Video monitoring data processing system based on cloud computing

Abstract

The invention relates to the technical field of data processing, in particular to a video monitoring data processing system based on cloud computing, which comprises a video quality synchronization module, an abnormal behavior identification module, a resource scheduling optimization module and a safety early warning trigger module, and synchronizes a video stream timestamp through a cloud parallel processing technology based on multi-source video data. According to the invention, the video stream timestamp and the image quality are precisely synchronized through the cloud parallel processing technology, the high quality of the data stream is ensured, the screening mechanism ensures that only the video stream meeting the quality standard is further analyzed, and the accuracy of subsequent abnormal behavior analysis is enhanced. The system not only can analyze and quickly identify abnormal behaviors deviating from a normal track in real time, but also can dynamically adjust, calculate and store resources according to actual resource use conditions, so that the processing speed and the cost efficiency are improved, the system can flexibly meet monitoring requirements of different scales, and the system is particularly suitable for a dynamic environment and has wide application prospects. And the early warning and response capabilities of video monitoring are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a video monitoring data processing system based on cloud computing. Background Art

[0002] Data processing technology is used to capture, store, sort, transform and extract information. This field plays a core role in modern information technology and is the basis for efficient and accurate data analysis. Data processing has a wide range of applications, including data mining, machine learning, image and signal processing, and big data technology. It can process and analyze large-scale data sets, enabling enterprises and individuals to use virtualized resources for complex data analysis tasks without investing in expensive physical hardware.

[0003] Among them, the cloud computing video surveillance data processing system is a system that uses cloud resources to store, process and analyze video surveillance data. Its main purpose is to improve the efficiency and effectiveness of security monitoring, allowing users to remotely access and manage surveillance videos. Through the cloud platform, it can not only provide large-scale data storage solutions, but also support advanced video analysis functions such as motion detection, facial recognition and abnormal behavior analysis. It is particularly suitable for environments that require high scalability and accessibility, such as urban safety, traffic monitoring and corporate security.

[0004] Traditional systems mostly rely on physical hardware and local storage. When processing large-scale video data, they are often constrained by insufficient resources and low processing capabilities, which limits the scalability and flexibility of the system. In particular, it is difficult to effectively expand monitoring capabilities when the amount of data surges. In addition, traditional technologies are often insufficient in real-time and abnormal behavior recognition and processing, especially in emergency situations, such as urban security and traffic monitoring scenarios. It is impossible to identify and handle emergencies in real time, resulting in untimely security warnings and responses. In large-scale events, insufficient real-time data processing capabilities will lead to security vulnerabilities, thus affecting the effectiveness of video surveillance. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a video surveillance data processing system based on cloud computing.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a video surveillance data processing system based on cloud computing, the system comprising:

[0007] The video quality synchronization module is based on multi-source video data. It synchronizes the video stream timestamp through cloud parallel processing technology, analyzes the clarity and signal integrity of each frame, verifies the quality of the video stream, and selects the video stream that meets the quality standards to obtain the video quality standard stream;

[0008] The abnormal behavior recognition module performs real-time analysis on the movement trajectories of pedestrians and vehicles in the video based on the video quality standard stream, identifies abnormal behaviors that deviate from normal trajectories, and compares and analyzes them with standard behaviors to generate abnormal behavior records;

[0009] The resource scheduling optimization module evaluates the utilization efficiency and load of the current cloud platform resources based on the abnormal behavior records, dynamically adjusts the computing resources and storage allocation according to the analysis results and the prediction of future demand, and generates resource allocation configuration;

[0010] The security warning trigger module analyzes the severity and urgency of abnormal behavior based on the resource allocation configuration, formulates a security warning process according to the risk rating standard, and automatically triggers an alarm to generate a warning response result.

[0011] The present invention is improved in that the step of synchronizing the video stream timestamp is specifically as follows:

[0012] Based on multi-source video data, original timestamps are collected and stored in the cloud library to obtain the original timestamp set;

[0013] Selecting a reference timestamp from the original timestamp set, calculating the deviation between the timestamps of multiple video streams and the reference timestamp, and obtaining a time deviation set;

[0014] Based on the time deviation set, the average deviation of the video source timestamp is calculated using the formula:

[0015]

[0016] Get the synchronized video stream timestamp t′ sync , where t base is the reference timestamp, Δt i is the deviation between the i-th video source and the reference time, and n is the number of video sources.

[0017] The present invention is improved in that the step of obtaining the video quality standard stream is specifically as follows:

[0018] Analyzing the clarity and signal integrity of each frame of the image, evaluating image clarity and data loss, and obtaining an image quality score;

[0019] Based on the image quality score, the formula is used:

[0020] QS=ws1×CS+ws2×IS

[0021] The quality of each frame of the image is evaluated, and the video frames exceeding the threshold are screened to obtain the video quality standard stream, where QS represents the quality score of each frame of video, CS represents clarity, IS represents signal integrity, and w1 and w2 are weight parameters.

[0022] The present invention is improved in that the step of identifying abnormal behavior that deviates from the normal trajectory is specifically:

[0023] Based on the video quality standard stream, pedestrians and vehicle targets in the video are tracked in real time, the motion trajectory of each pedestrian and vehicle is identified, and preliminary motion trajectory data is obtained;

[0024] Based on the preliminary motion trajectory data, the deviation between the current trajectory and the standard trajectory is calculated using the formula:

[0025]

[0026] Get the deviation score ΔSV for each target, where x act,t and act,t Represents the coordinate point currently monitored, x norm,t and norm,t represents the coordinate point of the standard trajectory, T represents the total number of time steps or frames in the detection cycle;

[0027] Based on the deviation score of each target, whether the behavior is abnormal is evaluated, and the target position and time are marked to obtain an abnormal behavior recognition result.

[0028] The present invention is improved in that the step of obtaining the abnormal behavior record is specifically as follows:

[0029] Based on the abnormal behavior, a preliminary comparative analysis of the data was conducted using the formula:

[0030]

[0031] Calculate the abnormal behavior index EV, where δX pos,i is the position deviation of target i from the standard trajectory, α x is the adjustment factor, n x is the total number of targets;

[0032] Based on the abnormal behavior index, the location, timestamp and degree of deviation of each target are recorded and formatted as an abnormal behavior record.

[0033] The present invention is improved in that the evaluation steps of the usage efficiency and load condition are specifically as follows:

[0034] Based on the abnormal behavior records, cloud platform resource data is collected, including CPU usage, memory usage, storage read and write rates, and network bandwidth usage, to obtain a resource usage data set;

[0035] Analyze the resource usage data set, calculate resource usage efficiency and load, and use the formula:

[0036]

[0037] According to the different resource types, the average cloud platform utilization EU is calculated, where UO i represents the current utilization rate of the i-th resource, WO i is the weight coefficient of the i-th resource, n o The total number of resource types.

[0038] The present invention is improved in that the step of obtaining the resource allocation configuration is specifically as follows:

[0039] Based on the analysis results, determine the current usage status and future demand of resources using the formula:

[0040] RG future =RG current ×(1+β g ×G y )

[0041] Predict future resource requirementsRG future , where RG current is the current resource usage, β g is the predicted business growth rate, G y is the coefficient of impact of business growth on resource demand;

[0042] Based on the future resource demand, computing resources and storage allocation are dynamically adjusted to obtain resource allocation configuration.

[0043] The present invention is improved in that the step of obtaining the early warning response result is specifically as follows:

[0044] Based on the resource allocation configuration, each abnormal behavior is detected and recorded, and the impact is evaluated according to the type, frequency and impact on resource usage of the abnormal behavior to obtain a temporary severity score;

[0045] Based on the provisional severity score, compared with the risk rating criteria, the formula is used:

[0046]

[0047] Get the determined warning level PM, where CM i is the severity score of a single anomaly, VM i is the risk weight coefficient of the abnormal type, n m is the total number of exception types;

[0048] Based on the determined warning level, an alarm process of a corresponding level is automatically activated to obtain a warning response result.

[0049] Compared with the prior art, the advantages and positive effects of the present invention are:

[0050] In the present invention, the cloud parallel processing technology is used to accurately synchronize the video stream timestamp and image quality, thereby ensuring the high quality of the data stream. The screening mechanism ensures that only the video stream that meets the quality standards is further analyzed, thereby enhancing the accuracy of subsequent abnormal behavior analysis. The system can not only analyze in real time and quickly identify abnormal behaviors that deviate from normal trajectories, but also dynamically adjust computing and storage resources according to actual resource usage, thereby improving processing speed and cost efficiency, and enabling the system to flexibly respond to monitoring needs of different scales, which is particularly suitable for dynamic environments and effectively improves the early warning and response capabilities of video surveillance. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a system flow chart of the present invention;

[0052] Figure 2 This is a flow chart of synchronizing video stream timestamps in the present invention;

[0053] Figure 3 This is a flow chart of obtaining a video quality standard stream in the present invention;

[0054] Figure 4 A flow chart for identifying abnormal behaviors that deviate from normal trajectories in the present invention;

[0055] Figure 5 A flowchart for obtaining abnormal behavior records in the present invention;

[0056] Figure 6 This is a flow chart for evaluating the efficiency and load conditions used in the present invention;

[0057] Figure 7 A flowchart for obtaining resource allocation configuration in the present invention;

[0058] Figure 8 This is a flow chart for obtaining early warning response results in the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0060] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0061] Example

[0062] See also Figure 1 The present invention provides a technical solution: a video surveillance data processing system based on cloud computing includes:

[0063] The video quality synchronization module is based on multi-source video data. It synchronizes the video stream timestamp through cloud parallel processing technology, analyzes the clarity and signal integrity of each frame, verifies the quality of the video stream, screens the video stream that meets the quality standards, and obtains the video quality standard stream; the quality standard indicators are specific: image clarity, signal integrity, timestamp synchronization, frame rate and color accuracy.

[0064] The abnormal behavior recognition module analyzes the movement trajectories of pedestrians and vehicles in the video in real time based on the video quality standard stream, identifies abnormal behaviors that deviate from the normal trajectory, and compares and analyzes them with the standard behaviors to generate abnormal behavior records; the scenes of normal trajectories may include urban traffic, public places, or schools and office buildings. Specifically, in urban road traffic monitoring, normal vehicle trajectories include vehicles driving along the road, obeying traffic lights, and changing lanes correctly within the lane. Abnormal behaviors include driving in the opposite direction, sudden emergency stops, or turning left at intersections where left turns are prohibited. In public places, normal pedestrian trajectories include walking along the aisles of shopping malls and entering and exiting stores. Abnormal behaviors may be attempts to enter during closed hours or frequent entry and exit of the same area in a very short period of time, which may be suspected to be signs of theft. In school and office building environments, normal trajectories may be in and out traffic during rush hour, as well as regular rest activities during work or study time. Abnormal behaviors include wandering around during non-working hours or staying in unauthorized areas.

[0065] The resource scheduling optimization module evaluates the utilization efficiency and load of current cloud platform resources based on abnormal behavior records, dynamically adjusts computing resources and storage allocation according to the analysis results and predictions of future demand, and generates resource allocation configurations;

[0066] The security warning trigger module analyzes the severity and urgency of abnormal behaviors based on resource allocation configuration, formulates security warning processes according to risk rating standards, and automatically triggers alarms to generate warning response results. The risk rating standards can be low, medium, and high risks. Low risk, such as pedestrians jogging at night or staying in the park, the response is only monitoring; medium risk, such as unknown persons wandering around the office building after office hours, the response is to send notifications to relevant personnel; high risk, such as unknown persons trying to enter private places multiple times, the response is to notify relevant personnel and alarm.

[0067] The video quality standard stream includes clarity level, color accuracy, and signal stability. The abnormal behavior record includes behavior category, occurrence time, and location information. The resource allocation configuration includes processing power, storage capacity, and network bandwidth. The early warning response results include alarm type and reaction speed.

[0068] See also Figure 2 , the steps to synchronize the video stream timestamp are as follows:

[0069] Based on multi-source video data, original timestamps are collected and stored in the cloud library to obtain the original timestamp set;

[0070] Collecting original timestamps from multi-source video data and storing them in the cloud library involves receiving the time tags of each video source and temporarily storing them in chronological order. The time tags are obtained through the built-in clock of each video device to ensure that the timestamp of each video stream is accurately recorded. The set reflects the time data of different video sources at the same moment. The generation of the original timestamp set is obtained by standardizing the time data of each video source to ensure that all data formats are consistent for subsequent processing.

[0071] Select a reference timestamp from the original timestamp set, calculate the deviation between the timestamps of multiple video streams and the reference, and obtain a time deviation set;

[0072] A reference timestamp is selected from the original timestamp set, and the deviation between the timestamp of each video stream and the reference is calculated. The earliest timestamp is usually used as the reference to ensure smooth time synchronization. The time deviation is calculated through the difference. The process involves a direct comparison of each timestamp with the reference timestamp to calculate the time delay or advance of each video source. The deviation will be used to adjust the timestamp of each video stream to ensure the accuracy of synchronization and obtain a time deviation set.

[0073] Based on the time deviation set, the average deviation of the video source timestamp is calculated using the formula:

[0074]

[0075] Get the synchronized video stream timestamp t′ sync, where t base is the reference timestamp, Δt i is the deviation between the i-th video source and the reference time, and n is the number of video sources;

[0076] If the following parameters are given, set the base timestamp t base The time span is 1500 seconds. There are three video sources with time deviations of -10 seconds, 5 seconds, and 3 seconds respectively. Substitute them into the formula for calculation:

[0077] Δt1=-10, Δt2=5, Δt3=3

[0078] Calculate the mean deviation:

[0079]

[0080] The synchronized timestamps are:

[0081] t′ sync =1500-0.67=1499.33

[0082] This result shows that a timestamp that takes into account the deviations of all video sources can be obtained and used for all video streams to ensure that display and analysis are based on the same time basis.

[0083] See also Figure 3 , the specific steps for obtaining the video quality standard stream are:

[0084] Analyze the clarity and signal integrity of each frame of the image, evaluate the image clarity and data loss, assign corresponding clarity and integrity indexes to each frame of the image, and obtain an image quality score;

[0085] Analyze the clarity and signal integrity of each frame of image, and conduct in-depth analysis of each frame of image. Clarity detection involves the use of edge detection algorithm to evaluate the clarity and definition of image edges. At the same time, signal integrity is assessed by detecting the color saturation and pixel integrity of the image, ensuring a comprehensive evaluation of the video frames so that the quality of each frame of image can be reliably quantified. The evaluation results assign clarity and integrity indexes to each frame of image, which reflect the visual clarity of the image and the complete transmission quality of the signal respectively.

[0086] Based on the image quality score, the formula is used:

[0087] QS=ws1×CS+ws2×IS

[0088] The quality of each frame of the image is evaluated, and the video frames exceeding the threshold are screened to obtain the video quality standard stream, where QS represents the quality score of each frame of video, CS represents clarity, IS represents signal integrity, and ws1 and ws2 are weight parameters used to adjust the influence of clarity CS and signal integrity IS on the total quality score QS;

[0089] There is a video frame with a clarity score CS of 80, a signal integrity index IS of 95, weight parameters ws1 set to 0.6, ws2 set to 0.4, and the scores are calculated by substituting into the formula:

[0090] QS=0.6×80+0.4×95=48+38=86

[0091] The result shows that the overall quality score of the video frame is 86, which means that it performs well in terms of clarity and signal integrity and is suitable for inclusion in a video quality standard stream.

[0092] See also Figure 4 ,The steps to identify abnormal behaviors that deviate from the normal trajectory are as follows:

[0093] Based on the video quality standard stream, the pedestrians and vehicles in the video are tracked in real time, the motion trajectory of each pedestrian and vehicle is identified, and preliminary motion trajectory data is obtained;

[0094] Based on the video quality standard stream, pedestrians and vehicle targets in the video are tracked in real time. Image recognition and tracking technology are used to identify the position of each pedestrian and vehicle. As the video stream continues to input, the system dynamically updates the target position to generate a continuous motion trajectory of the target. The trajectory data is a record of the movement path of each target in the video. The process includes detecting the starting position of the target, tracking its movement, and recording the position at each time point. The position data will serve as the basis for further analysis to ensure that each target is accurately tracked.

[0095] Based on the preliminary motion trajectory data, the deviation between the current trajectory and the standard trajectory is calculated using the formula:

[0096]

[0097] Get the deviation score ΔSV for each target, where x act,t and act,t Represents the coordinate point currently monitored, x norm,t and norm,t represents the coordinate point of the standard trajectory, T represents the total number of time steps or frames in the detection cycle;

[0098] If we have the following data, the standard trajectory coordinates are (100, 200) and the observed coordinates are (105, 205), and the time step T is 1, substitute the specific values ​​to calculate the deviation score:

[0099]

[0100] This result shows that the target deviates from the standard trajectory by 7.07, which can help determine whether to trigger an abnormal behavior alarm.

[0101] Based on the deviation score of each target, evaluate whether the behavior is abnormal, mark the target location and time, and obtain the abnormal behavior identification result;

[0102] Based on the deviation score of each target, the behavior is evaluated whether it is abnormal. If the deviation score exceeds the preset threshold, the system will mark the target as having abnormal behavior. The evaluation process is carried out by comparing the calculated deviation score with the set threshold. Exceeding the threshold means that the target's behavior deviates from the normal mode. The information is used to mark the target location and time to ensure that relevant security personnel can respond quickly when necessary, thereby improving the response efficiency and safety of monitoring.

[0103] See also Figure 5 , the specific steps for obtaining abnormal behavior records are:

[0104] Based on abnormal behavior, a preliminary comparative analysis of the data was conducted using the formula:

[0105]

[0106] Calculate the abnormal behavior index EV, where δX pos,i is the position deviation of target i from the standard trajectory, α x is the adjustment coefficient, which affects the sensitivity of abnormality determination, n x is the total number of targets;

[0107] There are three targets with position deviation δX pos,i are 2, 4 and 6 respectively, and the adjustment coefficient α x Set it to 0.5 and calculate:

[0108] EV=0.5·(2 2 +4 2 +6 2 )=0.5·(4+16+36)=0.5·56=28

[0109] The results showed that the cumulative abnormal behavior index was 28, which means that the monitoring targets in this area showed high behavioral deviations during a given period and required further attention or direct intervention.

[0110] Based on the abnormal behavior index, the location, timestamp and deviation degree of each target are recorded and formatted as abnormal behavior records;

[0111] Based on the abnormal behavior index, the location, timestamp and degree of deviation of each target are recorded. The records are formatted as structured data and input into the security monitoring system for further behavior analysis and decision support. In this way, the system can effectively track and manage potential risks and ensure timely response to security threats.

[0112] See also Figure 6 , the evaluation steps of usage efficiency and load conditions are as follows:

[0113] Based on abnormal behavior records, cloud platform resource data is collected, including CPU usage, memory usage, storage read and write rates, and network bandwidth usage, to obtain a resource usage data set;

[0114] Based on abnormal behavior records, cloud platform resource data is collected, including monitoring and data collection of various key resources in the cloud platform such as CPU usage, memory usage, storage read and write rates, and network bandwidth, and real-time capture of the usage of each resource. The data includes not only resource usage rate, but also records of peak usage periods and minimum usage periods. The collected data sets are stored in a structured manner for further analysis and evaluation.

[0115] Analyze resource usage data sets and calculate resource usage efficiency and load conditions using the formula:

[0116]

[0117] According to the different resource types, the average cloud platform utilization EU is calculated, where UO i represents the current utilization rate of the i-th resource, WO i is the weight coefficient of the i-th resource, n o is the total number of resource types;

[0118] The cloud platform has three resources, CPU (weight 0.5, usage 70%), memory (weight 0.3, usage 80%), and storage (weight 0.2, usage 60%). Substitute them into the formula to calculate the average usage of the cloud platform:

[0119]

[0120] This result shows that the average resource utilization efficiency of the cloud platform is 23.7%, which reflects the adaptability of the current resource configuration relative to the actual demand and guides the resource optimization and adjustment strategy.

[0121] See also Figure 7 , the specific steps for obtaining resource allocation configuration are:

[0122] Based on the analysis results, determine the current usage status and future demand of resources (computing, storage, network bandwidth) using the formula:

[0123] RG future =RG current ×(1+β g ×G y )

[0124] Predict future resource requirementsRG future , where RG current is the current resource usage, β g is the predicted business growth rate, G y is the coefficient of the impact of business growth on resource demand. Business refers to database, computing, storage, and network services related to video surveillance data processing.

[0125] Based on the analysis results, the current resources are verified, such as analyzing the historical data of CPU load, storage capacity usage trends, and the peak and valley of network traffic. The data is collected and sorted through the comprehensive monitoring of the cloud management platform to ensure the real-time and accuracy of the data. If the current resource usage is RG current is 100, the predicted business growth rate β g is 0.10 (10%), the impact coefficient of business growth on resource demand G y is 1.5, substitute it into the formula for calculation:

[0126] RG future =100×(1+0.10×1.5)=100×(1+0.15)

[0127] RG future =100×1.15=115

[0128] This result indicates that resource requirements will increase to 115 in the future, indicating the need to enhance resource allocation to cope with the expected business growth.

[0129] Based on future resource requirements, dynamically adjust computing resources and storage allocation to ensure that the configuration of each resource meets the predicted business needs to the greatest extent possible and obtain resource allocation configuration;

[0130] Based on future resource demands, dynamically adjust computing resources and storage allocation, and adjust resource allocation according to forecast results to ensure that the configuration of each resource can meet future business needs to the greatest extent possible, including increasing computing power, expanding storage capacity, and optimizing network architecture. Resource adjustments are made based on a comprehensive analysis of current resource utilization and future demand forecasts to obtain resource allocation configuration.

[0131] See also Figure 8, the specific steps for obtaining the early warning response results are:

[0132] Based on resource allocation configuration, detect and record each abnormal behavior, evaluate the impact based on the abnormal behavior type, frequency and impact on resource usage, and obtain a temporary severity score;

[0133] Based on the resource allocation configuration, a monitoring model is established to detect and record each abnormal behavior. The monitoring model relies on predefined resource usage indicators and behavior recognition algorithms to set baselines for resource usage, such as CPU usage, memory usage, and network traffic. Indicators are collected through real-time monitoring. By comparing real-time data with the baseline, it is determined when deviations occur. Whenever the system detects that an indicator anomaly exceeds the threshold, it automatically records detailed information about the event, including time, type, and resource usage. Then, these abnormal events are classified according to the type of anomaly, such as abnormal access frequency or sudden increase in resource usage. Each type of event is weighted according to its frequency of occurrence and resource usage impact. Each event is assigned a temporary severity score, which reflects the potential impact of the event on the overall performance of the system.

[0134] Based on the provisional severity score, compared with the risk rating criteria, the formula is used:

[0135]

[0136] Get the determined warning level PM, where CM i is the severity score of a single anomaly, VM i is the risk weight coefficient of the abnormal type, n m is the total number of exception types;

[0137] If there are three types of anomalies, their severity scores are 10, 20 and 30 respectively, and the corresponding risk weight coefficients are 1, 0.5 and 2 respectively. Substitute them into the formula for calculation:

[0138]

[0139] The result shows that the average warning level is 22.86, which means that the system needs to adjust the security warning and response strategy according to this level.

[0140] Based on the determined warning level, the corresponding level of alarm process is automatically activated to obtain the warning response result, including the alarm type, time, and response measures information;

[0141] Based on the determined warning level, the alarm process of the corresponding level is automatically activated, including configuring the alarm system to respond to the determined warning level, automatically selecting the alarm type and response measures according to the severity of the warning level, and ensuring that each alarm can be effectively handled as soon as possible. In this way, a quick response can be made to security threats, thereby improving overall security. The generated warning response results will record in detail the type, time and specific measures taken of the alarm.

[0142] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A video surveillance data processing system based on cloud computing, characterized in that: The system comprises: The video quality synchronization module is based on multi-source video data. It synchronizes the video stream timestamp through cloud parallel processing technology, analyzes the clarity and signal integrity of each frame, verifies the quality of the video stream, and selects the video stream that meets the quality standards to obtain the video quality standard stream; The abnormal behavior recognition module performs real-time analysis on the movement trajectories of pedestrians and vehicles in the video based on the video quality standard stream, identifies abnormal behaviors that deviate from normal trajectories, and compares and analyzes them with standard behaviors to generate abnormal behavior records; The resource scheduling optimization module evaluates the utilization efficiency and load of the current cloud platform resources based on the abnormal behavior records, dynamically adjusts the computing resources and storage allocation according to the analysis results and the prediction of future demand, and generates resource allocation configuration; The security warning trigger module analyzes the severity and urgency of abnormal behavior based on the resource allocation configuration, formulates a security warning process according to the risk rating standard, and automatically triggers an alarm to generate a warning response result.

2. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The steps of synchronizing the video stream timestamp are specifically as follows: Based on multi-source video data, original timestamps are collected and stored in the cloud library to obtain the original timestamp set; Selecting a reference timestamp from the original timestamp set, calculating the deviation between the timestamps of multiple video streams and the reference timestamp, and obtaining a time deviation set; Based on the time deviation set, the average deviation of the video source timestamp is calculated using the formula: Get the synchronized video stream timestamp t′ sync , where t base is the reference timestamp, Δt i is the deviation between the i-th video source and the reference time, and n is the number of video sources.

3. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The steps for obtaining the video quality standard stream are specifically as follows: Analyzing the clarity and signal integrity of each frame of the image, evaluating image clarity and data loss, and obtaining an image quality score; Based on the image quality score, the formula is used: QS=ws1×CS+ws2×IS The quality of each frame of the image is evaluated, and the video frames exceeding the threshold are screened to obtain the video quality standard stream, where QS represents the quality score of each frame of video, CS represents clarity, IS represents signal integrity, and ws1 and ws2 are weight parameters.

4. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The steps of identifying abnormal behaviors that deviate from the normal trajectory are specifically as follows: Based on the video quality standard stream, pedestrians and vehicle targets in the video are tracked in real time, the motion trajectory of each pedestrian and vehicle is identified, and preliminary motion trajectory data is obtained; Based on the preliminary motion trajectory data, the deviation between the current trajectory and the standard trajectory is calculated using the formula: Get the deviation score ΔSV for each target, where x act,t and act,t Represents the coordinate point currently monitored, x norm,t and norm,t represents the coordinate point of the standard trajectory, T represents the total number of time steps or frames in the detection cycle; Based on the deviation score of each target, whether the behavior is abnormal is evaluated, and the target position and time are marked to obtain an abnormal behavior recognition result.

5. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The steps for obtaining the abnormal behavior record are specifically as follows: Based on the abnormal behavior, a preliminary comparative analysis of the data was conducted using the formula: Calculate the abnormal behavior index EV, where δX pos,i is the position deviation of target i from the standard trajectory, α x is the adjustment factor, n x is the total number of targets; Based on the abnormal behavior index, the location, timestamp and degree of deviation of each target are recorded and formatted as an abnormal behavior record.

6. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The evaluation steps of the utilization efficiency and load condition are specifically as follows: Based on the abnormal behavior records, cloud platform resource data is collected, including CPU usage, memory usage, storage read and write rates, and network bandwidth usage, to obtain a resource usage data set; Analyze the resource usage data set, calculate resource usage efficiency and load, and use the formula: According to the different resource types, the average cloud platform utilization EU is calculated, where UO i represents the current utilization rate of the i-th resource, WO i is the weight coefficient of the i-th resource, n o The total number of resource types.

7. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The steps for obtaining the resource allocation configuration are specifically as follows: Based on the analysis results, determine the current usage status and future demand of resources using the formula: RG future =RG current ×(1+β g ×G y ) Predict future resource requirementsRG future , where RG current is the current resource usage, β g is the predicted business growth rate, G y is the coefficient of impact of business growth on resource demand; Based on the future resource demand, computing resources and storage allocation are dynamically adjusted to obtain resource allocation configuration.

8. The video surveillance data processing system based on cloud computing according to claim 1, characterized in that: The steps for obtaining the early warning response result are specifically as follows: Based on the resource allocation configuration, each abnormal behavior is detected and recorded, and the impact is evaluated according to the type, frequency and impact on resource usage of the abnormal behavior to obtain a temporary severity score; Based on the provisional severity score, compared with the risk rating criteria, the formula is used: Get the determined warning level PM, where CM i is the severity score of a single anomaly, VM i is the risk weight coefficient of the abnormal type, n m is the total number of exception types; Based on the determined warning level, an alarm process of a corresponding level is automatically activated to obtain a warning response result.

Images