Special operation spatial-temporal change dynamic scheduling method and device based on edge calculation

Abstract

The invention provides a special operation spatio-temporal change dynamic scheduling method and device based on edge computing, and the method comprises the steps: collecting spatio-temporal information, automatically matching nearby candidate camera groups, judging the level of a special operation task, and evaluating whether the current residual computing power of an edge end can meet a computing task of the number of candidate camera groups of a current request or not; dynamically allocating camera resources to ensure that a key area is covered; and selecting a corresponding hidden danger identification algorithm from the algorithm library, and updating the selected algorithm to the edge all-in-one machine in real time through a wireless network. According to the method, cameras at corresponding positions can be automatically matched according to the occurrence time and location of the special operation, an algorithm corresponding to hidden danger identification is automatically loaded and updated to the edge all-in-one machine for possible hidden danger types, restarting is not needed, the algorithm takes effect immediately, the algorithm can be dynamically updated, and which cameras can be dynamically allocated to be applied. A user on the application layer can flexibly adjust configuration at any time without changing hardware parameters and program configuration.

Description

Technical Field

[0001] The present application relates to the technical field of edge computing job scheduling, and in particular to a method and device for dynamic scheduling of special operations with temporal and spatial changes based on edge computing. Background Art

[0002] There are all kinds of special operations in the construction and operation stages of large-scale projects. Special operations are the focus of risk prevention, but the time and location of occurrence are random. In addition, the identification of hidden dangers in special operations is different from the simple and conventional single video behavior identification. For example, the identification of hidden dangers in special operations such as electric fire and high altitude operations often has high algorithm complexity. The resources of the edge computing AI all-in-one machine are limited. Generally, one machine is connected to 5-10 cameras. The number of cameras for concurrent application of complex algorithms is limited and often fixed, which is very inflexible.

[0003] In engineering scenarios with hundreds or thousands of cameras, the number of integrated machines required to access all cameras is large and the investment is large. In fact, the number of special operations carried out at the same time is not large, resulting in idle edge AI computing power.

[0004] Therefore, it is an important research topic for technical personnel in this field to design an edge computing method based on the dynamic scheduling of video AI algorithms based on the temporal and spatial changes of special operations, automatically match the corresponding location cameras according to the time and place of the special operations, and automatically load and update the corresponding hidden danger identification algorithm into the edge all-in-one machine according to the possible types of hidden dangers. Summary of the invention

[0005] In view of this, it is necessary to provide a dynamic scheduling method for special operations with spatiotemporal changes based on edge computing, including: It can reduce invasiveness, accurately position ultrasound-guided positioning, reduce errors and complications, is easy to promote and apply, is relatively simple to operate, and is suitable for promotion and application in more medical places.

[0006] The technical solution of the present invention combines a deep learning algorithm with oral panoramic films, so that oral panoramic films can assist in the diagnosis of osteoporosis, and realize a medical image segmentation algorithm with readable charts. Different deep learning models are used for testing and performance comparison, and new modules are added without reducing the segmentation effect of the original network model. The feature graph of the convolutional neural network Res-U-Net model is processed by improving its structure and a displayable model training curve graph is added to make it readable. The model is optimized and improved by the loss function, and a readable measurement index is set, so that the image segmentation network model and the readability network model are alternately optimized during the training process. The obtained evaluation performance is similar to the results of a specific detection instrument, and the image segmentation effect is similar to expert labeling.

[0007] In a first aspect, an embodiment of the present application provides a method for dynamic scheduling of special operations based on edge computing based on spatiotemporal changes, the method comprising: S1: Collect time and space information, obtain the special operation location through GPS or Beidou positioning system, and obtain the current time through the time synchronization system; S2: According to the location and time of the special operation, the edge end automatically matches the candidate camera group nearby from the edge end camera database. The edge end determines the special operation task level and evaluates whether the current edge end remaining computing power can meet the computing task of the number of candidate camera groups currently requested based on the candidate camera group call request; S3: When the special operation task level is high, if the current remaining computing power at the edge end meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure coverage of key areas and provide computing power to complete the current computing task; if not, the computing resources with low available or unloaded operation task levels are scheduled through the task scheduling algorithm to provide computing power to complete the current computing task and update the resource list; S4: When the level of the special operation task is low, if the current remaining computing power at the edge end meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure full coverage of the operation area and provide computing power to complete the current computing task; if not, the remaining computing power is scheduled through the task scheduling algorithm, and the computing task of the number of candidate camera groups that meet the remaining computing power is called, and the camera resources are dynamically allocated to ensure coverage of the key operation area; S5: Detect and identify the operations of the camera group through the current remaining computing power of the edge end, select the corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and update the selected algorithm to the edge integrated machine in real time through the wireless network; S6: The updated algorithm takes effect immediately without restarting the edge device. The system automatically verifies the algorithm loading status to ensure normal operation. Users can adjust camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for users to monitor and adjust.

[0008] Optionally, in another implementation of the first aspect of the present invention, the S1: collecting spatiotemporal information, obtaining a special operation location through a GPS or Beidou positioning system, and obtaining a current time through a time synchronization system includes: Through the GPS or Beidou timing module, the device can obtain the current time and keep time synchronized with other devices; Receive satellite signals through the built-in GPS module of the camera device; The latitude and longitude information of the current location of the special operation site is calculated in real time by combining the base station and the satellite signal data through a real-time dynamic positioning algorithm, including: Receive satellite signals at a reference station with known precise coordinates and calculate the error between the observed value and the theoretical value; The camera device receives the same satellite signal and the differential correction number from the base station; it uses the differential correction number to correct its own observation value and calculates the high-precision real-time position.

[0009] Optionally, in another implementation of the first aspect of the present invention, the step of determining the level of the special operation task and evaluating whether the remaining computing power of the current edge terminal can meet the computing task of the number of candidate camera groups currently requested includes: Set task priorities based on task urgency, task complexity, task size, and task consequence impact; Evaluate the remaining computing power at the edge from three aspects: hardware resources, task requirements, and real-time status, including: Use system monitoring tools to obtain real-time CPU, GPU, and memory usage; Estimate the required computing power based on the number of cameras and task types; Compare the task requirements with the remaining computing power to determine whether they are met.

[0010] Optionally, in another implementation of the first aspect of the present invention, the task scheduling algorithm is: Each task is assigned a time slice of fixed length, and tasks are executed in turn in the order of first come first served; Dynamically adjust the time slice length according to system load and task characteristics; Determine the priority level of tasks and ensure that high-priority tasks can be completed faster; When a task's time slice is used up, the operating system suspends it and assigns the CPU to the next task.

[0011] Optionally, in another implementation of the first aspect of the present invention, dynamically adjusting the time slice length according to the system load and task characteristics includes: Build a dynamic time slice adjustment model based on system load, task priority, task execution time, and task response time, and dynamically adjust the time slice length based on system load and task characteristics; The model formula is: , in, is the length of the time slice after dynamic adjustment, is the basic time slice length, is the current system load, is the maximum load of the system, is the priority of the current task in the system, The highest priority in the system. represents the load adjustment factor, Indicates the priority adjustment factor.

[0012] Optionally, in another implementation of the first aspect of the present invention, the S5: detecting and identifying the operation of the camera group by using the current remaining computing power of the edge end, selecting a corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and updating the selected algorithm to the edge integrated machine in real time through the wireless network, including: S5.1, assessment of remaining computing power at the edge: real-time monitoring of the CPU, GPU, and memory resource usage of the edge machine, and calculation of the remaining computing power; dynamically adjusting the detection task allocation of the camera group according to the remaining computing power to ensure optimal utilization of computing power resources; setting a computing power usage threshold, and when the remaining computing power is lower than the threshold, triggering algorithm optimization or task priority adjustment; S5.2, camera group operation detection and hidden danger identification: use the edge all-in-one machine to analyze the video stream of the camera group in real time to detect abnormal behavior or potential hidden dangers in the operation scene; according to the detection results, the hidden dangers are divided into different types, including equipment failure, personnel violation, and environmental abnormality; select the identification algorithm corresponding to the hidden danger type from the preset algorithm library, and the identification algorithm includes target detection, behavior recognition, and anomaly detection; S5.3, Algorithm library and dynamic update: Maintain a multifunctional hidden danger identification algorithm library in the cloud or locally, covering identification models for various hidden danger scenarios; select the most suitable algorithm based on the hidden danger type and the remaining computing power on the edge; transmit the selected algorithm to the edge machine in real time through the wireless network to ensure timely update and deployment of the algorithm; S5.4, continuous optimization: Dynamically adjust algorithms and task allocation based on detection results and computing power changes, and optimize edge computing power allocation strategies.

[0013] Optionally, in another implementation of the first aspect of the present invention, the optimizing edge computing power allocation strategy includes: S5.4.1, Task priority and scheduling optimization: Classify tasks into high priority and low priority according to their importance and real-time nature; use intelligent scheduling algorithms to dynamically allocate resources based on task priority and remaining computing power; implement task migration and load balancing between edge nodes to avoid overloading of a single node; S5.4.2, Algorithm lightweighting and model optimization: Use lightweight deep learning models to replace complex models to reduce computing overhead; compress models through pruning, quantization, and knowledge distillation techniques to improve reasoning speed; dynamic model selection: dynamically switch models of different complexities based on remaining computing power and task requirements; S5.4.3, Edge-Cloud Collaboration: Offload some high-computing demand tasks to the cloud to reduce the pressure on the edge; complete data preprocessing at the edge to reduce the amount of data transmitted; achieve resource complementarity through the edge-cloud collaborative framework; S5.4.4, Resource reservation and elastic expansion: reserve computing resources for core tasks to ensure that critical tasks are not affected; when computing power is insufficient, dynamically expand edge nodes or request cloud resource support; use container technology to achieve rapid deployment and resource isolation.

[0014] In a second aspect, an embodiment of the present application provides a special operation spatiotemporal dynamic scheduling device based on edge computing, which is applied to the special operation spatiotemporal dynamic scheduling method based on edge computing as described in the first aspect, and is characterized by comprising: Space-time information acquisition module: collects space-time information, obtains special operation locations through GPS or Beidou positioning system, and obtains current time through time synchronization system; Camera preliminary matching module: According to the location and time of special operations, the edge end automatically matches the candidate camera group nearby from the edge end camera database. The edge end determines the special operation task level and evaluates whether the current edge end remaining computing power can meet the current requested number of candidate camera groups based on the candidate camera group call request; Remaining computing power evaluation module: When the special operation task level is high, if the current remaining computing power at the edge end meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure coverage of key areas and provide computing power to complete the current computing task; if not, the computing resources with low available or unloaded operation task levels are scheduled through the task scheduling algorithm to provide computing power to complete the current computing task and update the resource list; Camera dynamic adjustment module: When the level of the special operation task is low, if the current edge terminal remaining computing power meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure full coverage of the operation area and provide computing power to complete the current computing task; if it does not meet the requirements, the task scheduling algorithm is used to schedule the remaining computing power and call the computing task of the number of candidate camera groups that meet the remaining computing power, and dynamically allocate camera resources to ensure coverage of key operation areas; Hidden danger identification module: detects and identifies the operation of the camera group through the current remaining computing power of the edge end, selects the corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and updates the selected algorithm to the edge all-in-one machine in real time through the wireless network; Application layer configuration module: The updated algorithm takes effect immediately without restarting the edge machine. The system automatically verifies the algorithm loading status to ensure normal operation. Users can adjust camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for users to monitor and adjust.

[0015] In a third aspect, an embodiment of the present application provides an electronic device, including: processor; a memory for storing processor-executable instructions; Among them, the processor is configured to implement the dynamic scheduling method for special operations with temporal and spatial changes based on edge computing as described in the first aspect when executing the instructions.

[0016] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program, wherein the program instructs a device to execute the method for dynamic scheduling of special operations based on edge computing with respect to temporal and spatial changes as described in the first aspect.

[0017] The technical solution provided by the present invention provides a method and device for dynamic scheduling of special operations based on edge computing based on spatiotemporal changes. By collecting spatiotemporal information, the special operation location is obtained through GPS or Beidou positioning system, and the current time is obtained through time synchronization system; according to the special operation location and time, the candidate camera group nearby is automatically matched from the edge camera database, and the edge determines the special operation task level and evaluates whether the current edge remaining computing power can meet the computing task of the number of candidate camera groups currently requested based on the candidate camera group call request; in the case of a high level of special operation task, if the current edge remaining computing power meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure coverage of key areas and provide computing power to complete the current computing task; if not, the computing resources with low operation task level are scheduled or unloaded through the task scheduling algorithm to provide computing power to complete the current computing task, and update the resource. Source list; when the level of the special operation task is low, if the current remaining computing power at the edge end meets the calculation task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure full coverage of the operation area and provide computing power to complete the current calculation task; if it does not meet the requirements, the task scheduling algorithm is used to schedule the remaining computing power and call the calculation tasks of some candidate camera groups that meet the remaining computing power, and dynamically allocate camera resources to ensure coverage of key operation areas; the operation of the camera group is detected and identified through the current remaining computing power at the edge end, and the corresponding hidden danger identification algorithm is selected from the algorithm library according to the possible type of hidden dangers, and the selected algorithm is updated to the edge all-in-one machine in real time through the wireless network; the updated algorithm does not need to restart the edge all-in-one machine and takes effect immediately. The system automatically verifies the algorithm loading status to ensure normal operation. Users can adjust the camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for user monitoring and adjustment.

[0018] Beneficial effects:

[0019] (1) According to the time and place of the special operation, the corresponding location camera is automatically matched, and the corresponding hidden danger identification algorithm is automatically loaded and updated to the edge all-in-one machine according to the possible hidden danger type. It takes effect immediately without restarting.

[0020] (2) The algorithm can be updated dynamically, and the cameras to be used can also be dynamically allocated. At the application layer, users can flexibly adjust the configuration at any time without having to change the hardware parameters and program configuration. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a method for dynamic scheduling of special operations with spatiotemporal changes based on edge computing provided in one embodiment of the present application.

[0022] Figure 2A schematic diagram of a module of a dynamic scheduling device for special operations with spatiotemporal changes based on edge computing provided in one embodiment of the present application.

[0023] Figure 3 A schematic diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0025] It should be noted that, in the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the present application. The terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application.

[0026] It should be noted that, in the embodiments of the present application, words such as "first" and "second" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order. Features defined as "first" and "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way.

[0027] Based on the implementations in this application, all other implementations obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] Embodiment 1

[0029] The present application provides a method for dynamic scheduling of special operations with temporal and spatial changes based on edge computing, comprising the following steps.

[0030] S1: Collect time and space information, obtain the special operation location through GPS or Beidou positioning system, and obtain the current time through the time synchronization system.

[0031] Understandably, collecting spatiotemporal information is an important part of modern special operations, especially in scenarios that require precise positioning and real-time time synchronization.

[0032] Specifically, in this embodiment, the S1: collecting spatiotemporal information, obtaining the special operation location through GPS or Beidou positioning system, and obtaining the current time through the time synchronization system includes: Through the GPS or Beidou timing module, the device can obtain the current time and keep time synchronized with other devices; Receive satellite signals through the built-in GPS module of the camera device; The latitude and longitude information of the current location of the special operation site is calculated in real time by combining the base station and the satellite signal data through a real-time dynamic positioning algorithm, including: Receive satellite signals at a reference station with known precise coordinates and calculate the error between the observed value and the theoretical value; The camera device receives the same satellite signal and the differential correction number from the base station; it uses the differential correction number to correct its own observation value and calculates the high-precision real-time position.

[0033] It is understandable that GPS (Global Positioning System) and BeiDou Satellite Navigation System are currently widely used positioning technologies around the world, capable of providing high-precision geographic location information. The positioning system receives signals from multiple satellites, calculates the signal propagation time difference, and combines the satellite position information to determine the spatial coordinates of the receiving device, such as latitude and longitude, altitude, etc.

[0034] Special operations such as field exploration, emergency rescue, military operations, etc. require real-time acquisition of accurate coordinates of the work site. In complex terrain or areas with weak signals, the short message function of the Beidou system can assist in data transmission. On the application platform, terminal devices with integrated GPS / Beidou modules (such as cameras, mobile devices, etc.) can be used. Combined with map software or GIS system, the coordinate information can be visualized to facilitate rapid positioning of the work site.

[0035] The time synchronization system is the key to ensuring the accuracy and consistency of time information in special operations. GPS and Beidou satellites not only provide positioning services, but also send high-precision time signals through atomic clocks. After receiving satellite signals, the equipment can be synchronized to standard time (such as UTC time), and the error is usually in nanoseconds. In operations that require precise time records (such as event recording and data collection), the time synchronization system ensures the accuracy of timestamps. In multi-device collaborative operations, time synchronization can avoid operational errors caused by equipment time differences.

[0036] In this embodiment, a terminal device or a network time protocol (NTP) server supporting time synchronization is used. In an environment without satellite signals, time synchronization can be performed through a local area network or an external clock source.

[0037] Combining positioning information with time information can achieve more efficient operation management. The time and space information of the operation site can be recorded to facilitate subsequent data analysis and task review. In the collaborative operation of multiple devices, unified time and space information ensures consistent operation of each device. Real-time monitoring of the time and space information of operators ensures that they are in a safe area.

[0038] S2: According to the location and time of the special operation, the nearby candidate camera groups are automatically matched from the edge camera database. The edge determines the special operation task level and evaluates whether the current remaining computing power of the edge can meet the computing task of the number of candidate camera groups currently requested based on the candidate camera group call request.

[0039] Specifically, in this embodiment, the calculation task of determining the special operation task level and evaluating whether the current remaining computing power of the edge end can meet the number of candidate camera groups currently requested includes: Set task priorities based on task urgency, task complexity, task size, and task consequence impact; Evaluate the remaining computing power at the edge from three aspects: hardware resources, task requirements, and real-time status, including: Use system monitoring tools to obtain real-time CPU, GPU, and memory usage; Estimate the required computing power based on the number of cameras and task types; Compare the task requirements with the remaining computing power to determine whether they are met.

[0040] It is understandable that after inputting the longitude and latitude coordinates of the special operation site and the current time, the distance to all cameras in the edge camera database is calculated based on the longitude and latitude of the operation site. Set a distance threshold (such as 500 meters) to filter out the candidate camera group closest to the operation site. According to the current time, exclude cameras that are not in working hours or are turned off. Give priority to cameras that run in real time or support remote activation. According to the nature of the special operation, the tasks are divided into different levels (such as low, medium, and high). According to the task level, dynamically adjust the calling priority of the camera group.

[0041] Calculate the total computing power requirements of all cameras based on the number of candidate camera groups. Monitor the CPU, GPU, memory and other resource usage at the edge in real time. Compare the total computing power requirements with the remaining computing power at the edge to determine whether the task requirements are met. If the computing power is insufficient, dynamically adjust the number of candidate camera groups or reduce the task level.

[0042] Use edge computing platforms (such as KubeEdge and EdgeX Foundry) to manage cameras and computing resources. When computing power is sufficient, you can also optimize strategies, such as expanding the range of candidate camera groups to obtain more comprehensive monitoring data. Upgrade medium-level or low-level tasks to high-level tasks to improve analysis accuracy.

[0043] S3: When the level of the special operation task is high, if the current remaining computing power at the edge meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure coverage of key areas and provide computing power to complete the current computing task; if not, the task scheduling algorithm is used to schedule the surrounding available or unload computing resources with low operation task levels, provide computing power to complete the current computing task, and update the resource list.

[0044] Specifically, in this embodiment, when the remaining computing power at the edge is sufficient, camera resources are dynamically allocated to cover key areas. Specifically, according to the needs of special operations, camera resources are preferentially allocated to cover key areas (such as core operation areas and high-risk areas). The camera layout is optimized through spatial analysis algorithms (such as Voronoi diagrams) to ensure that there are no blind spots. Computing power can also be allocated, such as dynamically allocating the remaining computing power at the edge to candidate camera groups to support high-level tasks (such as real-time video analysis and multi-target tracking). Use load balancing technology to ensure that each camera has sufficient computing resources.

[0045] After adjusting the camera resources, update the resource list, update the edge resource list, and record the allocated computing power and camera status to facilitate subsequent task scheduling.

[0046] When the remaining computing power at the edge is not sufficient, the available resources in the surrounding area are called for scheduling through the task scheduling algorithm, such as retrieving available computing resources in the surrounding area (such as other edge nodes and cloud servers) through the edge computing network. Unload some high-level tasks to available resources to relieve local computing pressure. You can also release low-level task resources and suspend or reduce the computing power allocation of low-level tasks (such as simple monitoring) through the task scheduling algorithm. Reallocate the released computing power to high-level tasks to ensure their priority execution. You can also unload and collaborate tasks, unload some computing tasks to the cloud or other edge nodes, and complete high-level tasks through collaborative computing. Use distributed computing frameworks (such as Kubernetes and Apache Spark) to manage task allocation and execution. Finally, update the resource list, update the edge resource list, and record the computing power allocation and task status after scheduling to facilitate subsequent monitoring and adjustment.

[0047] S4: When the level of the special operation task is low, if the current remaining computing power at the edge end meets the calculation task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure full coverage of the operation area and provide computing power to complete the current calculation task; if it does not meet the requirements, the task scheduling algorithm is used to schedule the remaining computing power and call the calculation tasks of some candidate camera groups that meet the remaining computing power, and dynamically allocate camera resources to ensure coverage of key operation areas.

[0048] Specifically, in this embodiment, the task scheduling algorithm is: Each task is assigned a time slice of fixed length, and tasks are executed in turn in the order of first come first served; Dynamically adjust the time slice length according to system load and task characteristics; Determine the priority level of tasks and ensure that high-priority tasks can be completed faster; When a task's time slice is used up, the operating system suspends it and assigns the CPU to the next task.

[0049] It can be understood that the dynamically adjusting the time slice length according to the system load and task characteristics includes: Build a dynamic time slice adjustment model based on system load, task priority, task execution time, and task response time, and dynamically adjust the time slice length based on system load and task characteristics; The model formula is: , in, is the length of the time slice after dynamic adjustment, is the basic time slice length, is the current system load, is the maximum load of the system, is the priority of the current task in the system, The highest priority in the system. represents the load adjustment factor, Indicates the priority adjustment factor.

[0050] S5: Detect and identify the operations of the camera group through the current remaining computing power of the edge end. According to the possible hidden danger types, select the corresponding hidden danger identification algorithm from the algorithm library, and update the selected algorithm to the edge all-in-one machine in real time through the wireless network.

[0051] Specifically, in this embodiment, S5: detecting and identifying the operation of the camera group by using the current remaining computing power of the edge end, selecting a corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and updating the selected algorithm to the edge integrated machine in real time through the wireless network, including: S5.1, assessment of remaining computing power at the edge: real-time monitoring of the CPU, GPU, and memory resource usage of the edge machine, and calculation of the remaining computing power; dynamically adjusting the detection task allocation of the camera group according to the remaining computing power to ensure optimal utilization of computing power resources; setting a computing power usage threshold, and when the remaining computing power is lower than the threshold, triggering algorithm optimization or task priority adjustment; S5.2, camera group operation detection and hidden danger identification: use the edge all-in-one machine to analyze the video stream of the camera group in real time to detect abnormal behavior or potential hidden dangers in the operation scene; according to the detection results, the hidden dangers are divided into different types, including equipment failure, personnel violation, and environmental abnormality; select the identification algorithm corresponding to the hidden danger type from the preset algorithm library, and the identification algorithm includes target detection, behavior recognition, and anomaly detection; S5.3, Algorithm library and dynamic update: Maintain a multifunctional hidden danger identification algorithm library in the cloud or locally, covering identification models for various hidden danger scenarios; select the most suitable algorithm based on the hidden danger type and the remaining computing power on the edge; transmit the selected algorithm to the edge machine in real time through the wireless network to ensure timely update and deployment of the algorithm; S5.4, continuous optimization: Dynamically adjust algorithms and task allocation based on detection results and computing power changes, and optimize edge computing power allocation strategies.

[0052] The strategy for optimizing edge computing power allocation includes: S5.4.1, Task priority and scheduling optimization: Classify tasks into high priority and low priority according to their importance and real-time nature; use intelligent scheduling algorithms to dynamically allocate resources based on task priority and remaining computing power; implement task migration and load balancing between edge nodes to avoid overloading of a single node; S5.4.2, Algorithm lightweighting and model optimization: Use lightweight deep learning models to replace complex models to reduce computing overhead; compress models through pruning, quantization, and knowledge distillation techniques to improve reasoning speed; dynamic model selection: dynamically switch models of different complexities based on remaining computing power and task requirements; S5.4.3, Edge-Cloud Collaboration: Offload some high-computing demand tasks to the cloud to reduce the pressure on the edge; complete data preprocessing at the edge to reduce the amount of data transmitted; achieve resource complementarity through the edge-cloud collaborative framework; S5.4.4, Resource reservation and elastic expansion: reserve computing resources for core tasks to ensure that critical tasks are not affected; when computing power is insufficient, dynamically expand edge nodes or request cloud resource support; use container technology to achieve rapid deployment and resource isolation.

[0053] S6: The updated algorithm takes effect immediately without restarting the edge device. The system automatically verifies the algorithm loading status to ensure normal operation. Users can adjust camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for users to monitor and adjust.

[0054] Specifically, in this embodiment, algorithm updates and immediate effectiveness can be dynamically loaded using hot update technology, including the use of dynamic link libraries (DLLs) or modular design to support dynamic loading and unloading of algorithms at runtime. There is no need to restart the edge all-in-one to ensure continuous operation of the system. During the update process, the system automatically releases the memory occupied by the old algorithm, loads the new algorithm and allocates resources. Memory isolation technology is used to avoid memory leaks or conflicts during the update process. After the update is completed, the system automatically switches to the new algorithm to ensure immediate effectiveness. Through the status monitoring mechanism, verify whether the new algorithm is running normally.

[0055] It is understandable that the system has an automatic verification mechanism. After the update is completed, the system automatically performs a health check to verify the integrity and functionality of the algorithm. This includes the detection of indicators such as memory usage, computing performance, and output results. If the verification fails, the system automatically rolls back to the old algorithm and records an error log. An exception notification function is provided to prompt users of update failures and reasons. During the operation of the algorithm, the system continuously monitors its status to ensure long-term stable operation.

[0056] Embodiment 2 like Figure 2 As shown, the present application provides a device for dynamic scheduling of special operations based on edge computing and time-space changes, which is applied to the method for dynamic scheduling of special operations based on edge computing and time-space changes as described in Example 1, including: a time-space information acquisition module 11, a camera preliminary matching module 12, a remaining computing power evaluation module 13, a camera dynamic adjustment module 14, a hidden danger identification module 15, and an application layer configuration module 16.

[0057] It can be understood that, in this embodiment, the spatiotemporal information acquisition module 11 is used to collect spatiotemporal information, obtain the special operation location through the GPS or Beidou positioning system, and obtain the current time through the time synchronization system.

[0058] It can be understood that in this embodiment, the camera preliminary matching module 12 is used to automatically match nearby candidate camera groups from the edge camera database according to the location and time of the special operation. The edge determines the special operation task level and evaluates whether the current remaining computing power of the edge can meet the computing task of the currently requested number of candidate camera groups based on the candidate camera group call request.

[0059] It can be understood that in this embodiment, the remaining computing power evaluation module 13 is used to dynamically allocate camera resources to ensure coverage of key areas and provide computing power to complete the current computing task if the current edge-end remaining computing power meets the computing task of the number of candidate camera groups currently requested when the special operation task level is high; if it does not meet the requirements, the task scheduling algorithm is used to schedule the surrounding available or unload computing resources with low operation task levels to provide computing power to complete the current computing task and update the resource list.

[0060] It can be understood that in this embodiment, the camera dynamic adjustment module 14 is used to dynamically allocate camera resources to ensure full coverage of the operation area and provide computing power to complete the current computing task if the current edge-end remaining computing power meets the computing task of the currently requested number of candidate camera groups when the level of the special operation task is low; if it does not meet the requirements, the task scheduling algorithm is used to schedule the remaining computing power and call the computing task of part of the number of candidate camera groups that meet the remaining computing power, and dynamically allocate camera resources to ensure coverage of key operation areas.

[0061] It can be understood that in this embodiment, the hidden danger identification module 15 is used to detect and identify the operation of the camera group through the current remaining computing power of the edge end, select the corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and update the selected algorithm to the edge all-in-one machine in real time through the wireless network.

[0062] It can be understood that in this embodiment, the application layer configuration module 16 is used for the updated algorithm, and there is no need to restart the edge all-in-one machine, and the updated algorithm takes effect immediately. The system automatically verifies the algorithm loading status to ensure normal operation. The user can adjust the camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for user monitoring and adjustment.

[0063] Figure 3 This is an electronic device provided by an embodiment of the present application. Figure 3 As shown, the electronic device includes at least the following parts: a processor 101 and a memory 100 , a communication interface 103 , and a bus 102 .

[0064] In the embodiment of the present application, the memory 100 is used to store instructions executable by the processor 101, and the processor 101 is configured to implement the following when executing the instructions: Figure 2 The equipment module of the dynamic scheduling device for special operations with spatiotemporal changes based on edge computing is shown.

[0065] In an embodiment of the present application, a computer-readable storage medium includes instructions, and the instructions instruct a device to execute the method of the first aspect. For example, the instructions instruct the device to execute Figure 1The process steps in the figure show the dynamic scheduling method of special operations with temporal and spatial changes based on edge computing.

[0066] The program running in the electronic device involved in one embodiment of the present application can be a program (a program that enables a computer to function) that controls a central processing unit (CPU) and the like to realize the functions of the above-mentioned implementation method involved in one scheme of the present invention. Then, the information processed by these devices is temporarily stored in a random access memory (RAM) during its processing, and then stored in various ROMs such as a read-only memory (Flash ROM) and a hard disk drive (HDD), and is read, corrected, and written by the CPU as needed.

[0067] It should be noted that a part of the electronic device of the above embodiment can also be realized by a computer. In this case, a program for realizing the control function can be recorded in a computer-readable recording medium, and the program recorded in the recording medium can be read into the computer and executed to realize it.

[0068] It should be noted that the "computer" mentioned here refers to a computer built into an electronic device, and uses a computer including hardware such as an OS and peripheral devices. In addition, "computer-readable recording medium" refers to removable media such as floppy disks, magneto-optical disks, ROMs, CD-ROMs, and storage devices such as hard disks built into computers.

[0069] Furthermore, "computer-readable recording media" may include: media that dynamically store programs for a short period of time, such as a communication line when the program is transmitted via a network such as the Internet or a communication line such as a telephone line; and media that store programs for a fixed period of time, such as a volatile memory inside a computer that serves as a server or client in this case. In addition, the above-mentioned program may be a program for realizing a part of the above-mentioned functions, and may also be a program that can realize the above-mentioned functions by combining with a program already recorded in a computer.

[0070] In addition, the electronic device in the above-mentioned embodiment can also be implemented as a collection (device group) composed of multiple devices. Each device constituting the device group can have a part or all of the functions or functional blocks of the electronic device in the above-mentioned embodiment. As a device group, it is sufficient to have all the functions or functional blocks of the electronic device.

[0071] Those skilled in the art should recognize that the above embodiments are only used to illustrate the present application and are not intended to be limiting of the present application. As long as they are within the spirit and scope of the present application, appropriate changes and modifications to the above embodiments are within the scope of protection claimed in the present application.

Claims

1. A method for dynamic scheduling of special operations based on edge computing, characterized in that: The method comprises: S1: Collect time and space information, obtain the special operation location through GPS or Beidou positioning system, and obtain the current time through the time synchronization system; S2: According to the location and time of the special operation, the edge end automatically matches the candidate camera group nearby from the edge end camera database. The edge end determines the special operation task level and evaluates whether the current edge end remaining computing power can meet the computing task of the number of candidate camera groups currently requested based on the candidate camera group call request; S3: When the special operation task level is high, if the current remaining computing power at the edge end meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure coverage of key areas and provide computing power to complete the current computing task; if not, the computing resources with low available or unloaded operation task levels are scheduled through the task scheduling algorithm to provide computing power to complete the current computing task and update the resource list; S4: When the level of the special operation task is low, if the current remaining computing power at the edge end meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure full coverage of the operation area and provide computing power to complete the current computing task; if not, the remaining computing power is scheduled through the task scheduling algorithm, and the computing task of the number of candidate camera groups that meet the remaining computing power is called, and the camera resources are dynamically allocated to ensure coverage of the key operation area; S5: Detect and identify the operations of the camera group through the current remaining computing power of the edge end, select the corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and update the selected algorithm to the edge integrated machine in real time through the wireless network; S6: The updated algorithm takes effect immediately without restarting the edge device. The system automatically verifies the algorithm loading status to ensure normal operation. Users can adjust camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for users to monitor and adjust.

2. The method for dynamic scheduling of special operations based on edge computing in time and space according to claim 1 is characterized in that: S1: collecting time and space information, obtaining the special operation location through GPS or Beidou positioning system, and obtaining the current time through the time synchronization system, including: Through the GPS or Beidou timing module, the device can obtain the current time and keep time synchronized with other devices; Receive satellite signals through the built-in GPS module of the camera device; The latitude and longitude information of the current location of the special operation site is calculated in real time by combining the base station and the satellite signal data through a real-time dynamic positioning algorithm, including: Receive satellite signals at a reference station with known precise coordinates and calculate the error between the observed value and the theoretical value; The camera device receives the same satellite signal and the differential correction number from the base station; it uses the differential correction number to correct its own observation value and calculates the high-precision real-time position.

3. The method for dynamic scheduling of special operations based on edge computing according to claim 2 is characterized in that: The computing task of determining the special operation task level and evaluating whether the current remaining computing power of the edge end can meet the number of candidate camera groups currently requested includes: Set task priorities based on task urgency, task complexity, task size, and task consequence impact; Evaluate the remaining computing power at the edge from three aspects: hardware resources, task requirements, and real-time status, including: Use system monitoring tools to obtain real-time CPU, GPU, and memory usage; Estimate the required computing power based on the number of cameras and task types; Compare the task requirements with the remaining computing power to determine whether they are met.

4. The method for dynamic scheduling of special operations based on edge computing according to claim 3 is characterized in that: The task scheduling algorithm is: Each task is assigned a time slice of fixed length, and tasks are executed in turn in the order of first come first served; Dynamically adjust the time slice length according to system load and task characteristics; Determine the priority level of tasks and ensure that high-priority tasks can be completed faster; When a task's time slice is used up, the operating system suspends it and assigns the CPU to the next task.

5. The method for dynamic scheduling of special operations based on edge computing according to claim 4 is characterized in that: The dynamically adjusting the time slice length according to the system load and task characteristics includes: Build a dynamic time slice adjustment model based on system load, task priority, task execution time, and task response time, and dynamically adjust the time slice length based on system load and task characteristics; The model formula is: ， in, is the length of the time slice after dynamic adjustment, is the basic time slice length, is the current system load, is the maximum load of the system, is the priority of the current task in the system, The highest priority in the system. represents the load adjustment factor, Indicates the priority adjustment factor.

6. The method for dynamic scheduling of special operations based on edge computing according to claim 5 is characterized in that: S5: detecting and identifying the operation of the camera group by using the current remaining computing power of the edge end, selecting a corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and updating the selected algorithm to the edge integrated machine in real time through the wireless network, including: S5.1, assessment of remaining computing power at the edge: real-time monitoring of the CPU, GPU, and memory resource usage of the edge machine, and calculation of the remaining computing power; dynamically adjusting the detection task allocation of the camera group according to the remaining computing power to ensure optimal utilization of computing power resources; setting a computing power usage threshold, and when the remaining computing power is lower than the threshold, triggering algorithm optimization or task priority adjustment; S5.2, camera group operation detection and hidden danger identification: use the edge all-in-one machine to analyze the video stream of the camera group in real time to detect abnormal behavior or potential hidden dangers in the operation scene; according to the detection results, the hidden dangers are divided into different types, including equipment failure, personnel violation, and environmental abnormality; select the identification algorithm corresponding to the hidden danger type from the preset algorithm library, and the identification algorithm includes target detection, behavior recognition, and anomaly detection; S5.3, Algorithm library and dynamic update: Maintain a multifunctional hidden danger identification algorithm library in the cloud or locally, covering identification models for various hidden danger scenarios; select the most suitable algorithm based on the hidden danger type and the remaining computing power on the edge; transmit the selected algorithm to the edge machine in real time through the wireless network to ensure timely update and deployment of the algorithm; S5.4, continuous optimization: Dynamically adjust algorithms and task allocation based on detection results and computing power changes, and optimize edge computing power allocation strategies.

7. The method for dynamic scheduling of special operations based on edge computing according to claim 6 is characterized in that: The strategy for optimizing edge computing power allocation includes: S5.4.1, Task priority and scheduling optimization: Classify tasks into high priority and low priority according to their importance and real-time nature; use intelligent scheduling algorithms to dynamically allocate resources based on task priority and remaining computing power; implement task migration and load balancing between edge nodes to avoid overloading of a single node; S5.4.2, Algorithm lightweighting and model optimization: Use lightweight deep learning models to replace complex models to reduce computing overhead; compress models through pruning, quantization, and knowledge distillation techniques to improve reasoning speed; dynamic model selection: dynamically switch models of different complexities based on remaining computing power and task requirements; S5.4.3, Edge-Cloud Collaboration: Offload some high-computing demand tasks to the cloud to reduce the pressure on the edge; complete data preprocessing at the edge to reduce the amount of data transmitted; achieve resource complementarity through the edge-cloud collaborative framework; S5.4.4, Resource reservation and elastic expansion: reserve computing resources for core tasks to ensure that critical tasks are not affected; when computing power is insufficient, dynamically expand edge nodes or request cloud resource support; use container technology to achieve rapid deployment and resource isolation.

8. A device for dynamic scheduling of special operations based on edge computing, applied to the method for dynamic scheduling of special operations based on edge computing according to any one of claims 1 to 7, characterized in that: include: Space-time information acquisition module: collects space-time information, obtains special operation locations through GPS or Beidou positioning system, and obtains current time through time synchronization system; Camera preliminary matching module: According to the location and time of special operations, the edge end automatically matches the candidate camera group nearby from the edge end camera database. The edge end determines the special operation task level and evaluates whether the current edge end remaining computing power can meet the current requested number of candidate camera groups based on the candidate camera group call request; Remaining computing power evaluation module: When the special operation task level is high, if the current remaining computing power at the edge end meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure coverage of key areas and provide computing power to complete the current computing task; if not, the computing resources with low available or unloaded operation task levels are scheduled through the task scheduling algorithm to provide computing power to complete the current computing task and update the resource list; Camera dynamic adjustment module: when the special operation task level is low, if the current edge end remaining computing power meets the computing task of the number of candidate camera groups currently requested, the camera resources are dynamically allocated to ensure full coverage of the operation area and provide computing power to complete the current computing task; If it is not satisfied, the task scheduling algorithm is used to schedule the remaining computing power, call the computing tasks of some candidate camera groups that meet the remaining computing power, and dynamically allocate camera resources to ensure coverage of key operation areas; Hidden danger identification module: detects and identifies the operation of the camera group through the current remaining computing power of the edge end, selects the corresponding hidden danger identification algorithm from the algorithm library according to the possible hidden danger type, and updates the selected algorithm to the edge all-in-one machine in real time through the wireless network; Application layer configuration module: The updated algorithm takes effect immediately without restarting the edge machine. The system automatically verifies the algorithm loading status to ensure normal operation. Users can adjust camera allocation and algorithm configuration at any time through the application layer interface. The system provides a visual interface for users to monitor and adjust.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the dynamic scheduling method for special operations with spatiotemporal changes based on edge computing as described in any one of claims 1 to 7 when executing the instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, and the program instructs the device to execute the method for dynamic scheduling of special operations with temporal and spatial changes based on edge computing as described in any one of claims 1 to 7.

Images