoring, which is not only inefficient, but also difficult to effectively identify and warn of potential security risks. With the development of video analysis technology, although some systems can achieve automatic anomaly detection, their effectiveness in practical applications is often limited by problems such as high false alarm rate and poor adaptability to complex scenarios. Especially in large public places or complex environments, the risk levels and transaction activity types in different areas vary. A single threshold or fixed rule is difficult to accurately distinguish and respond to various risk situations. For this reason, a manual information collection and early warning method based on video data is proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a manual information verification and early warning method based on video data to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for artificial information verification and early warning based on video data, comprising the following steps:

[0006] S1. Collect video data from cameras installed at different locations through the recognition system;

[0007] S2, preprocessing the collected videos;

[0008] S3, extracting key features from the preprocessed video;

[0009] S4, use key features to perform portrait recognition and behavior pattern recognition, and record the activity range and behavior trajectory of the target object;

[0010] S5, mapping the behavior trajectory to a scene view generated based on the object scene, and dividing the predetermined risk area and different types of predetermined transaction areas based on the scene view;

[0011] S6. For behavioral trajectories with scores higher than a predetermined threshold, manual information verification is performed;

[0012] S7. When manual nuclear mining confirms the existence of potential risks, issue early warning information.

[0013] Preferably, in the above: in S1, the recognition system includes a video acquisition module, a preprocessing module, a target detection module, an abnormal behavior recognition module, a risk score calculation module and an early warning generation module;

[0014] The video acquisition module is used to obtain real-time video data from the surveillance camera;

[0015] The preprocessing module is used to preprocess the collected video data;

[0016] The target detection module is used for target detection and tracking;

[0017] The abnormal behavior recognition module is used to recognize abnormal behavior in the video;

[0018] The risk score calculation module is used to quantify the degree of potential risk;

[0019] The warning generation module is used to trigger the manual information verification process and send warning information to relevant personnel when the risk score exceeds a preset threshold.

[0020] Preferably, the above-mentioned: the video acquisition module is connected to the preprocessing module, the preprocessing module is connected to the target detection module, the target detection module is connected to the abnormal behavior recognition module, the abnormal behavior recognition module is connected to the risk score calculation module, and the risk score calculation module is connected to the early warning generation module.

[0021] Preferably, the preprocessing module comprises a denoising unit, a frame rate adjustment unit and an image enhancement unit;

[0022] The denoising unit is connected to the frame rate adjustment unit, and the frame rate adjustment unit is connected to the image enhancement unit;

[0023] The denoising unit is used to reduce noise in the video;

[0024] The frame rate adjustment unit is used to adjust the frame rate of the video to match the requirements of subsequent analysis;

[0025] The image enhancement unit is used to enhance the contrast and brightness of the image.

[0026] Preferably, the target detection module comprises a detection unit and a tracking unit;

[0027] The detection unit is connected to the tracking unit;

[0028] The detection unit is used to detect key targets in the video frame;

[0029] The tracking unit is used to continuously track the detected target and obtain its motion trajectory.

[0030] Preferably, the abnormal behavior identification module comprises a feature extraction unit and an abnormal behavior matching unit;

[0031] The feature extraction unit is connected to the abnormal behavior matching unit;

[0032] The feature extraction unit is used to extract features related to the target behavior from the results of target detection and tracking;

[0033] The abnormal behavior matching unit is used to match the extracted features with a preset abnormal behavior pattern library.

[0034] Preferably, in the feature extraction unit, the calculation formula of feature extraction (F) is:

[0035] F=CNN(V)

[0036] Among them, CNN is a convolutional neural network and V is a video frame.

[0037] Preferably, in the above-mentioned abnormal behavior matching unit, the calculation formula of abnormal detection (D) in the abnormal behavior matching is:

[0038] D=SVM(M)

[0039] Among them, SVM is support vector machine, and M is the fused information.

[0040] Preferably, the above-mentioned warning generation module includes a risk score (R), a warning trigger (W) and a risk threshold (T);

[0041] The risk score (R) calculation formula is:

[0042]

[0043] Among them, ω i is the weight of the i-th feature, X i is the value of the ith feature.

[0044] Preferably, the calculation formula of the warning trigger (W) is:

[0045] W=h(R,T)

[0046] Among them, h represents the warning decision function, T represents the preset threshold, and when R>T, the warning is triggered.

[0047] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0048] 1. Real-time and integrity monitoring ensures that there is no delay or loss in video data, as well as data synchronization between multiple cameras. The anomaly detection function in the preprocessing stage can detect and handle abnormal situations in a timely manner to prevent potential security risks.

[0049] Second, by extracting key features and fusing them to form a more comprehensive feature description, the accuracy of subsequent recognition and analysis can be improved. By combining multi-camera data to achieve continuous tracking and recording of the target object's behavior trajectory and considering the target object's environment and contextual information, the behavior pattern can be more accurately interpreted and judged, and the behavioral characteristics of the target object can be more deeply understood.

[0050] 3. Divide the predetermined risk areas and different types of predetermined transaction areas based on the scenario view, and dynamically divide the risk areas according to historical data and real-time monitoring results, focus on monitoring high-risk areas, improve the pertinence and effectiveness of risk management, conduct manual information verification for behavioral trajectories with scores higher than the predetermined threshold, and promptly feed back the results to the system, adjust and optimize subsequent analysis and early warning, thereby improving the intelligence level and accuracy of the early warning system, improving the overall safety and reliability of the monitoring system, effectively preventing and handling potential safety risks, and providing strong protection for the safety of people and property. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 It is a schematic diagram of the early warning method of the present invention;

[0053] Figure 2 This is a connection diagram of the identification system of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe 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.

[0055] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the conditions under which this application can be implemented. Therefore, they have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should still fall within the scope of the technical content disclosed in this application without affecting the effects and purposes that can be achieved by this application.

[0056] Example

[0057] See also Figure 1-2 The present invention provides a technical solution: a method for artificial information collection and early warning based on video data, comprising the following steps:

[0058] S1. Collect video data from cameras installed at different locations through the recognition system; increase the real-time and integrity monitoring of video data, ensure that the video data collected from each camera is not delayed or lost, and synchronize and coordinate data processing between multiple cameras, so as to continuously track the behavior of the same target object under different cameras;

[0059] S2. Preprocess the collected videos; add anomaly detection functions in the preprocessing stage, such as sudden appearance of objects, color anomalies, etc., so as to detect and handle abnormal situations in time;

[0060] S3. Extract key features from the preprocessed video; fuse different features to form a more comprehensive feature description to improve the accuracy of subsequent recognition and analysis; and continuously optimize and update the features according to actual application scenarios and data analysis results to meet different monitoring needs.

[0061] S4. Use key features to perform portrait recognition and behavior pattern recognition, record the target object's activity range and behavior trajectory; combine multi-camera data to achieve continuous tracking and recording of the target object's behavior trajectory, consider the target object's environment and context information, and make a more accurate interpretation and judgment of the behavior pattern;

[0062] S5. Mapping the behavior trajectory to a scene view generated based on the object scene, dividing the predetermined risk area and different types of predetermined transaction areas based on the scene view; dynamically dividing the risk area based on historical data and real-time monitoring results, so as to focus on monitoring high-risk areas;

[0063] S6. For behavioral trajectories with scores higher than the predetermined threshold, manual information verification is carried out; behavioral trajectories suspected of being abnormal or high-risk are marked for manual verification, and the results of manual verification are promptly fed back to the system to adjust and optimize subsequent analysis and early warning.

[0064] S7. When manual nuclear mining confirms the existence of potential risks, early warning information will be issued. According to the severity of the potential risks, different warning levels will be divided, such as minor, moderate, severe, etc., so that relevant personnel can take corresponding response measures according to the warning level, establish a follow-up processing mechanism for early warning, track, process and feedback the early warning information, so as to continuously improve and optimize the early warning process.

[0065] The recognition system includes a video acquisition module, a preprocessing module, a target detection module, an abnormal behavior recognition module, a risk score calculation module and an early warning generation module; the video acquisition module is connected to the preprocessing module, the preprocessing module is connected to the target detection module, the target detection module is connected to the abnormal behavior recognition module, the abnormal behavior recognition module is connected to the risk score calculation module, and the risk score calculation module is connected to the early warning generation module.

[0066] Video acquisition module, used to obtain real-time video data from surveillance cameras;

[0067] A preprocessing module, used for preprocessing the collected video data;

[0068] The preprocessing module includes a denoising unit, a frame rate adjustment unit and an image enhancement unit; the denoising unit is connected to the frame rate adjustment unit, and the frame rate adjustment unit is connected to the image enhancement unit;

[0069] De-noising unit, used to reduce noise in the video;

[0070] A frame rate adjustment unit, used to adjust the frame rate of the video to match the requirements of subsequent analysis;

[0071] The image enhancement unit is used to improve the contrast and brightness of the image.

[0072] The target detection module is used for target detection and tracking; the target detection module includes a detection unit and a tracking unit; the detection unit is connected to the tracking unit;

[0073] A detection unit, used to detect key targets in video frames;

[0074] The tracking unit is used to continuously track the detected target and obtain its motion trajectory.

[0075] The abnormal behavior recognition module is used to recognize abnormal behavior in the video; the abnormal behavior recognition module includes a feature extraction unit and an abnormal behavior matching unit; the feature extraction unit is connected to the abnormal behavior matching unit;

[0076] A feature extraction unit, used to extract features related to target behavior from the results of target detection and tracking;

[0077] The abnormal behavior matching unit is used to match the extracted features with a preset abnormal behavior pattern library.

[0078] In the feature extraction unit, the calculation formula of feature extraction (F) is:

[0079] F=CNN(V)

[0080] Among them, CNN is a convolutional neural network and V is a video frame.

[0081] In the abnormal behavior matching unit, the calculation formula of anomaly detection (D) in abnormal behavior matching is:

[0082] D=SVM(M)

[0083] Among them, SVM is support vector machine, and M is the fused information.

[0084] The risk score calculation module is used to quantify the degree of potential risk. Based on the actual application scenario and data analysis results, it dynamically selects the features that have the greatest impact on the risk score, and uses machine learning algorithms such as gradient boosting and random forest to adaptively adjust feature weights to reflect the relative importance of different features in the current scenario.

[0085] The early warning generation module is used to trigger the manual information verification process and send early warning information to relevant personnel when the risk score exceeds the preset threshold.

[0086] In the warning generation module, it includes risk score (R), warning trigger (W) and risk threshold (T);

[0087] The risk score (R) calculation formula is:

[0088]

[0089] Among them, ω i is the weight of the i-th feature, X i is the value of the ith feature.

[0090] The calculation formula for early warning trigger (W) is:

[0091] W=h(R,T)

[0092] Among them, h represents the warning decision function, T represents the preset threshold, and when R>T, the warning is triggered.

[0093] Set multiple warning thresholds corresponding to different warning levels, such as minor, moderate, severe, etc., so that relevant personnel can take corresponding response measures according to the warning level, and include more useful information in the warning information, such as risk score, trigger conditions, relevant feature descriptions, historical risk records, etc., so that relevant personnel can quickly understand the risk situation and make decisions.

[0094] In summary, real-time and integrity monitoring ensures that video data is free of delay and loss, as well as data synchronization between multiple cameras. In addition, the anomaly detection function in the preprocessing stage can detect and handle abnormal situations in a timely manner to prevent potential security risks.

[0095] By extracting key features and fusing them to form a more comprehensive feature description, the accuracy of subsequent recognition and analysis can be improved. By combining multi-camera data to achieve continuous tracking and recording of the target object's behavior trajectory and taking into account the target object's environment and contextual information, the behavior pattern can be more accurately interpreted and judged, and the behavioral characteristics of the target object can be more deeply understood.

[0096] Based on the scenario view, the predetermined risk areas and different types of predetermined transaction areas are divided, and the risk areas are dynamically divided according to historical data and real-time monitoring results. High-risk areas are monitored in key areas to improve the pertinence and effectiveness of risk management. For behavioral trajectories with scores higher than the predetermined threshold, manual information verification is carried out, and the results are fed back to the system in a timely manner. The subsequent analysis and early warning are adjusted and optimized, thereby improving the intelligence level and accuracy of the early warning system, improving the overall safety and reliability of the monitoring system, effectively preventing and handling potential safety risks, and providing strong protection for the safety of people and property.

[0097] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present invention may be combined and / or combined in various ways, even if such combinations and / or combinations are not explicitly described in the present invention. In particular, the features described in the various embodiments and / or claims of the present invention may be combined and / or combined in various ways without departing from the spirit and teachings of the present invention. All of these combinations and / or combinations fall within the scope of the present invention.

Claims

1. A method for artificial information collection and early warning based on video data, characterized in that: The following steps are involved: S1. Collect video data from cameras installed at different locations through the recognition system; S2, preprocessing the collected videos; S3, extracting key features from the preprocessed video; S4, use key features to perform portrait recognition and behavior pattern recognition, and record the activity range and behavior trajectory of the target object; S5, mapping the behavior trajectory to a scene view generated based on the object scene, and dividing the predetermined risk area and different types of predetermined transaction areas based on the scene view; S6. For behavioral trajectories with scores higher than a predetermined threshold, manual information verification is performed; S7. When manual nuclear mining confirms the existence of potential risks, issue early warning information.

2. The method for manual information collection and early warning based on video data according to claim 1 is characterized by: In S1, the recognition system includes a video acquisition module, a preprocessing module, a target detection module, an abnormal behavior recognition module, a risk score calculation module and an early warning generation module; The video acquisition module is used to obtain real-time video data from the surveillance camera; The preprocessing module is used to preprocess the collected video data; The target detection module is used for target detection and tracking; The abnormal behavior recognition module is used to recognize abnormal behavior in the video; The risk score calculation module is used to quantify the degree of potential risk; The warning generation module is used to trigger the manual information verification process and send warning information to relevant personnel when the risk score exceeds a preset threshold.

3. The method for manual information collection and early warning based on video data according to claim 2 is characterized in that: The video acquisition module is connected to the preprocessing module, the preprocessing module is connected to the target detection module, the target detection module is connected to the abnormal behavior recognition module, the abnormal behavior recognition module is connected to the risk score calculation module, and the risk score calculation module is connected to the early warning generation module.

4. The method for manual information collection and early warning based on video data according to claim 3 is characterized in that: The pre-processing module includes a denoising unit, a frame rate adjustment unit and an image enhancement unit; The denoising unit is connected to the frame rate adjustment unit, and the frame rate adjustment unit is connected to the image enhancement unit; The denoising unit is used to reduce noise in the video; The frame rate adjustment unit is used to adjust the frame rate of the video to match the requirements of subsequent analysis; The image enhancement unit is used to enhance the contrast and brightness of the image.

5. The method for manual information collection and early warning based on video data according to claim 3 is characterized by: The target detection module includes a detection unit and a tracking unit; The detection unit is connected to the tracking unit; The detection unit is used to detect key targets in the video frame; The tracking unit is used to continuously track the detected target and obtain its motion trajectory.

6. The method for manual information collection and early warning based on video data according to claim 3 is characterized by: The abnormal behavior identification module includes a feature extraction unit and an abnormal behavior matching unit; The feature extraction unit is connected to the abnormal behavior matching unit; The feature extraction unit is used to extract features related to the target behavior from the results of target detection and tracking; The abnormal behavior matching unit is used to match the extracted features with a preset abnormal behavior pattern library.

7. The method for manual information collection and early warning based on video data according to claim 6 is characterized by: In the feature extraction unit, the calculation formula of feature extraction (F) is: F=CNN(V) Among them, CNN is a convolutional neural network and V is a video frame.

8. The method for manual information collection and early warning based on video data according to claim 6 is characterized by: In the abnormal behavior matching unit, the calculation formula of the abnormal detection (D) in the abnormal behavior matching is: D=SVM(M) Among them, SVM is support vector machine, and M is the fused information.

9. The method for manual information collection and early warning based on video data according to claim 3 is characterized by: The warning generation module includes risk score (R), warning trigger (W) and risk threshold (T); The risk score (R) calculation formula is: Among them, ω i is the weight of the i-th feature, X i is the value of the ith feature.

10. The method for manual information collection and early warning based on video data according to claim 9, characterized in that: The calculation formula of the early warning trigger (W) is: W=h(R,T) Among them, h represents the warning decision function, T represents the preset threshold, and when R>T, the warning is triggered.

Images