Data storage method and system based on monitoring memory

Abstract

The present invention relates to the technical fields of semiconductor devices, data storage, etc., and provides a data storage method and system based on a monitoring memory. By configuring different priorities and data complexities for data streams from different video sources, a communication connection is established with front-end monitoring cameras of different video sources, and data streams of different video sources collected by the front-end monitoring cameras are received. The front-end monitoring camera includes an encoder, and the encoder encodes the video data collected by the front-end monitoring camera to generate a data stream, automatically identifies the received data streams of different video sources, dynamically adjusts the CPU resources and memory resources of the processing thread according to the priority of the configured data stream, and stores and processes the received data streams of different video sources, thereby reducing dependence on a storage server, reducing system costs, ensuring data processing speed and storage quality, optimizing data storage, and reducing single point failures and performance bottlenecks of data storage.

Description

Technical Field

[0001] The present invention relates to the technical fields of semiconductor devices, data storage, and more particularly to a data storage method and system based on a monitoring memory. Background Art

[0002] In traditional surveillance systems, storage servers can present single points of failure and performance bottlenecks. A single point of failure refers to a failure in any critical component of a system, potentially halting the entire system. In traditional surveillance systems, all data flows through a central storage server. Any hardware or software failure on the central storage server can cause the entire surveillance system to fail. For example, a damaged storage server hard drive, power supply issues, or network interface card failure can render surveillance data inaccessible. Furthermore, since surveillance data typically requires high real-time performance, server failures can prevent the ability to review critical video footage, compromising security monitoring effectiveness. Furthermore, a performance bottleneck refers to a performance limitation in a specific component of the data processing and transmission process that impacts the performance of the entire system. In surveillance systems, this often manifests as insufficient processing and storage capacity to meet the demands of high-speed, high-volume data processing and storage. For example, a surveillance system may need to process video data from hundreds of cameras. Insufficient storage server processing power or network bandwidth can lead to data transmission delays and even the loss of important data. For example, when writing and reading data, the storage server may encounter input / output operation bottlenecks. Especially in a multi-tasking environment, the simultaneous writing of multiple video streams may significantly reduce the system's response speed.

[0003] In summary, the data storage technology of existing monitoring systems has technical problems such as single point failure and performance bottleneck. Summary of the Invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a data storage method and system based on a monitoring memory to optimize data storage.

[0005] In a first aspect, the present invention provides a data storage method based on a monitoring memory, comprising:

[0006] Configure different priorities and data complexity for data streams from different video sources;

[0007] The front-end surveillance camera is communicatively connected to a front-end surveillance camera of different video sources, and receives data streams of different video sources collected by the front-end surveillance camera, wherein the front-end surveillance camera includes an encoder, and the encoder encodes the video data collected by the front-end surveillance camera to generate the data stream;

[0008] Automatically identify the data streams received from different video sources, dynamically adjust the CPU resources and memory resources of the processing thread according to the configured priorities of the data streams, and store the data streams received from different video sources in the monitoring memory.

[0009] In a second aspect, the present invention provides a data storage system based on a monitoring memory, comprising:

[0010] Monitoring memory, used to store data streams from different video sources;

[0011] An execution processor is communicatively connected with the monitoring memory and is used to run the above-mentioned data storage method based on the monitoring memory.

[0012] Compared with the prior art, the present invention has the following beneficial effects:

[0013] The present invention provides a data storage method and system based on a monitoring memory. By configuring different priorities and data complexities for data streams from different video sources, the system communicates with front-end monitoring cameras of different video sources, receives data streams from different video sources captured by the front-end monitoring cameras, and includes an encoder that encodes the video data captured by the front-end monitoring cameras to generate the data streams. The system automatically identifies the received data streams from different video sources and dynamically adjusts the CPU resources and memory resources of the processing thread based on the configured priorities of the data streams. The system stores the received data streams from different video sources in the monitoring memory, thereby reducing dependence on the storage server, lowering system costs, and preventing the server from becoming a single point of failure for the monitoring system. Furthermore, since the video data is stored in the form of a stream, file fragmentation can be avoided, which helps to improve data reading and writing efficiency and the service life of the memory. In addition, by dynamically adjusting resource allocation based on the priority of the data stream, it can ensure that high-priority data streams receive sufficient processing resources, thereby ensuring their processing speed and storage quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their description are used to explain the present invention and do not constitute an undue limitation of the present invention. Some specific embodiments of the present invention will be described in detail in an illustrative and non-restrictive manner with reference to the drawings. The same reference numerals in the drawings indicate the same or similar components or parts. It should be understood by those skilled in the art that these drawings are not necessarily drawn to scale. In the drawings:

[0015] Figure 1 This is a flow chart of a data storage method based on a monitoring memory according to an embodiment of the present invention;

[0016] Figure 2 A schematic diagram of an architecture for communicating between an execution memory and different front-end surveillance cameras according to an embodiment of the present invention;

[0017] Figure 3 This is a schematic diagram of the architecture of a data storage system based on a monitoring memory according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0019] Example 1

[0020] See also Figure 1-Figure 3 This embodiment provides a data storage method based on a monitoring memory, including steps S101, S102, and S103. By configuring different priorities and data complexities for data streams from different video sources, a method is provided to communicate with front-end monitoring cameras of different video sources, receive data streams from different video sources captured by the front-end monitoring cameras, the front-end monitoring cameras including an encoder that encodes the video data captured by the front-end monitoring cameras to generate the data streams, automatically identify the received data streams from different video sources, dynamically adjust the CPU resources and memory resources of the processing thread based on the configured priorities of the data streams, and store the received data streams from different video sources in the monitoring memory, thereby reducing dependence on the storage server, lowering system costs, and preventing the server from becoming a single point of failure for the monitoring system. Furthermore, since the video data is stored in the form of streams, file fragmentation can be avoided, which helps to improve data reading and writing efficiency and the service life of the memory. In addition, dynamically adjusting resource allocation based on the priority of the data stream can ensure that high-priority data streams receive sufficient processing resources, thereby ensuring their processing speed and storage quality.

[0021] Step S101: Configure different priorities and data complexities for data streams from different video sources. In some preferred embodiments, when configuring different data complexities for data streams from different video sources, different resolutions and frame rates can be set for each video source based on the video surveillance scenario requirements. Different resolutions and frame rates correspond to different data complexities for the data streams. For example, surveillance cameras in important areas (such as the vault entrance) are configured with high resolutions and high frame rates to capture more details and movements; less important areas (such as parking lots) are configured with lower resolutions and frame rates. Furthermore, the different data complexities of the data streams are automatically adjusted according to a preset schedule. For example, the resolution and frame rate of the front-end surveillance camera are automatically increased at night.

[0022] In some preferred embodiments, assigning different priorities to data streams from different video sources may include: configuring monitoring areas for front-end surveillance cameras in different regions, categorizing these monitoring areas according to security sensitivity to create areas of varying security sensitivity; and assigning different priorities to data streams from video sources in areas of varying security sensitivity, with higher-priority data streams corresponding to areas of higher security sensitivity. It should be noted that the security sensitivity of monitoring areas varies. For example, locations such as the entrances to financial institutions and shopping malls are more sensitive, while warehouses and office areas are less so. Categorizing data streams based on security sensitivity allows for more efficient resource allocation, ensuring adequate surveillance coverage for high-risk areas. Processor resources (such as memory and computing power) are limited and need to be allocated rationally to ensure efficiency. Configuring video data streams from areas of high security sensitivity as high priority means that video from these areas is processed first during data processing and storage, ensuring that events in critical areas are captured and recorded promptly and clearly. By prioritizing security sensitivity, processor resources are focused where they are most needed, reducing resource waste in non-critical areas.

[0023] In some preferred embodiments, assigning different priorities to data streams from different video sources may also include: configuring monitoring areas for front-end surveillance cameras in different regions, categorizing these monitoring areas according to accident rates to identify regions with different accident rates; and assigning different priorities and data complexities to data streams from video sources in regions with different accident rates, with regions with higher accident rates corresponding to higher-priority data streams. It should be noted that categorizing monitoring areas by accident rate allows identification of areas that are more prone to accidents or other safety issues and therefore require more intensive monitoring. The resources (memory and computing power) of the execution processor are limited and need to be properly allocated based on priority to maximize their utility. Video streams from regions with high accident rates are assigned higher priorities, ensuring that video data from these regions receives priority in processing and storage, thereby ensuring the quality of recording critical events.

[0024] In some preferred embodiments, configuring different priorities for data streams from different video sources may further include: configuring a set of trigger events for triggering video data acquisition, the set of trigger events including trigger events of varying urgency, wherein the trigger events of varying urgency trigger the front-end surveillance camera to acquire video data; and configuring different priorities and data complexities for the data streams corresponding to the video data acquired by the trigger events of varying urgency, with more urgent trigger events corresponding to higher-priority data streams. Furthermore, the trigger events of varying urgency include fire detector alarm events, illegal security door opening alarm events, sudden increase in crowd density events, and abnormal facility operation alarm events, with the urgency of the fire detector alarm events, illegal security door opening alarm events, sudden increase in crowd density events, and abnormal facility operation alarm events decreasing in order. Furthermore, the fire detector alarm event indicates that when a fire detector detects fire or smoke within the monitored area and is triggered, a recording start signal is transmitted to the front-end surveillance camera to trigger the front-end surveillance camera to acquire video data. The illegal security door opening alarm event indicates that when an unauthorized intrusion is detected by the security door, a recording start signal is transmitted to the front-end surveillance camera to trigger the front-end surveillance camera to acquire video data. It should be noted that the urgency of an event determines the priority of its response. For example, a fire alarm has a higher urgency than other events, so a higher data stream priority is assigned to ensure rapid and prioritized processing. This configuration strategy allows resources to be effectively allocated to where they are most needed, especially when resources are limited. In this embodiment, event-triggered video data collection allows for more targeted resource utilization, avoiding unnecessary resource consumption. It should also be noted that a sudden increase in crowd density occurs when the number of people in a specific monitoring area increases rapidly within a short period of time (e.g., a preset short period of time), exceeding a preset threshold. This sudden increase is typically caused by a specific event, emergency, or other social gathering. Using video analysis technologies, such as facial recognition and motion tracking, image processing technologies can accurately monitor crowd dynamics. A facility operation abnormality alarm event indicates an abnormality in the operation of facilities or equipment within the monitoring area, potentially due to mechanical failure, system malfunction, or unauthorized interference. When a facility operates abnormally, sensor or communication technologies can be used to transmit an abnormality signal to a surveillance camera, triggering the camera to collect video data.

[0025] Step S102: Communicate with a front-end surveillance camera of different video sources and receive data streams from different video sources captured by the front-end surveillance camera. The front-end surveillance camera includes an encoder that encodes the video data captured by the front-end surveillance camera to generate the data stream. In some preferred embodiments, the encoder converts the video signal captured by the front-end surveillance camera into video data and compresses the video data using the H.264 video coding standard or the H.265 video coding standard to generate the data stream. The video data includes a data header and a data body. The data header records the ID or location code of the front-end surveillance camera, and the data body records video content information.

[0026] It should be noted that, in this embodiment, the encoder of the front-end surveillance camera encodes the video data collected by the front-end surveillance camera to generate the data stream. In the subsequent steps, the generated data stream is directly input into the monitoring memory for storage, thereby reducing dependence on the storage server, reducing system costs, and preventing the server from becoming a single point of failure for the monitoring system. In addition, the use of H.264 or H.265 video encoding standards can significantly reduce the size of video files, thereby reducing network bandwidth requirements and storage space usage. The data header records the camera ID or location code, which facilitates the system to quickly locate the source of the video data and facilitates subsequent data identification and analysis.

[0027] Step S103 automatically identifies the received data streams from different video sources, dynamically adjusts the CPU and memory resources of the processing thread based on the configured data stream priorities and data complexity, and stores the received data streams from different video sources in the monitoring memory. It should be noted that different video data streams have different priorities due to factors such as the importance or urgency of their content. Dynamically adjusting processing resources (such as CPU and memory) ensures that critical data streams receive sufficient resources for efficient processing, thereby prioritizing the most important or urgent video data. Furthermore, video streams of varying data complexity have different resource requirements. For example, high-resolution video requires more processing power and storage space. Dynamic resource adjustment allows the system to allocate resources based on actual needs, improving processing efficiency. In this embodiment, dynamic resource allocation can reduce system latency and increase processing speed. This can significantly improve overall system performance, particularly when multiple video sources are input. This ensures that high-priority tasks receive the necessary resources, guarantees data processing speed and storage quality, optimizes data storage, and reduces single points of failure and performance bottlenecks in data storage.

[0028] In some preferred embodiments, automatically identifying the received data streams from different video sources may include: after receiving the data streams from different video sources, parsing the data header of the video data in the data stream to obtain the ID or location code of the front-end surveillance camera; and automatically identifying the received data streams from different video sources based on the ID or location code of the front-end surveillance camera obtained by parsing. It should be noted that in this embodiment, the camera ID or location code contained in the data header can distinguish and process video streams from different locations or devices. By identifying the video source, corresponding processing resources can be allocated according to the specific location or importance of the camera. For example, for cameras in security-sensitive or accident-prone areas, more computing resources can be allocated preferentially to ensure that video data is processed in a timely and effective manner.

[0029] In some preferred embodiments, automatically identifying the data streams received from different video sources includes: collecting data streams from different video sources to extract color histogram features, gray level co-occurrence matrix features, and inter-frame difference features to obtain training sample data; using the training sample data to train an SVM model to obtain a trained SVM model that can distinguish data streams from different video sources; extracting color histogram features, gray level co-occurrence matrix features, and inter-frame difference features from the data streams received from different video sources to obtain feature data to be identified, inputting the feature data to be identified into the trained SVM model, and the trained SVM model automatically identifying the data streams received from different video sources. It should be noted that in this embodiment, the use of image features such as color histogram, gray level co-occurrence matrix, and inter-frame difference can provide rich information to describe and distinguish video content. These features are more descriptive and discriminative than simple data header information. The SVM model can effectively process high-dimensional data and can learn the best way to distinguish different video sources from complex features through training. In a changing surveillance environment, especially when visually similar or missing data header information, a single data header is insufficient to accurately identify all video sources. The SVM model can better adapt to environmental changes and cope with the complexity of visual information. It should be noted that in this embodiment, color histogram features, gray-level co-occurrence matrix features, and inter-frame difference features can describe video content from different dimensions, such as color, texture, and temporal variation, thereby providing more comprehensive information about video source characteristics and facilitating the construction of a more comprehensive data description model. By extracting these complex features, rich training sample data can be generated, enabling the trained SVM model to more accurately learn and distinguish characteristics from different video sources. By extracting the same feature set during runtime as during the training phase, the consistency and comparability of the data to be identified with the training data at the feature level are ensured, thereby ensuring model accuracy. It is understood that color histogram features can describe the distribution and frequency of colors in a video. Color histograms capture the color composition of video frames and are very effective in distinguishing video sources with different color patterns, such as natural scenes and indoor environments. In this embodiment, the color histogram feature can be used to identify and distinguish video streams with unique color distributions. For example, different monitoring areas may have different color characteristics, and the color histogram helps the model learn these differences. The gray-level co-occurrence matrix feature (GLCM) can describe the texture information in the image. By analyzing the grayscale relationship between pixels and extracting the image's texture statistical characteristics, such as contrast, uniformity, and entropy, it can help identify specific types of scenes or activities, such as densely populated and non-dense areas, or external environments under different weather conditions. The inter-frame difference feature can be used to capture information that changes over time in a video sequence, for example, to detect and describe motion information in a video.In this embodiment, the inter-frame difference feature can effectively reveal dynamic changes in the area, such as moving vehicles or active people, and can effectively identify abnormal events or behavior patterns.

[0030] In some preferred embodiments, dynamically adjusting the CPU resources and memory resources of the processing thread according to the configured priority and data complexity of the data flow may include:

[0031] Definition The priority of a data stream is , No. The data complexity of a data stream is , No. The amount of CPU resources required for each data flow is , the sum of the CPU resources of all processing threads is , No. The amount of memory resources required for a data stream is , the sum of memory resources of all processing threads is ; The sum of CPU resources of all processing threads , the sum of memory resources Satisfy the constraint formula:

[0032] ; ;

[0033] Setting resource requirement function and , resource demand function and Satisfies the following formula:

[0034] ;

[0035] ; ;

[0036] parameter , , , , , The resource requirement function is determined through system analysis and actual measurement. and Dynamically calculate the required CPU and memory resources based on the priority and complexity of the data flow;

[0037] Prioritize all data flows Sort in descending order and allocate the CPU resources required for each data flow in turn and memory resources .

[0038] It should be noted that in this embodiment, all data streams are sorted in descending order of priority, and the calculated CPU resources and memory resources are allocated to each data stream in turn to ensure that the total resources of all processing threads do not exceed the maximum available resources of the system, and dynamically adjust resource allocation to ensure that critical data streams (such as high-priority or high-complexity data streams) obtain sufficient resources, thereby ensuring their processing speed and storage quality.

[0039] In some preferred embodiments, when data streams originate from both areas of high security sensitivity and areas with high accident rates, the highest priority is assigned to overlapping data streams from these areas, prioritizing CPU and memory resources in the processing thread for storage and processing of these overlapping data streams. It should be noted that when data streams originate from areas with both high security sensitivity and a high accident rate, this indicates a dual risk, potentially involving the safety of high-value assets and the monitoring of high-accident areas. In this embodiment, by giving these data streams the highest priority, CPU and memory resources are prioritized for processing video data from these high-risk areas, thereby maximizing the effectiveness of security measures and ensuring the speed and quality of data processing and storage. By prioritizing the most critical data streams, emergency response time is reduced and the ability to handle critical incidents is improved. For example, consider a large commercial center. This center is not only a busy commercial area (high security sensitivity), but also a frequent accident area (high accident rate) due to its high pedestrian flow and complex traffic conditions. For example, a certain entrance and exit is near a valuable commodity area and a major traffic node. Therefore, the video data stream collected by the surveillance camera at this entrance and exit belongs to an area with high security sensitivity and a high accident rate. In this case, the camera data stream at this entrance and exit is configured as the highest priority, and CPU and memory resources are allocated to it first. This ensures that any abnormal events occurring at this entrance and exit are captured and stored immediately.

[0040] Example 2

[0041] See also Figure 1-Figure 3 This embodiment provides a data storage system based on a monitoring memory, including:

[0042] Monitoring memory, used to store data streams from different video sources;

[0043] An execution processor is communicatively connected to the monitoring memory and is used to execute the data storage method based on the monitoring memory described in any of the above embodiments. The method configures different priorities and data complexities for data streams from different video sources, communicates with front-end monitoring cameras of different video sources, receives data streams from different video sources captured by the front-end monitoring cameras, and includes an encoder that encodes the video data captured by the front-end monitoring cameras to generate the data streams. The encoder automatically identifies the received data streams from different video sources and dynamically adjusts the CPU resources and memory resources of the processing thread based on the configured priorities of the data streams. The received data streams from different video sources are stored in the monitoring memory, thereby reducing dependence on the storage server, reducing system costs, and preventing the server from becoming a single point of failure for the monitoring system. At the same time, because the video data is stored in the form of a stream, file fragmentation can be avoided, which helps to improve data reading and writing efficiency and the service life of the memory. In addition, dynamically adjusting resource allocation based on the priority of the data stream can ensure that high-priority data streams obtain sufficient processing resources, thereby ensuring their processing speed and storage quality. It should be noted that the related equipment of the data storage method based on the monitoring memory includes but is not limited to a data storage system based on the monitoring memory, a monitoring memory, and a front-end monitoring camera.

[0044] It should be pointed out that the above embodiments are only preferred specific implementation methods of the present invention, and the protection scope of the present invention is not limited thereto. Any changes or replacements that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A data storage method based on monitoring memory, characterized in that: The monitoring memory is communicatively connected to the execution processor, and the data storage method based on the monitoring memory includes: Configure different priorities and data complexity for data streams from different video sources; A front-end monitoring camera is communicatively connected with different video sources to receive data streams of different video sources collected by the front-end monitoring camera, wherein the front-end monitoring camera includes an encoder, and the encoder encodes the video data collected by the front-end monitoring camera to generate the data stream; Automatically identify the data streams received from different video sources, dynamically adjust the CPU resources and memory resources of the processing thread according to the configured priority and data complexity of the data streams, and store the data streams received from different video sources into the monitoring memory; Configuring different priorities for data streams from different video sources, including: configuring monitoring areas of front-end monitoring cameras in different areas, classifying the monitoring areas according to security sensitivity to obtain areas with different security sensitivities; configuring different priorities and data complexities for data streams of video sources in areas with different security sensitivities, and areas with higher security sensitivity correspond to data streams with higher priorities; configuring different priorities for data streams from different video sources, further including: configuring monitoring areas of front-end monitoring cameras in different areas, classifying the monitoring areas according to accident incidence rates to obtain areas with different accident incidence rates; configuring different priorities and data complexities for data streams of video sources in areas with different accident incidence rates, and areas with higher accident incidence rates correspond to data streams with higher priorities; When the data streams come from both the area with higher security sensitivity and the area with higher accident incidence rate, the highest priority is assigned to the regional overlapping data streams coming from both the area with higher security sensitivity and the area with higher accident incidence rate, the CPU resources and memory resources of the processing thread are adjusted preferentially, and the regional overlapping data streams are stored and processed; When configuring different data complexities for data streams from different video sources, different resolutions and frame rates are set for each video source according to the scene requirements of video surveillance. Different resolutions and frame rates correspond to different data complexities of data streams; different data complexities of data streams are automatically adjusted according to a preset schedule.

2. The data storage method based on monitoring memory according to claim 1, characterized in that: Configure different priorities for data streams from different video sources, including: Configuring a trigger event set for triggering video data collection, wherein the trigger event set includes trigger events of different urgency levels, and the trigger events of different urgency levels trigger a front-end surveillance camera to collect video data; Different priorities and data complexities are configured for data streams corresponding to video data collected by triggering events of different urgency, and more urgent triggering events correspond to data streams with higher priorities.

3. The data storage method based on monitoring memory according to claim 2, characterized in that: The triggering events of different emergency levels include fire detector alarm events, safety door illegal opening alarm events, crowd density sudden increase events and facility operation abnormality alarm events; the emergency levels of the fire detector alarm events, the safety door illegal opening alarm events, the crowd density sudden increase events and the facility operation abnormality alarm events decrease in sequence.

4. The data storage method based on monitoring memory according to any one of claims 1 to 3, characterized in that: The encoder converts the video signal collected by the front-end surveillance camera into video data, and compresses the video data using the H.264 video encoding standard or the H.265 video encoding standard to generate the data stream; the video data includes a data header and a data body; the data header records the ID or location code of the front-end surveillance camera, and the data body records the video content information.

5. The data storage method based on monitoring memory according to claim 4, characterized in that: Automatically identifying the received data streams of different video sources, including: After receiving the data streams from different video sources, parsing the data header of the video data in the data streams to obtain the ID or location code of the front-end surveillance camera; According to the ID or position code of the front-end monitoring camera obtained by parsing, the data streams received from different video sources are automatically identified.

6. The data storage method based on monitoring memory according to claim 1, characterized in that: Automatically identifying the received data streams of different video sources, including: Collect data streams from different video sources to extract color histogram features, gray-level co-occurrence matrix features, and inter-frame difference features to obtain training sample data; Using the training sample data to train the SVM model, so as to obtain a trained SVM model capable of distinguishing data streams from different video sources; The color histogram features, gray-level co-occurrence matrix features and inter-frame difference features are extracted from the data streams received from different video sources to obtain feature data to be identified, and the feature data to be identified is input into the trained SVM model. The trained SVM model automatically identifies the data streams received from different video sources.

7. The data storage method based on monitoring memory according to claim 1, characterized in that: Dynamically adjust the CPU resources and memory resources of the processing thread according to the configured priority and data complexity of the data flow, including: Define the priority of the i-th data stream as p i , the data complexity of the i-th data stream is d i , the amount of CPU resources required for the i-th data stream is r cpu (i) The sum of the CPU resources of all processing threads is R cpu , the amount of memory resources required for the i-th data stream is r mem (i) The sum of the memory resources of all processing threads is R mem ; The sum of CPU resources of all processing threads R cpu , the total memory resource R mem Satisfy the constraint formula: Set the resource requirement function f cpu and f mem , resource demand function f cpu and f mem Satisfies the following formula: f cpu (d,p)=a·e b·p +c·log(d+1) r cpu (i)=f cpu (d i ,p i );r mem (i)=f mem (d i ,p i ); Parameters a, b, c, x, y, z are determined through system analysis and actual measurement, and the resource demand function f cpu and f mem Dynamically calculate the required CPU and memory resources based on the priority and complexity of the data flow; Sort all data flows in descending order of priority pi, and allocate the required CPU resources r to each data flow in turn cpu (i) and the amount of memory resources r mem (i).

8. A data storage system based on monitoring memory, characterized in that: include: Monitoring memory for storing data streams from different video sources; An execution processor is communicatively connected to the monitoring memory, and is used to run the data storage method based on the monitoring memory as described in any one of claims 1-7.

Images