Video edge gateway AI algorithm task generation method based on large model

Abstract

The invention discloses a video edge gateway AI algorithm task generation method based on a large model, and the method comprises the following steps: inputting the requirements of an AI algorithm task through a man-machine interaction input box, the requirements comprising at least one of a camera address, a frame extraction frequency, a frame resolution, and an algorithm name; natural language processing is carried out on input AI algorithm task requirements through the large model, structured data is generated, and the structured data comprises task parameter configuration information; transmitting the structured data to a background application service; the background application service analyzes the structured data, extracts task parameter configuration information, and generates an AI algorithm task according to the parameter configuration information; and issuing the AI algorithm task to a video edge gateway for execution.

Description

Technical Field

[0001] The present invention is applied to the field of artificial intelligence, and specifically relates to a method for generating AI algorithm tasks of a video edge gateway based on a large model. Background Art

[0002] With the rapid development of artificial intelligence technology, video edge computing has been widely applied in fields such as security monitoring, intelligent transportation, and industrial inspection. As the core device connecting cameras and cloud computing, the video edge gateway can achieve real-time video analysis by deploying AI algorithms, such as object detection, face recognition, behavior analysis, etc. However, there are still many problems in the configuration and management of AI algorithm tasks for existing video edge gateways:

[0003] Complex task configuration:

[0004] Traditional video edge gateways require operation and maintenance personnel to manually configure various parameters of AI algorithm tasks, such as camera addresses, frame extraction frequencies, frame resolutions, algorithm types, etc. This configuration method is not only cumbersome but also prone to task configuration errors due to human mistakes, affecting the task execution effect.

[0005] Lack of intelligent support:

[0006] Existing systems usually rely on fixed configuration templates or scripts and cannot automatically generate task configurations according to the natural language descriptions of operation and maintenance personnel. This lack of intelligent support limits the flexibility and usability of the system, especially for non-technical personnel, and the operation threshold is relatively high.

[0007] Insufficient task management and monitoring:

[0008] After task generation, existing systems lack a real-time monitoring and feedback mechanism for task execution status. It is difficult for operation and maintenance personnel to timely understand the task execution progress and results and cannot quickly adjust task parameters to optimize performance.

[0009] Limited scalability and compatibility:

[0010] Existing systems are usually designed for specific AI algorithms or hardware devices and lack compatibility support for multiple algorithms and multiple devices. This limitation makes it difficult for the system to adapt to rapidly changing technical requirements and business scenarios. Summary of the Invention

[0011] The technical problem to be solved by the present invention is to provide a method for generating AI algorithm tasks of a video edge gateway based on a large model in view of the deficiencies of the prior art.

[0012] To solve the above technical problem, a method for generating AI algorithm tasks of a video edge gateway based on a large model of the present invention includes the following steps:

[0013] Input the requirements of the AI algorithm task through the human-computer interaction input box, and the requirements include at least one of the camera address, frame extraction frequency, frame resolution, and algorithm name;

[0014] Perform natural language processing on the input AI algorithm task requirements through a large model to generate structured data, and the structured data includes task parameter configuration information;

[0015] Transmit the structured data to the background application service;

[0016] The background application service parses the structured data, extracts the task parameter configuration information, and generates an AI algorithm task according to the parameter configuration information;

[0017] Send the AI algorithm task to the video edge gateway for execution.

[0018] As a possible implementation, further, the human-computer interaction input box supports multi-line text input and provides a real-time verification function to ensure that the input content meets the preset format requirements.

[0019] As a possible implementation, further, the structured data is in JSON or XML format and includes at least one of the following fields: camera address, frame extraction frequency, frame resolution, algorithm name.

[0020] As a possible implementation, further, the background application service generates a specific AI algorithm task by parsing the structured data, including the following steps:

[0021] Determine the target video source according to the camera address;

[0022] Configure video processing parameters according to the frame extraction frequency and frame resolution;

[0023] Select the corresponding AI algorithm model according to the algorithm name;

[0024] Send the configured AI algorithm task to the video edge gateway.

[0025] As a possible implementation, further, the video edge gateway supports multiple AI algorithm tasks, including but not limited to object detection, face recognition, and behavior analysis.

[0026] As a possible implementation, further, the method further includes a task status monitoring and result feedback function, specifically including:

[0027] The video edge gateway reports the task execution status in real time;

[0028] The background application service feeds back the task execution status to the front-end interface;

[0029] Operation and maintenance personnel view the task execution progress and results through the front-end interface.

[0030] As a possible implementation, further, the method further includes a task parameter optimization function, specifically including:

[0031] The background application service automatically adjusts the task parameter configuration according to the task execution result;

[0032] Feed back the optimized task parameter configuration to the large model for parameter recommendation in subsequent task generation.

[0033] As a possible implementation, further, the method further includes a task log recording function, specifically including:

[0034] Record the full process logs of task generation, distribution, and execution;

[0035] Provide log query and analysis functions for fault troubleshooting and performance optimization.

[0036] A video edge gateway AI algorithm task generation system based on a large model, including:

[0037] A human-computer dialogue input module for receiving the AI algorithm task requirements input by operation and maintenance personnel;

[0038] A large model service module for performing natural language processing on the input task requirements to generate structured data;

[0039] A background application service module for parsing the structured data and generating AI algorithm tasks;

[0040] A video edge gateway module for executing the generated AI algorithm tasks.

[0041] The present invention adopts the above technical solutions and has the following beneficial effects:

[0042] Simplify the task configuration process: By introducing large model technology, operation and maintenance personnel only need to describe the AI algorithm task requirements in natural language, and the system can automatically generate structured data and configure task parameters, significantly reducing the complexity of task configuration and the operation threshold.

[0043] Improve the usability of the system: The present invention supports natural language interaction, and operation and maintenance personnel do not need to have professional algorithm configuration knowledge, especially suitable for non-technical personnel to use, greatly improving the usability and popularity of the system.

[0044] Enhance the flexibility of task management: Through the automatic generation and parsing of structured data, the system can quickly adapt to different task requirements, support the flexible configuration of multiple AI algorithms and hardware devices, and improve the compatibility and scalability of the system.

[0045] Implement full - process monitoring of tasks: The present invention provides task status monitoring and result feedback functions. Operation and maintenance personnel can view the task execution progress and results in real time, which is convenient for quickly adjusting task parameters to optimize performance. Description of the Drawings

[0046] The following further elaborates on the present invention in detail in conjunction with the drawings and specific embodiments:

[0047] Figure 1 It is a schematic flowchart of an embodiment of the present invention. Specific Embodiments

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention.

[0049] Embodiment 1

[0050] A method for generating an AI algorithm task of a video edge gateway based on a large model includes the following steps:

[0051] Input the requirements of the AI algorithm task through a human - machine interaction input box, and the requirements include at least one of the camera address, frame extraction frequency, frame resolution, and algorithm name;

[0052] Perform natural language processing on the input requirements of the AI algorithm task through a large model to generate structured data, and the structured data includes task parameter configuration information;

[0053] Transmit the structured data to the background application service;

[0054] The background application service parses the structured data, extracts the task parameter configuration information, and generates an AI algorithm task according to the parameter configuration information;

[0055] Send the AI algorithm task to the video edge gateway for execution.

[0056] Among them, the human - machine interaction input box supports multi - line text input and provides a real - time verification function to ensure that the input content meets the preset format requirements.

[0057] Among them, the structured data is in JSON or XML format and includes at least one of the following fields: camera address, frame extraction frequency, frame resolution, and algorithm name.

[0058] Among them, the background application service generates a specific AI algorithm task by parsing the structured data, including the following steps:

[0059] Determine the target video source according to the camera address;

[0060] Configure video processing parameters according to the frame extraction frequency and frame resolution;

[0061] Select the corresponding AI algorithm model according to the algorithm name;

[0062] Send the configured AI algorithm task to the video edge gateway.

[0063] Among them, the video edge gateway supports multiple AI algorithm tasks, including but not limited to object detection, face recognition, and behavior analysis.

[0064] Among them, the method also includes a task status monitoring and result feedback function, specifically including:

[0065] The video edge gateway reports the task execution status in real time;

[0066] The background application service feeds back the task execution status to the front-end interface;

[0067] The operation and maintenance personnel view the task execution progress and results through the front-end interface.

[0068] Among them, the method also includes a task parameter optimization function, specifically including:

[0069] The background application service automatically adjusts the task parameter configuration according to the task execution results;

[0070] Feed back the optimized task parameter configuration to the large model for parameter recommendation in subsequent task generation.

[0071] Among them, the method also includes a task log recording function, specifically including:

[0072] Record the full process logs of task generation, distribution, and execution;

[0073] Provide log query and analysis functions for fault troubleshooting and performance optimization.

[0074] A video edge gateway AI algorithm task generation system based on a large model, including:

[0075] A human-computer dialogue input module for receiving the AI algorithm task requirements input by the operation and maintenance personnel;

[0076] A large model service module for performing natural language processing on the input task requirements to generate structured data;

[0077] A background application service module for parsing the structured data and generating AI algorithm tasks;

[0078] A video edge gateway module for executing the generated AI algorithm tasks.

[0079] Embodiment 2

[0080] As Figure 1 shown, the present invention provides a method for arranging the literal description of the video edge gateway algorithm application based on the privatized deployment of large models, and the process is as follows:

[0081] S1 Create a human-computer dialogue input box for the video edge gateway;

[0082] S2 Input requirements related to the AI algorithm task, such as descriptions of camera address, frame extraction frequency, frame resolution, algorithm name, etc.;

[0083] S3 The large model analyzes the description of the algorithm task and generates a structured data format file to forward to the background application service;

[0084] S4 The background application service parses the structured data, extracts the setting values of each parameter configuration, and creates an algorithm task.

[0085] Example 3

[0086] Step 1: Task input

[0087] The operation and maintenance personnel input the following task requirements in the human-computer dialogue input box on the front-end interface:

[0088] Analyze the video stream of the camera at the east gate of the park (address: 192.168.1.101), with a frame extraction frequency of 5 frames per second, a frame resolution of 1080p, and use the object detection algorithm (YOLOv5) to detect vehicles and pedestrians.

[0089] Step 2: Large model processing

[0090] The large model service module receives the input text, performs natural language processing, and identifies the following key parameters:

[0091] Camera address: 192.168.1.101

[0092] Frame extraction frequency: 5fps

[0093] Frame resolution: 1080p

[0094] Algorithm name: YOLOv5

[0095] Detection targets: vehicles, pedestrians

[0096] Generate the structured data (JSON format) as follows:

[0097] {

[0098] "camera_address":"192.168.1.101",

[0099] "frame_rate":5,

[0100] "resolution": "1080p",

[0101] "algorithm_name": "YOLOv5",

[0102] "targets": ["vehicle", "person"]

[0103] }

[0104] Step 3: Task Generation

[0105] The background application service module receives the structured data, parses it, and generates an AI algorithm task: Determine the video source based on the camera address (192.168.1.101).

[0106] Configure the frame extraction frequency to 5fps and the frame resolution to 1080p.

[0107] Select the YOLOv5 algorithm model and set the detection targets to vehicles and pedestrians. The generated task configuration file is as follows:

[0108] {

[0109] "task_id": "task_001",

[0110] "camera_address": "192.168.1.101",

[0111] "frame_rate": 5,

[0112] "resolution": "1080p",

[0113] "algorithm": "YOLOv5",

[0114] "parameters": {

[0115] "targets": ["vehicle", "person"]

[0116] }

[0117] }

[0118] Step 4: Task Distribution and Execution

[0119] The background application service distributes the task configuration file to the video edge gateway.

[0120] After receiving the task, the video edge gateway starts the YOLOv5 algorithm and performs real-time analysis on the video stream of the specified camera to detect vehicles and pedestrians.

[0121] Step 5: Task Monitoring and Feedback

[0122] The video edge gateway reports the task execution status (such as the number of processed frames, detection results) to the background application service in real time.

[0123] The background application service feeds back the task status to the front-end interface, and the operation and maintenance personnel can view the task progress and results in real time.

[0124] For example, the front-end interface displays:

[0125] Task ID: task_001

[0126] Status: Running

[0127] Number of processed frames: 1200

[0128] Detection results: Vehicles: 25, Pedestrians: 10

[0129] Step 6: Task Optimization and Log Recording

[0130] The background application service automatically adjusts the frame extraction frequency to 10fps according to the task execution results to improve the detection accuracy.

[0131] The system records the full process logs of task generation, distribution, and execution for subsequent analysis and optimization.

[0132] The above are the embodiments of the present invention. For those of ordinary skill in the art, according to the teachings of the present invention, any equivalent changes, modifications, substitutions, and variations made within the scope of the patent application of the present invention without departing from the principles and spirit of the present invention shall fall within the scope of the present invention.

Claims

1. A method for generating AI algorithm tasks of a video edge gateway based on a large model, characterized in that It includes the following steps: Input the requirements of the AI algorithm task through the human-computer interaction input box, where the requirements include at least one of the camera address, frame extraction frequency, frame resolution, and algorithm name; Perform natural language processing on the input AI algorithm task requirements through a large model to generate structured data, where the structured data includes task parameter configuration information; Transmit the structured data to the background application service; The background application service parses the structured data, extracts the task parameter configuration information, and generates an AI algorithm task based on the parameter configuration information; Send the AI algorithm task to the video edge gateway for execution.

2. The method for generating AI algorithm tasks of a video edge gateway based on a large model according to claim 1, wherein: The human-computer interaction input box supports multi-line text input and provides a real-time verification function to ensure that the input content meets the preset format requirements.

3. The AI algorithm task generation method for a video edge gateway based on a large model according to claim 1, wherein: The structured data is in JSON or XML format and includes at least one of the following fields: camera address, frame extraction frequency, frame resolution, and algorithm name.

4. The method for generating an AI algorithm task of a video edge gateway based on a large model according to claim 1, wherein: The background application service generates a specific AI algorithm task by parsing the structured data, including the following steps: Determine the target video source according to the camera address; Configure video processing parameters according to the frame extraction frequency and frame resolution; Select the corresponding AI algorithm model according to the algorithm name; Send the configured AI algorithm task to the video edge gateway.

5. The AI algorithm task generation method for a video edge gateway based on a large model according to claim 1, wherein: The video edge gateway supports multiple AI algorithm tasks, including but not limited to object detection, face recognition, and behavior analysis.

6. A method for generating AI algorithm tasks of a video edge gateway based on a large model according to claim 1, characterized in that: The method also includes a task status monitoring and result feedback function, specifically including: The video edge gateway reports the task execution status in real time; The background application service feeds back the task execution status to the front-end interface; The operation and maintenance personnel view the task execution progress and results through the front-end interface.

7. The method for generating an AI algorithm task of a video edge gateway based on a large model according to claim 1, characterized in that: The method also includes a task parameter optimization function, specifically including: The background application service automatically adjusts the task parameter configuration according to the task execution results; Feed back the optimized task parameter configuration to the large model for parameter recommendation in subsequent task generation.

8. A method for generating AI algorithm tasks of a video edge gateway based on a large model according to claim 1, characterized in that: The method also includes a task log recording function, specifically including: Record the full process logs of task generation, distribution, and execution; Provide log query and analysis functions for fault troubleshooting and performance optimization.

9. A video edge gateway AI algorithm task generation system based on a large model, characterized in that, It includes: A human-computer dialogue input module for receiving the requirements of the AI algorithm task input by the operation and maintenance personnel; A large model service module for performing natural language processing on the input task requirements to generate structured data; A background application service module for parsing the structured data and generating an AI algorithm task; A video edge gateway module for executing the generated AI algorithm task.

Images