Dynamic video monitoring resource management system

Abstract

The invention relates to the technical field of video monitoring resource management, in particular to a dynamic video monitoring resource management system, which comprises a task emergency degree calculation module, a video storage resource measurement and calculation module, a video storage distribution and optimization module, a monitoring target residence analysis module and a video storage hierarchy adjustment module. According to the method, the task emergency degree is calculated through the event triggering information, the task triggering source, the event category and the coverage radius, accurate task priority division is achieved, the storage resource distribution proportion is dynamically adjusted, the high-priority task obtains better storage guarantee, the storage requirement of the high-emergency-degree task is calculated, the storage structure is optimized in combination with the data fragmentation rate, and the storage efficiency is improved. The storage space utilization rate is improved, the storage retention duration is adjusted according to the target entry frequency, low-priority task storage compression is executed in combination with the load state of the storage server, efficient release of the storage space is achieved, storage redundancy is reduced, and the storage stability and the task scheduling efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of video surveillance resource management, and in particular to a dynamic video surveillance resource management system. Background Art

[0002] The field of video surveillance resource management technology includes the core contents of the collection, storage, scheduling, retrieval and management of surveillance video data, involving the organization of massive video data, index structure, access mechanism and storage optimization strategy to improve the data utilization efficiency and management capabilities of the surveillance system. It also covers data access permission control, storage medium optimization configuration, distributed storage architecture, data flow scheduling strategy and data consistency management to ensure the stable storage, efficient transmission and fast call of surveillance data. With the growth of video surveillance demand, it is gradually developing towards automated and intelligent management, making the use of surveillance resources more efficient and accurate.

[0003] Among them, the dynamic video surveillance resource management system refers to a system that can dynamically allocate, schedule and optimize the management of video surveillance resources. In view of the dynamic change characteristics of monitoring data, a storage management strategy based on real-time video data analysis is adopted, and video data is hierarchically archived through data labeling processing to optimize the storage structure. In view of the access requirements of video resources, the time series feature analysis method is used to retrieve and schedule video data, and the resource positioning accuracy is improved through the regional video content indexing method. For data storage optimization, a hierarchical storage management mechanism is adopted to automatically classify and store video data according to factors such as time, frequency, and importance. In view of data consistency issues, a data synchronization control mechanism is used to coordinate resources in a distributed storage environment to ensure data integrity and consistency when accessed by multiple nodes.

[0004] In the existing video surveillance resource management process, there is a problem of rigid storage strategy in the management of surveillance video data. The storage priority division method is relatively simple and fails to be adjusted in real time according to the urgency of the task, resulting in a lack of pertinence in the allocation of storage resources, which affects the execution efficiency of the surveillance task. In the storage management process, the storage structure is not dynamically optimized in combination with the data fragmentation rate. After long-term operation, storage fragments accumulate, affecting the data retrieval efficiency and the overall performance of the storage system. The target storage management lacks in-depth analysis of the activity characteristics of the monitored object, and only relies on static rules for data storage. It fails to fully utilize parameters such as target residence time, entry frequency, and departure interval to optimize the storage level, resulting in low-frequency target data occupying storage resources, while high-frequency target data may be difficult to be stored preferentially due to the rigid storage strategy. The storage level adjustment depends on the fixed storage duration and lacks an adaptation mechanism for the storage load status. It is difficult to reasonably release storage resources under high load conditions, affecting the long-term operation stability and data management efficiency of the system. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a dynamic video monitoring resource management system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A dynamic video monitoring resource management system comprises: The task urgency calculation module obtains the event triggering information of the video monitoring task, calculates the monitoring task coverage radius, analyzes the monitoring task urgency, and obtains the video monitoring task urgency analysis result; The video storage resource calculation module reads the storage space usage rate of the storage server based on the urgency analysis result of the video surveillance task, calculates the storage proportion that can be allocated to each video task, and obtains the video task storage proportion allocation result; The video storage allocation and optimization module calculates the video storage requirements of high-urgency tasks based on the video task storage weight allocation result, performs video data block merging, and obtains an optimized video task storage allocation state; The monitoring target resident analysis module calculates the target video recording duration in the monitoring area, the target entry frequency, and the target leaving time interval based on the optimized video task storage allocation state, determines the video storage level requirement, and obtains the video monitoring target storage level requirement analysis result.

[0007] As a further solution of the present invention, the video surveillance task urgency analysis result includes the task trigger source category, event severity, monitoring task coverage radius, and monitoring task priority; the video task storage proportion allocation result includes the high priority task storage proportion, medium priority task storage proportion, low priority task storage proportion, and storage server storage load proportion; the video task storage allocation status includes the high urgency task storage allocation status, storage server data fragmentation rate, and video data block optimization status; the video surveillance target storage level requirement analysis result includes the target video recording duration, target entry frequency, target departure time interval, and storage level requirement level.

[0008] As a further solution of the present invention, the task urgency calculation module includes: The event trigger information acquisition submodule acquires the event trigger information of the video surveillance task, reads the task trigger source and event category, analyzes the time and space characteristics of the event category, extracts the correlation information between the event location, event time, event type and task source, and obtains the event correlation information parameter set; The monitoring task coverage radius calculation submodule analyzes and determines the coverage required for the monitoring task based on the event association information parameter set, using the formula: ; Calculate the monitoring task coverage radius value ,in, Represents the area affected by the event, Represents the speed of event propagation, Represents the time required for response; The monitoring task urgency analysis submodule calculates the urgency index of the current monitoring task according to the monitoring task coverage radius value, combined with the event type risk coefficient, event occurrence density, and historical urgency distribution, and obtains the video monitoring task urgency analysis result.

[0009] As a further solution of the present invention, the video storage resource calculation module includes The storage resource acquisition submodule obtains the available storage capacity of all current video storage servers, reads the storage space usage rate of the storage servers, counts the total number of tasks in the current video task queue, extracts the storage resource status data of each server, and obtains the storage resource status parameter set; The task priority calculation submodule is based on the storage resource status parameter set and the video surveillance task urgency analysis result, and divides the tasks into high, medium and low priority tasks according to the task urgency, and counts the number of high, medium and low priority tasks, using the formula ; Calculate the The proportion of tasks , integrate to get the proportion data of each priority task, among which, Representative The number of class tasks, Represents the total amount of high, medium, and low priority tasks; The storage proportion allocation submodule adjusts the video storage resource allocation proportion based on the priority task proportion data, calculates the storage proportion that can be allocated to each video task, and obtains the video task storage proportion allocation result.

[0010] As a further solution of the present invention, the video storage allocation and optimization module includes The high-urgency task storage requirement calculation submodule extracts the storage allocation data corresponding to the high-urgency task in combination with the video task storage weight distribution result, calculates the video storage requirement of the high-urgency task, and obtains the high-urgency task storage requirement data; The video data fragmentation rate calculation submodule obtains the existing video data storage status in the storage server based on the high-urgency task storage demand data, extracts the video data occupancy of each storage unit, and adopts the formula: ; Calculate video data fragmentation rate , get the data fragmentation rate data of the storage server, where, Representative The amount of video data stored in the storage unit, Representative The free storage capacity of the storage unit, Represents the total number of storage cells that store data, Represents the total number of storage units of the current server; The video data block merging submodule performs video data block merging based on the data fragmentation rate data of the storage server, adjusts the arrangement of fragmented data blocks and the video storage structure in the storage unit, and obtains an optimized video task storage allocation state.

[0011] As a further solution of the present invention, the monitoring target resident analysis module includes: The monitoring target residence time calculation submodule obtains the entry timestamp of the monitoring target according to the optimized video task storage allocation state, reads the target exit timestamp, calculates the video recording time of the target in the monitoring area, screens the corresponding relationship between continuous entry and exit records, calculates the residence time of different targets in different time periods, summarizes the residence time distribution of targets in the monitoring area, and obtains the monitoring target residence time data; The target entry frequency statistics submodule counts the number of target entries based on the residence time data of the monitored target, analyzes the activity of different targets in the monitored area, calculates the target entry frequency based on the observation period, analyzes the time distribution characteristics of the target entering the area, and obtains the target entry frequency data; The storage level demand judgment submodule calculates the target's departure time interval based on the target entry frequency data, counts the retention periods of different targets, combines the target's video residence time, entry frequency, and departure time interval in the monitoring area, compares the storage priorities of different targets, and adjusts the video storage level according to the storage demand classification to obtain the video surveillance target storage level demand analysis results.

[0012] As a further solution of the present invention, the system further includes a video storage level adjustment module; The video storage level adjustment module adjusts the video storage retention time based on the storage level demand analysis result of the video surveillance target according to the target's access frequency, and performs storage compression on low-priority video tasks in combination with the storage server load status to release storage space and obtain the video storage level adjustment result; The video storage level adjustment result includes the video storage retention time adjustment, storage server load balancing status, low priority task storage compression ratio, and storage space release amount.

[0013] As a further solution of the present invention, the video storage level adjustment module includes: The target entry frequency analysis submodule combines the storage level demand analysis results of the video surveillance target to obtain the entry records of the surveillance target, count the number of entries of the target within a specified time period, calculate the entry frequency values ​​of different targets, filter the targets whose entry frequency is higher than the set threshold, and calculate the entry frequency distribution interval to obtain the target entry frequency distribution data; The storage retention time adjustment submodule sets a storage retention time adjustment factor for targets in different frequency intervals based on the target entry frequency distribution data, calculates the storage time adjustment value corresponding to each target, and adjusts the storage retention time to obtain storage retention time adjustment data; The storage server load monitoring submodule monitors the current load of the storage server, collects storage occupancy, access rate and storage pressure value, calculates the storage occupancy, determines whether the load status exceeds the load threshold, and obtains the storage server load status data; The low-priority video storage compression submodule screens low-priority video data based on the storage retention time adjustment data and the storage server load status data, using the formula: ; Calculate the compression ratio of low priority video data storage , get the video storage level adjustment result; where, Representative Storage size of video files, Represents the remaining storage time of the video. Represents the target entry frequency of the video, Represents the load ratio of the server where the video is located. Represents the bandwidth usage of the video storage server. Represents the average data flow during the video storage period, Represents the current bandwidth consumption of video storage, Represents the frequency of video access. Represents the number of compression histories of the video. The total number of low priority videos.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by obtaining the event trigger information of the video monitoring task, and combining the task trigger source, event category and monitoring task coverage radius, a dynamic priority evaluation mechanism is formed to make the classification of monitoring tasks more accurate and ensure that resource allocation meets actual needs. Combined with the available storage capacity, storage space utilization rate and total amount of task queues of the storage server, based on the dynamic calculation method of task urgency, the task storage priority is divided, and the resource allocation ratio is adjusted, so that the storage ratio of high-priority tasks is optimized, and the storage guarantee of key video data is improved. The storage requirements of high-urgency tasks are calculated, and the storage optimization direction is analyzed in combination with the data fragmentation rate of the storage server, and the video data blocks are intelligently merged to reduce the storage fragmentation rate and improve data access efficiency. Based on the entry timestamp and exit timestamp of the monitoring target, the target residence time, entry frequency and exit interval are calculated, so that the storage level demand division is more accurate, and the video data of important targets can be stored at the appropriate level. The storage retention time is dynamically adjusted according to the target entry frequency, and the storage compression of low-priority tasks is performed in combination with the storage server load status, so that the storage resource release is more reasonable, while optimizing the storage space utilization, the system load is reduced, and the system storage stability and scheduling flexibility are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a system flow chart of the present invention; Figure 2 This is a flow chart of the task urgency calculation module of the present invention; Figure 3 This is a flow chart of the video storage resource calculation module of the present invention; Figure 4 The video storage allocation and optimization module flow chart of the present invention; Figure 5 This is a flow chart of the monitoring target resident analysis module of the present invention; Figure 6 This is a flow chart of the video storage level adjustment module of the present invention. DETAILED DESCRIPTION

[0016] 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.

[0017] 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" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the 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, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0018] See also Figure 1 , a dynamic video surveillance resource management system comprises: The task urgency calculation module obtains the event triggering information of the video surveillance task, reads the task triggering source and event category, calculates the monitoring task coverage radius, analyzes the monitoring task urgency, and obtains the video surveillance task urgency analysis result; The video storage resource calculation module obtains the available storage capacity of all current video storage servers based on the urgency analysis results of video surveillance tasks, reads the storage space usage rate of the storage servers, counts the total number of tasks in the current video task queue, divides high, medium and low priority tasks according to the urgency of the tasks, calculates the proportion of high, medium and low priority tasks, adjusts the allocation proportion of video storage resources, calculates the storage proportion that can be allocated to each video task, and obtains the video task storage proportion allocation result; The video storage allocation and optimization module calculates the video storage requirements of high-urgency tasks based on the video task storage weight allocation results, calculates the video data fragmentation rate of the storage server, performs video data block merging, and obtains the optimized video task storage allocation status; The monitoring target residence analysis module obtains the entry timestamp of the monitoring target, reads the target exit timestamp, calculates the video recording time of the target in the monitoring area, counts the target entry frequency, calculates the target departure time interval, and determines the video storage level requirement based on the target video residence time, entry frequency, and departure time interval in the monitoring area, and obtains the video monitoring target storage level requirement analysis result; The video storage level adjustment module is based on the storage level demand analysis results of the video surveillance target, adjusts the video storage retention time according to the target's access frequency, and performs storage compression on low-priority video tasks in combination with the storage server load status to release storage space and obtain the video storage level adjustment results.

[0019] The results of the video surveillance task urgency analysis include the task trigger source category, event severity, monitoring task coverage radius, and monitoring task priority. The video task storage proportion allocation results include the high-priority task storage proportion, medium-priority task storage proportion, low-priority task storage proportion, and storage server storage load proportion. The video task storage allocation status includes the high-urgency task storage allocation status, storage server data fragmentation rate, and video data block optimization status. The results of the video surveillance target storage level requirement analysis include the target video recording duration, target entry frequency, target departure time interval, and storage level requirement level. The video storage level adjustment results include the video storage retention time adjustment status, storage server load balancing status, low-priority task storage compression ratio, and storage space released.

[0020] See also Figure 2 ,The task urgency calculation module includes: The event trigger information acquisition submodule acquires the event trigger information of the video surveillance task, reads the task trigger source and event category, analyzes the time and space characteristics of the event category, extracts the correlation information between the event location, event time, event type and task source, and obtains the event correlation information parameter set; The event trigger information of video surveillance tasks comes from multiple data sources, including traffic monitoring cameras, public security cameras, intelligent sensor devices, etc. First, it is necessary to collect monitoring data from these devices, parse the data format, and extract key information, such as target recognition results in image frames, timestamp data in video streams, and geographic location information of sensors. Assuming that a monitoring task involves urban road traffic events, the system will extract image data containing the area, time and event category of the event from the traffic monitoring camera, and calculate the occurrence time of the event based on the timestamp information. After parsing the data, the system matches the corresponding preset category according to the event type (such as vehicle collision, pedestrian running a red light, traffic jam, etc.), and analyzes the source of the task trigger, such as whether the event is triggered by an automatic monitoring system, or by data input provided by a manual alarm or other external system, and further parses the time characteristics of the event category, such as some events have For persistence, such as traffic congestion, its duration can be calculated from historical data, while the characteristics of sudden events such as vehicle collisions may be limited to mutation phenomena within a short period of time. In terms of spatial characteristic analysis, the system will estimate the impact area of ​​the event based on the GPS coordinates, camera angles and event range provided by the sensor equipment. Assuming that an accident occurs at a road intersection, the system needs to calculate the number of lanes affected by the accident and the scope of congestion based on the camera angle and traffic flow data. In order to ensure the relevance of the data, the system compares multi-source data when extracting the correlation information between the event location, event time, event type and task source. For example, the accident report provided by the traffic police system can be time-matched with the video surveillance record. If the error in the accident time recorded by the two is less than 1 second, it can be identified as the same event. As shown in Table 1, the time error range of different types of events is listed, and finally the event correlation information parameter set is obtained.

[0021] Table 1 Event type and time error range As shown in Table 1, different event types have different time error ranges, and these error values ​​affect the matching degree of event data and ensure the accuracy of event triggering information.

[0022] The monitoring task coverage radius calculation submodule analyzes and determines the coverage required for the monitoring task based on the event association information parameter set, using the formula: ; Calculate the monitoring task coverage radius value ,in, Represents the area affected by the event, Represents the speed of event propagation, Represents the time required for response; Based on the event correlation information parameter set, it is necessary to calculate the coverage required for the monitoring task. The calculation involves multiple key parameters, including the area of ​​the event impact area, the speed of event propagation, and the time required for response. First, the system determines the area of ​​the event impact area. If the event is a traffic jam, the system can calculate its impact area through historical traffic flow data. For example, a one-way congestion on a main road may affect traffic within a 1-kilometer range. It can be approximated as 1000 meters × 20 meters (road width) = 20,000 square meters, and the event propagation speed It can be determined according to different event types. For example, the impact propagation speed at the accident scene can be estimated by monitoring the vehicle moving speed. Assuming that an accident causes the vehicle speed to drop to 10km / h, and the normal speed is 60km / h, it can be calculated that the impact propagation speed is about 50km / h (i.e. 13.89m / s), and the response time is It can be calculated from historical response data. For example, the average response time of traffic police in handling traffic accidents is 5 minutes (300 seconds). Substituting the above data into the formula: ; Perform the calculation: ; Finally, the coverage radius of the monitoring task is obtained. The results show that the minimum coverage radius of the monitoring task should be set to about 88 meters to ensure comprehensive monitoring of the accident affected area.

[0023] The monitoring task urgency analysis submodule calculates the urgency index of the current monitoring task based on the monitoring task coverage radius value, combined with the event type risk coefficient, event occurrence density, and historical urgency distribution, and obtains the video monitoring task urgency analysis result; Call the monitoring task coverage radius value, combine the event type risk coefficient, event occurrence density, and historical urgency distribution to calculate the urgency index of the current monitoring task and the event type risk coefficient. The value is assigned by analyzing historical data. For example, the risk factor of a traffic accident may be set to 1.5, while that of a normal traffic jam is set to 1.0. The value is determined based on the impact of the type of accident on road safety. For example, traffic accidents are usually accompanied by casualties and road congestion, and the risk is significantly higher than that of normal congestion events. Therefore, the value is assigned to be higher than 1.0. If the accident involves a multi-vehicle chain collision or a complete closure of a main road, the risk factor may be further increased to 2.0 or above. The frequency of the event is determined by the risk factor. It can be calculated by the number of events per unit time. If there are 3 traffic accidents in a certain area every hour, the density is 3 events / hour. The value comes from the actual monitoring data of the area in the past 24 hours. If the frequency of accidents increases in a short period of time, the value will also increase synchronously. For example, if there are 3 accidents in 30 minutes, the density may be adjusted to 6 events / hour. The historical urgency distribution Provided by historical monitoring data, for example, in the past week, the average urgency index of the region is 2.3. The value is calculated by arithmetic average of the urgency index of different time periods, and the urgency index of different time periods is calculated according to the daytime peak period (such as 7:00-9:00, 17:00-19:00) and the night low traffic period (such as 0:00-5:00) to evaluate the trend of urgency. Based on these parameters, the urgency index of the current task is calculated. , using the formula: ; Substitute specific values: ; ; The criteria for setting the judgment are as follows: (1) Low urgency (0.0-1.9): The monitoring tasks corresponding to this interval usually involve minor congestion or a single violation. For example, occasional illegal parking on a road section causes a short stop of vehicles. If the incident density is high, Less than 1 incident / hour, and the accident risk factor Less than 1.2, the urgency index is usually Below 1.9, in this case, the response priority of the monitoring task is low, and only regular video inspections are required, without emergency deployment of resources; (2) Medium urgency (2.0-3.9): This interval usually involves general traffic accidents or medium-scale traffic congestion. For example, a small scratch accident occurs at an intersection, affecting the traffic of 1-2 lanes. If the accident risk coefficient of the area is Between 1.2-1.5, and the accident density per unit time In the range of 1-2 events / hour, the urgency index is calculated Usually falls in the range of 2.0-3.9. In this case, the monitoring task needs to be prioritized, key areas need to be inspected, and the traffic management department needs to be notified to intervene if necessary; (3) High urgency (4.0-5.9): This range covers serious traffic accidents or large-scale traffic jams, such as a chain reaction on an urban expressway that blocks three or more lanes. The risk factor for such accidents is Usually greater than 1.5, if the accident density If more than 3 incidents / hour occur, or multiple incidents occur within a short period of time, the urgency index It may exceed 4.0. In this case, the monitoring task must be responded to first, monitoring resources must be dispatched in real time, and the traffic police or emergency rescue team must be notified to intervene; (4) Extreme emergency (6.0 and above): This interval represents serious abnormalities in the traffic system, such as a large-scale traffic accident on a highway, resulting in the closure of the entire highway, or multiple serious accidents on a city's main road, affecting the passage of thousands of vehicles. The risk coefficient of such events is The accident density can reach 2.0 or above. It may exceed 5 events / hour, resulting in an emergency index If it exceeds 6.0, the monitoring task will enter the highest response level, requiring real-time monitoring of all relevant areas, coordination with traffic management departments and rescue agencies to intervene simultaneously, and take emergency control measures.

[0024] Table 2 Task urgency index determination criteria As shown in Table 2, the results show that the current urgency index of the region is 4.526. According to the urgency index judgment standard (see Table 2), the value falls into the "high urgency" range (4.0 and above), which means that the regional monitoring task should be set as a high priority, and an immediate response and deployment of monitoring resources are required for real-time tracking and analysis. At the same time, the value is affected by the event density. The impact of the changes is significant. If the number of new accidents reaches 5 or more in the next hour, the urgency index may exceed 6.0, entering an extreme emergency state and requiring an upgrade of the emergency response level.

[0025] See also Figure 3 , the video storage resource calculation module includes The storage resource acquisition submodule obtains the available storage capacity of all current video storage servers, reads the storage space usage rate of the storage servers, counts the total number of tasks in the current video task queue, extracts the storage resource status data of each server, and obtains the storage resource status parameter set; To obtain the available storage capacity of all current video storage servers, first call data for each storage server to count the storage capacity of the server and the current available storage space. For example, a data center is equipped with 5 storage servers, and the total storage capacity of each server is 100TB, 120TB, 90TB, 110TB, and 130TB respectively. The current available storage space is 30TB, 40TB, 20TB, 35TB, and 50TB respectively. The total available storage capacity is 175TB. To obtain the storage space utilization rate of the storage server, the storage space utilization rate is calculated as (total storage capacity-available storage capacity) / total storage capacity. For example, for a server with a total storage capacity of 100TB and available storage space of 30TB, the storage space utilization rate is calculated as (100TB-30TB) / 100TB=70%. The average storage space utilization rate of all servers is calculated as follows: ; Count the total number of tasks in the current video task queue, extract the number of tasks to be stored from the task database, for example, the current task queue contains 240 video storage tasks, arrange them by time priority, extract the storage resource status data of each server, including the server's available storage space, storage utilization rate and task queue task volume, and obtain the storage resource status parameter set.

[0026] Table 3 Storage server status table As shown in Table 3, the storage status parameters of the current storage server have been obtained and can be used for subsequent calculation of task priority and storage proportion.

[0027] The task priority calculation submodule is based on the storage resource status parameter set and the video surveillance task urgency analysis results. It divides the tasks into high, medium and low priority tasks according to the task urgency, counts the number of high, medium and low priority tasks, and uses the formula ; Calculate the The proportion of tasks , integrate to get the proportion data of each priority task, among which, Representative The number of class tasks, Represents the total amount of high, medium, and low priority tasks; Based on the storage resource status parameter set, the video surveillance task urgency analysis results are called, and the priority of the video task is divided according to the urgency. According to the task urgency, it is divided into high priority, medium priority, and low priority, which respectively indicate that the task urgency is in the range of 80-100, 50-79, and 0-49. The number of tasks with different priorities is counted. For example, among the 240 tasks in the current task queue, there are 80 high priority tasks, 100 medium priority tasks, and 60 low priority tasks. The proportion of tasks with different priorities is calculated and the actual numerical value is brought in for calculation: ; ; ; Get the proportion of tasks of each priority level.

[0028] Table 4 Task priority ratio As shown in Table 4, the task priority distribution has been determined and can be used for subsequent storage weight allocation calculation.

[0029] The storage ratio allocation submodule adjusts the video storage resource allocation ratio based on the priority task ratio data, calculates the storage ratio that can be allocated to each video task, and obtains the video task storage ratio allocation result; According to the data of the proportion of tasks of each priority, the proportion of storage resource allocation is adjusted. According to the proportion of high, medium and low priority tasks, the allocation weight of storage resources is set. The setting of storage resource allocation weight is based on the urgency level of the task, the storage demand of the task and the timeliness requirement of task completion. In actual monitoring scenarios, high-priority tasks usually involve emergency tasks such as emergency monitoring and abnormal behavior detection. Their storage demand is generally large and requires higher storage bandwidth support. Therefore, the allocated storage weight is higher; medium-priority tasks involve daily monitoring or general security patrol tasks. The storage demand is relatively balanced and the weight is in the middle; low-priority tasks are mainly used for non-urgent tasks, such as historical data archiving, daily background data storage, etc. The storage demand is relatively small, so the weight is the lowest. The specific weight is set to 0.5 for high-priority tasks, 0.3 for medium-priority tasks, and 0.2 for low-priority tasks. According to the proportion of task urgency and storage demand statistical data, the storage proportion that can be allocated to each task is calculated using the formula: ; in, Representative The average storage allocation for class tasks, Representative task ratio, Represents the task priority storage weight, Represents a total available storage capacity of 175TB, Represents the number of priority tasks.

[0030] The task storage weight is set based on the following: High-priority tasks (weight 0.5): Urgent tasks usually need to be stored for a long time for subsequent analysis, such as abnormal behavior recognition or crime tracking tasks. The general storage time can be up to 30-60 days, and the amount of video data for a single task is large (usually 50GB-200GB / hour), so it is given the highest weight.

[0031] Medium priority tasks (weight 0.3): Daily monitoring tasks are mainly used for real-time viewing and short-term archiving. The storage period is generally 15-30 days. The data volume is small (usually 20GB-100GB / hour), and the storage demand is relatively medium.

[0032] Low priority tasks (weight 0.2): Low priority tasks are mainly used for data archiving or long-term storage. The storage period is usually more than 60 days, but the data access frequency is low, so it occupies less storage resources and has the lowest storage demand.

[0033] Bring in data calculation: ; ; ; After the calculation is completed, the video task storage weight distribution results are obtained. The results show that a single high-priority task can be allocated 0.365TB, a single medium-priority task can be allocated 0.219TB, and a single low-priority task can be allocated 0.146TB. This data can be used for storage task scheduling and allocation.

[0034] See also Figure 4 , the video storage allocation and optimization module includes The high-urgency task storage demand calculation submodule extracts the storage allocation data corresponding to the high-urgency task in combination with the video task storage weight distribution result, calculates the video storage demand of the high-urgency task, and obtains the high-urgency task storage demand data; According to the video task storage weight distribution results, the storage allocation data corresponding to the high-urgency tasks is determined. The calculation of storage requirements first requires identifying the storage duration, video encoding format, resolution, and frame rate of each high-urgency task. The storage requirements of a single task are calculated based on these parameters. Assuming that the average storage duration of a high-urgency task is 2 hours, the video encoding format is H.265, the resolution is 1080P, the frame rate is 30 frames / second, and the size of each frame is about 0.5MB, the storage requirement of a single task is calculated as follows: ; ; Assuming that there are 100 high-urgency tasks in the current system, the total storage requirement is calculated as follows: ; ; The final storage requirement value for high-urgency tasks is 10.8TB.

[0035] The video data fragmentation rate calculation submodule obtains the existing video data storage status in the storage server based on the high-urgency task storage demand data, extracts the video data occupancy of each storage unit, and uses the formula: Calculate video data fragmentation rate , get the data fragmentation rate data of the storage server, where, Representative The amount of video data stored in the storage unit, Representative The free storage capacity of the storage unit, Represents the total number of storage cells that store data, Represents the total number of storage units of the current server; Get the storage status of existing video data in the storage server and extract the video data occupancy of each storage unit. The distribution status of storage blocks inside the storage server affects the data fragmentation rate. Assume that a storage server contains 500 storage units, of which 250 storage units have stored video data and 250 storage units are idle. Assume that the amount of stored video data in each storage unit is as follows (partial example) Table 5 Video data storage unit usage table As shown in Table 5, the storage capacity of storage units varies. Assuming the total storage data volume is 20,000 GB and the total free capacity is 10,000 GB, the data fragmentation rate is calculated as follows: ; ; The calculated data fragmentation rate of the storage server is 0.187, which means the fragmentation rate is 18.7%.

[0036] The video data block merging submodule performs video data block merging based on the data fragmentation rate data of the storage server, adjusts the arrangement of fragmented data blocks and the video storage structure in the storage unit, and obtains the optimized video task storage allocation state; According to the data fragmentation rate value of the storage server, the video data blocks are merged. The fragmented data blocks of the storage server affect the storage efficiency. Assuming that there are multiple data blocks in a server, each block has different sizes. Now the data block merging strategy is adopted to rearrange the storage blocks and fill the free storage units to optimize the storage structure. Assuming that the available storage space of the server is 10000GB and the current high-urgency task storage demand is 10800GB, the adjustment plan is: Identify the amount of video data stored in the storage unit, as shown in Table 5; Calculate the fill ratio of each storage unit to ensure that high-urgency tasks are stored first; Sort by task urgency and reallocate storage blocks, for example, merge scattered data blocks into continuous storage blocks.

[0037] After the adjustment, the fragmentation rate after storage optimization is recalculated using the above formula. Assuming that the total fragmentation rate is reduced to 10.5% after merging, it means that the fragmentation rate has dropped from 18.7% to 10.5%, indicating that the optimization effect is significant. Finally, the optimized video task storage allocation status is obtained.

[0038] See also Figure 5 , the monitoring target resident analysis module includes: The monitoring target residence time calculation submodule obtains the entry timestamp of the monitoring target according to the optimized video task storage allocation status, reads the target exit timestamp, calculates the video recording time of the target in the monitoring area, screens the corresponding relationship between continuous entry and exit records, calculates the residence time of different targets in different time periods, summarizes the residence time distribution of targets in the monitoring area, and obtains the monitoring target residence time data; Call the optimized video task storage allocation status, obtain the entry timestamp of the monitored target, read the target exit timestamp, ensure data integrity and then match. If there are multiple entry and exit records for the target, filter the corresponding entry and exit times in chronological order, and calculate the video recording time of the monitored target in the monitored area. First, extract the historical timestamp data for each target, filter abnormal data, such as duplicate, missing or invalid timestamp data, and establish the target time series after correcting the data. Then calculate the residence time based on the timestamp difference. For example, if a target enters the monitored area at 08:30:00 and leaves at 09:15:00, the residence time is 45 minutes. If the target enters and exits multiple times in one day, the time period is merged to calculate its total residence time. The residence time of multiple monitored targets in different time periods is counted, and the residence situation is summarized by time range. For example, 15 targets reside in the interval of 08:00-10:00, with a total residence time of 8 hours. Record the daily residence time distribution, and finally obtain the residence time data of the monitored target.

[0039] The target entry frequency statistics submodule counts the number of target entries based on the target residence time data, analyzes the activity of different targets in the monitoring area, calculates the target entry frequency based on the observation period, analyzes the time distribution characteristics of the target entering the area, and obtains the target entry frequency data; Based on the target's residence time data, the number of times the target enters is counted, the target's activity records in the monitoring area are obtained, and the daily or hourly entry records are extracted. The duplicate records are removed and organized into entry time series. For example, if a target enters the monitoring area at 10:05, 12:30, and 16:20, then the number of entries per day is 3 times. Targets with high-frequency entries are screened out, and an observation period is set, such as 7 days or 30 days. The total number of entries within the period is counted. For example, if the target enters the monitoring area 28 times in 7 days, its average entry frequency is calculated, and the time interval is divided into hours, days, and weeks. The time distribution characteristics of the target entering the area are summarized. For example, the number of entries during the 08:00-10:00 period accounts for 40%, forming entry behavior trend statistics, and finally obtaining the target entry frequency data.

[0040] The storage level demand judgment submodule calculates the target's departure time interval based on the target entry frequency data, counts the retention periods of different targets, and compares the storage priorities of different targets based on the target's video stay time, entry frequency, and departure time interval in the monitoring area. It adjusts the video storage level according to the storage demand classification and obtains the storage level demand analysis results of the video monitoring target; Combined with the target entry frequency data, the target's departure time interval is calculated, the target's residence pattern in the monitoring area is analyzed, and all monitored targets are classified according to entry frequency, residence time and departure interval. The specific classification basis is as follows: High-frequency, short-term residence targets: These targets enter the monitoring area frequently, but stay for a short time each time. For example, a target enters the monitoring area 10 times a day, and stays for no more than 5 minutes each time. These targets usually appear in crowded areas, such as shopping mall entrances and transportation hubs. Storage requirements mainly rely on short-term storage, such as cache or temporary storage. Low-frequency long-stay targets: These targets enter the monitoring area less frequently, but stay longer each time. For example, a target enters the monitoring area once a day on average, but stays for more than 3 hours each time. These targets may be staff members or resident personnel, and their storage requirements tend to be long-term storage for subsequent data tracing. Random entry targets: The entry time and residence time of such targets do not have obvious patterns. For example, a target may enter the monitoring area multiple times in one day, but the entry situation on different days varies greatly. The storage demand of such targets needs to be dynamically adjusted according to the entry records, and the storage priority is at a medium level. High-frequency and long-stay targets: These targets have both high entry frequency and long-stay time. For example, a target enters the monitoring area 8 times a day and stays for more than 2 hours each time. These targets usually have a higher storage priority and need to be fully recorded and stored at a high level. Low-frequency and short-stay targets: These targets have low entry frequency and short stay time. For example, a target enters the monitoring area twice a week on average and stays for only 3 minutes each time. The storage requirements for these targets are low, and the lowest level of storage or interval storage strategy can be arranged.

[0041] During the classification process, we first calculate the average daily number of entries for each target, count the residence time of each target, and classify the targets based on the set thresholds. For example, the high-frequency standard is set to more than 5 entries per day, the long-stay standard is set to a residence time of more than 1 hour, and the short-stay standard is set to a residence time of less than 10 minutes. Finally, the storage level is matched according to the classification results, and the storage strategy is adjusted to obtain the video surveillance target storage level demand analysis results.

[0042] See also Figure 6 ,The video storage level adjustment module includes: The target entry frequency analysis submodule combines the video surveillance target storage level demand analysis results, obtains the entry records of the monitored targets, counts the number of targets’ entries within a specified time period, calculates the entry frequency values ​​of different targets, filters out targets with entry frequencies higher than the set threshold, and calculates the entry frequency distribution interval to obtain the target entry frequency distribution data; First, the monitoring system collects data on the target's entry records, with a 24-hour period, recording the target's entry once an hour to form a complete data set, counting the number of entries of each target within the period, calculating the target's entry frequency, and setting the entry frequency interval, such as low frequency (0-2 times), medium frequency (3-6 times), and high frequency (7 times and above), and classifying them based on this. In actual application scenarios, taking shopping mall entrance monitoring as an example, the number of times different customers enter the mall can be recorded. If a customer enters the mall 5 times in a day, his entry frequency belongs to the medium frequency interval. All high-frequency targets are screened, and the mean and standard deviation of their entry frequencies are calculated to obtain the target entry frequency distribution data.

[0043] The storage retention time adjustment submodule sets storage retention time adjustment factors for targets in different frequency intervals based on the target entry frequency distribution data, calculates the storage time adjustment value corresponding to each target, and adjusts the storage retention time to obtain storage retention time adjustment data; Based on the target entry frequency distribution data, the storage retention time adjustment factor is set according to the entry frequency range. For example, the video of high-frequency targets is retained for 72 hours, medium-frequency targets for 48 hours, and low-frequency targets for 24 hours. The storage retention time of each target is calculated, and the adjustment factor is determined according to the entry frequency value. The formula is: ; in is the default storage time, is the target entry frequency, and are the minimum and maximum entry frequencies respectively. If the default storage time is 24 hours, the target entry frequency is 5, the maximum entry frequency is 10, and the minimum is 1, then the calculation is: ; The final target video storage duration is adjusted to 34.56 hours, and the storage retention duration adjustment data is obtained.

[0044] The storage server load monitoring submodule monitors the current load of the storage server, collects storage occupancy, access rate and storage pressure value, calculates the storage occupancy, determines whether the load status exceeds the load threshold, and obtains the storage server load status data; Monitor the storage occupancy, access rate and storage pressure value of the storage server, record the storage usage every 10 minutes, and set the storage occupancy threshold based on the storage device's capacity limit and data write rate. Under the current storage architecture, if the actual available capacity of the storage device is less than 15% of the total storage capacity, the storage occupancy exceeds 85%. This value is calculated based on the storage space allocation strategy. Considering the redundant storage overhead of the RAID mechanism, when the storage space reaches 85%, the write rate usually decreases, affecting the stability of data storage. Therefore, 85% is set as the high load threshold, and the access rate is set based on data throughput. Under the current storage architecture, the data write rate peak can reach 120MB / s. When the storage pressure is high, the I / O processing capacity of the storage controller is close to the upper limit. Therefore, when the access rate exceeds 100MB / s, the storage load tends to be saturated, resulting in reduced data access efficiency. The storage pressure value is low, so 100MB / s is set as the high load threshold of the access rate. The setting of the storage pressure value is based on the analysis of IOPS (input / output operations per second) and queue length. When IOPS exceeds 5000 and the queue depth is greater than 10, the response time of the storage system will be significantly delayed. At this time, the calculated storage pressure value is usually higher than 0.8, so 0.8 is set as the storage pressure high load threshold. Under high load, the storage occupancy rate exceeds 85%, the access rate is higher than 100MB / s, and the storage pressure value is greater than 0.8. Under normal load, the storage occupancy rate is between 60%-85%, the access rate is 50MB / s-100MB / s, and the storage pressure value is between 0.5-0.8. If the storage occupancy rate of a server at a certain moment is 88%, the access rate is 110MB / s, and the storage pressure value is 0.85, it is judged to be in a high load state, and the storage server load status data is obtained.

[0045] The low-priority video storage compression submodule screens low-priority video data based on the storage retention time adjustment data and the storage server load status data, using the formula: ; Calculate the compression ratio of low priority video data storage , get the video storage level adjustment result; where, Representative Storage size of video files, Represents the remaining storage time of the video. Represents the target entry frequency of the video, Represents the load ratio of the server where the video is located. Represents the bandwidth usage of the video storage server. Represents the average data flow during the video storage period, Represents the current bandwidth consumption of video storage, Represents the frequency of video access. Represents the number of compression histories of the video. is the total number of low priority videos; Based on the storage retention time adjustment data and the storage server load status data, filter the video data with a storage time of less than 48 hours and video data with a high server load, and calculate the storage compression ratio for low-priority videos. Assume that the stored video data is as follows: Table 6 Low priority video storage parameter table Substitute the data into the formula to calculate: ; ; ; The storage compression ratio adjustment value is 17.9, indicating that the low-priority video data can be compressed by 17.9 times, thereby obtaining the video storage level adjustment result.

[0046] 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 dynamic video surveillance resource management system, characterized in that: The system comprises: The task urgency calculation module obtains the event triggering information of the video monitoring task, calculates the monitoring task coverage radius, analyzes the monitoring task urgency, and obtains the video monitoring task urgency analysis result; The video storage resource calculation module reads the storage space usage rate of the storage server based on the urgency analysis result of the video surveillance task, calculates the storage proportion that can be allocated to each video task, and obtains the video task storage proportion allocation result; The video storage allocation and optimization module calculates the video storage requirements of high-urgency tasks based on the video task storage weight allocation result, performs video data block merging, and obtains an optimized video task storage allocation state; The monitoring target resident analysis module calculates the target video recording duration in the monitoring area, the target entry frequency, and the target leaving time interval based on the optimized video task storage allocation state, determines the video storage level requirement, and obtains the video monitoring target storage level requirement analysis result.

2. The dynamic video surveillance resource management system according to claim 1, characterized in that: The video surveillance task urgency analysis result includes the task trigger source category, event severity, monitoring task coverage radius, and monitoring task priority; the video task storage proportion allocation result includes the high priority task storage proportion, medium priority task storage proportion, low priority task storage proportion, and storage server storage load proportion; the video task storage allocation status includes the high urgency task storage allocation status, storage server data fragmentation rate, and video data block optimization status; the video surveillance target storage level requirement analysis result includes the target video recording duration, target entry frequency, target departure time interval, and storage level requirement level.

3. The dynamic video surveillance resource management system according to claim 1, characterized in that: The task urgency calculation module includes: The event trigger information acquisition submodule acquires the event trigger information of the video surveillance task, reads the task trigger source and event category, analyzes the time and space characteristics of the event category, extracts the correlation information between the event location, event time, event type and task source, and obtains the event correlation information parameter set; The monitoring task coverage radius calculation submodule analyzes and determines the coverage required for the monitoring task based on the event association information parameter set, using the formula: ； Calculate the monitoring task coverage radius value ,in, Represents the area affected by the event, Represents the speed of event propagation, Represents the time required for response; The monitoring task urgency analysis submodule calculates the urgency index of the current monitoring task according to the monitoring task coverage radius value, combined with the event type risk coefficient, event occurrence density, and historical urgency distribution, and obtains the video monitoring task urgency analysis result.

4. The dynamic video surveillance resource management system according to claim 1, characterized in that: The video storage resource calculation module includes The storage resource acquisition submodule obtains the available storage capacity of all current video storage servers, reads the storage space usage rate of the storage servers, counts the total number of tasks in the current video task queue, extracts the storage resource status data of each server, and obtains the storage resource status parameter set; The task priority calculation submodule is based on the storage resource status parameter set and the video surveillance task urgency analysis result, and divides the tasks into high, medium and low priority tasks according to the task urgency, and counts the number of high, medium and low priority tasks, using the formula ； Calculate the The proportion of tasks , integrate to get the proportion data of each priority task, among which, Representative The number of class tasks, Represents the total amount of high, medium, and low priority tasks; The storage proportion allocation submodule adjusts the video storage resource allocation proportion based on the priority task proportion data, calculates the storage proportion that can be allocated to each video task, and obtains the video task storage proportion allocation result.

5. The dynamic video surveillance resource management system according to claim 1, characterized in that: The video storage allocation and optimization module includes The high-urgency task storage requirement calculation submodule extracts the storage allocation data corresponding to the high-urgency task in combination with the video task storage weight distribution result, calculates the video storage requirement of the high-urgency task, and obtains the high-urgency task storage requirement data; The video data fragmentation rate calculation submodule obtains the existing video data storage status in the storage server based on the high-urgency task storage demand data, extracts the video data occupancy of each storage unit, and adopts the formula: ； Calculate video data fragmentation rate , get the data fragmentation rate data of the storage server, where, Representative The amount of video data stored in the storage unit, Representative The free storage capacity of the storage unit, Represents the total number of storage cells that store data, Represents the total number of storage units of the current server; The video data block merging submodule performs video data block merging based on the data fragmentation rate data of the storage server, adjusts the arrangement of fragmented data blocks and the video storage structure in the storage unit, and obtains an optimized video task storage allocation state.

6. The dynamic video surveillance resource management system according to claim 1, characterized in that: The monitoring target resident analysis module includes: The monitoring target residence time calculation submodule obtains the entry timestamp of the monitoring target according to the optimized video task storage allocation state, reads the target exit timestamp, calculates the video recording time of the target in the monitoring area, screens the corresponding relationship between continuous entry and exit records, calculates the residence time of different targets in different time periods, summarizes the residence time distribution of targets in the monitoring area, and obtains the monitoring target residence time data; The target entry frequency statistics submodule counts the number of target entries based on the residence time data of the monitored target, analyzes the activity of different targets in the monitored area, calculates the target entry frequency based on the observation period, analyzes the time distribution characteristics of the target entering the area, and obtains the target entry frequency data; The storage level demand judgment submodule calculates the target's departure time interval based on the target entry frequency data, counts the retention periods of different targets, combines the target's video residence time, entry frequency, and departure time interval in the monitoring area, compares the storage priorities of different targets, and adjusts the video storage level according to the storage demand classification to obtain the video surveillance target storage level demand analysis results.

7. The dynamic video surveillance resource management system according to claim 1, characterized in that: The system also includes a video storage level adjustment module; The video storage level adjustment module adjusts the video storage retention time based on the storage level demand analysis result of the video surveillance target according to the target's access frequency, and performs storage compression on low-priority video tasks in combination with the storage server load status to release storage space and obtain the video storage level adjustment result; The video storage level adjustment result includes the video storage retention time adjustment, storage server load balancing status, low priority task storage compression ratio, and storage space release amount.

8. The dynamic video monitoring resource management system according to claim 7, characterized in that: The video storage level adjustment module includes: The target entry frequency analysis submodule combines the storage level demand analysis results of the video surveillance target to obtain the entry records of the surveillance target, count the number of entries of the target within a specified time period, calculate the entry frequency values ​​of different targets, filter the targets whose entry frequency is higher than the set threshold, and calculate the entry frequency distribution interval to obtain the target entry frequency distribution data; The storage retention time adjustment submodule sets storage retention time adjustment factors for targets in different frequency intervals based on the target entry frequency distribution data, calculates storage time adjustment values ​​corresponding to each target, and adjusts the storage retention time to obtain storage retention time adjustment data; The storage server load monitoring submodule monitors the current load of the storage server, collects storage occupancy, access rate and storage pressure value, calculates the storage occupancy, determines whether the load status exceeds the load threshold, and obtains the storage server load status data; The low-priority video storage compression submodule screens low-priority video data based on the storage retention time adjustment data and the storage server load status data, using the formula: ； Calculate the compression ratio of low priority video data storage , get the video storage level adjustment result; where, Representative Storage size of video files, Represents the remaining storage time of the video. Represents the target entry frequency of the video, Represents the load ratio of the server where the video is located. Represents the bandwidth usage of the video storage server. Represents the average data flow during the video storage period, Represents the current bandwidth consumption of video storage, Represents the frequency of video access. Represents the number of compression histories of the video. The total number of low priority videos.

Images