Multi-modal large model-based optimization method for edge computing equipment in reservoir intelligent measurement

Abstract

The invention discloses a method for optimizing edge computing equipment based on a multi-modal large model in reservoir intelligent measurement, and the method achieves the comprehensive and real-time monitoring of a reservoir water source protection region through the combination of the multi-modal large model and an edge computing technology. A high-performance edge computing device is adopted as a platform, a multi-mode large model is deployed to perform real-time analysis on a monitoring video, and illegal behaviors such as burning prohibition, cultivation, water surface floating objects, catering business point garbage and personnel invasion are identified; meanwhile, a small model is designed for specific illegal behaviors for special identification, and the identification precision and efficiency are improved; the edge computing platform supports access of Internet of Things equipment, and integrated analysis of multi-dimensional data such as meteorological data and water quality data is achieved. According to the method, a manager is notified in time through an early warning mechanism, remote access and management are supported, and data transmission security is ensured; the whole scheme is efficient and accurate, is suitable for various reservoir intelligent monitoring scenes, and provides powerful support for reservoir management and protection.

Description

Technical Field

[0001] The present invention relates to the fields of computer technology, edge computing and artificial intelligence, and specifically to an optimization method for edge computing equipment in intelligent measurement of reservoirs based on a multimodal large model. By combining the edge computing equipment with the multimodal large model, comprehensive and real-time monitoring of a reservoir water source protection area can be achieved. Background Art

[0002] Traditional monitoring of reservoir water source protection areas mainly relies on manual inspections and video surveillance, which has the following disadvantages:

[0003] 1. Low efficiency: The scope of manual inspection is limited, making it difficult to achieve all-weather and all-round monitoring;

[0004] 2. High missed detection rate: Video surveillance has blind spots and it is difficult to identify subtle violations;

[0005] 3. Untimely response: After the problem is discovered, it needs to be manually confirmed and processed, and the response speed is slow;

[0006] In view of the above shortcomings, the present invention aims to provide an intelligent analysis algorithm service based on edge computing to achieve real-time monitoring and early warning of violations such as prohibition of burning, farming, floating objects on the water surface, garbage from catering outlets, and human intrusion in reservoir water source protection areas. Summary of the invention

[0007] In order to solve the problems existing in the prior art, the present invention provides an optimization method for edge computing equipment in reservoir intelligent measurement based on a multimodal large model, which is characterized by:

[0008] The following steps are involved:

[0009] a) Deploy edge computing platform: Use China Tower Smart Edge Gateway or other equivalent edge computing devices as the platform to move computing tasks from the cloud to edge nodes in the reservoir monitoring area to reduce data transmission delays and improve the response speed of real-time monitoring;

[0010] b) Build a multimodal large model: Combined with a deep learning network, a comprehensive analysis model capable of processing multimodal data of images and videos is built. This model analyzes the surveillance video of the reservoir water source protection area in real time to identify violations such as prohibited burning, farming, floating objects on the water surface, garbage at catering outlets, and human intrusion;

[0011] c) Introducing small models to assist in identification: Design and deploy small models for special identification of specific types of violations, such as specific types of floating objects or burning bans in specific areas, to improve the accuracy and efficiency of identification;

[0012] d) Real-time monitoring and early warning: Use the multimodal large and small models on the edge computing platform to analyze the surveillance video data in real time. Once a violation is identified, the early warning mechanism is immediately triggered and the early warning information is sent to the management personnel through the preset communication method for timely response and processing;

[0013] e) Data transmission and storage: The key data and warning information processed on the edge computing platform are uploaded to the cloud platform for storage and analysis. At the same time, the cloud platform also provides remote access and management functions to support managers to remotely configure and monitor the monitoring system.

[0014] Furthermore, the multimodal large model includes but is not limited to convolutional neural networks, recurrent neural networks and their variants, and deep learning models based on attention mechanisms, which are used to extract key features in images and videos and improve recognition accuracy and generalization capabilities.

[0015] Furthermore, the small model is designed based on a lightweight neural network architecture, such as MobileNet and EfficientNet, so as to reduce computational complexity and resource consumption while ensuring recognition accuracy.

[0016] Furthermore, the edge computing platform also includes an IoT device access function, which can access meteorological monitoring equipment, pump station soundprint monitoring equipment, and water quality monitoring equipment to achieve comprehensive monitoring of the reservoir environment.

[0017] Furthermore, the early warning mechanism includes but is not limited to sound alarm, SMS notification, email reminder and APP push to ensure that management personnel can receive early warning information in a timely manner.

[0018] Furthermore, the method also includes continuous training and optimization of the multimodal large model and small model to adapt to changes in the reservoir monitoring environment and updates of violation behavior patterns.

[0019] Compared with the traditional monitoring method, the present invention has the following significant advantages:

[0020] 1. High efficiency: It realizes all-weather and all-round monitoring, significantly improving the monitoring efficiency;

[0021] 2. High recognition accuracy: The combination of multimodal large models and deep learning networks improves recognition accuracy;

[0022] 3. Timely response: Ability to analyze data in real time, detect and warn of violations in a timely manner;

[0023] 4. High degree of intelligence: It can automatically identify and handle violations, reducing human intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a flow chart of the optimization method of edge computing equipment based on multimodal large model in reservoir intelligent measurement of the present invention;

[0025] Figure 2 This is a system flow chart of the optimization method of edge computing equipment based on a multimodal large model in the reservoir intelligent measurement of the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, 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. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0027] In the description of the embodiments of the present invention, it should be noted that if the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the drawings, or the orientation or position relationship in which the invented product is usually placed when used. It is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0028] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components must be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0029] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0030] Example:

[0031] 1. Overview

[0032] This technical solution aims to provide an efficient and accurate optimization method for edge computing devices in reservoir intelligent measurement based on multimodal large models. Through the combination of edge computing devices and multimodal large models, comprehensive and real-time monitoring of reservoir water source protection areas can be achieved. It can automatically identify and warn of violations, such as prohibition of burning, farming, floating objects on the water surface, garbage from catering outlets, and human intrusion, thereby effectively protecting the water quality and ecological environment of the reservoir.

[0033] 2. Technical Solution

[0034] Edge computing platform deployment

[0035] Platform selection: Use China Tower Smart Edge Gateway or other high-performance edge computing devices as the core platform.

[0036] Functional configuration: Ensure that the edge computing platform has data processing, storage and transmission capabilities, and supports the access of IoT devices and the operation of AI algorithms.

[0037] Deployment location: Deploy the edge computing platform at the edge nodes in the reservoir monitoring area to reduce data transmission delays and improve the response speed of real-time monitoring.

[0038] Multimodal large model construction

[0039] Model selection: Combined with deep learning networks, build a comprehensive analysis model that can process multimodal data such as images and videos.

[0040] Feature extraction: Use deep learning models such as convolutional neural networks (CNNs) to extract key features in images and videos, such as color, shape, texture, etc.

[0041] Behavior recognition: The extracted features are further analyzed through recurrent neural networks (RNNs) and their variants or attention-based deep learning models to identify violations.

[0042] Model optimization: Use transfer learning, incremental learning and other methods to continuously optimize the model to adapt to changes in the reservoir monitoring environment and updates to violation behavior patterns.

[0043] Small model assisted recognition

[0044] Model design: Design and deploy small models with lightweight neural network architectures, such as MobileNet and EfficientNet, for specific types of violations.

[0045] Specialized identification: Use small models to perform specialized identification on specific scenarios, such as identifying specific types of floating objects or burning bans in specific areas.

[0046] Resource optimization: On the basis of ensuring recognition accuracy, reduce the computational complexity and resource consumption of small models to adapt to the resource limitations of edge computing devices.

[0047] Real-time monitoring and early warning

[0048] Data analysis: Use multimodal large and small models on the edge computing platform to perform real-time analysis of surveillance video data.

[0049] Early warning mechanism: Once a violation is identified, the early warning mechanism is immediately triggered and the early warning information is sent to the management personnel through preset communication methods (such as sound alarm, SMS notification, email reminder, and APP push, etc.).

[0050] Response and processing: After receiving the early warning information, the management personnel can immediately take corresponding measures to respond and handle it.

[0051] Data transmission and storage

[0052] Data upload: Upload key data and warning information processed on the edge computing platform to the cloud platform for storage and analysis.

[0053] Remote access: The cloud platform provides remote access and management functions, supporting managers to remotely configure and monitor the monitoring system.

[0054] Data security: Encryption technology is used to ensure the security of data transmission and storage to prevent data leakage and tampering.

[0055] IoT device access and integration

[0056] Device access: The edge computing platform supports the access of IoT devices, such as meteorological monitoring equipment, pump station soundprint monitoring equipment, water quality monitoring equipment, etc.

[0057] Data integration: Integrate and analyze the data collected by IoT devices with surveillance video data to achieve comprehensive monitoring of the reservoir environment.

[0058] 3. Implementation Effect

[0059] This technical solution realizes comprehensive and real-time monitoring of reservoir water source protection areas through the combination of edge computing devices and multimodal large models. It can automatically identify and warn of violations, improve monitoring efficiency and accuracy, and reduce missed detection and false alarm rates. At the same time, through the access and integration of IoT devices, it realizes comprehensive monitoring and data analysis of the reservoir environment, providing strong support for reservoir management and protection.

[0060] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. The optimization method of edge computing equipment in reservoir intelligent measurement based on multimodal large model is characterized by: The following steps are involved: a) Deploy edge computing platform: Use China Tower Smart Edge Gateway or other equivalent edge computing devices as the platform to move computing tasks from the cloud to edge nodes in the reservoir monitoring area to reduce data transmission delays and improve the response speed of real-time monitoring; b) Build a multimodal large model: Combined with a deep learning network, a comprehensive analysis model capable of processing multimodal data of images and videos is built. This model analyzes the surveillance video of the reservoir water source protection area in real time to identify violations such as prohibited burning, farming, floating objects on the water surface, garbage at catering outlets, and human intrusion; c) Introducing small models to assist in identification: Design and deploy small models for special identification of specific types of violations, such as specific types of floating objects or burning bans in specific areas, to improve the accuracy and efficiency of identification; d) Real-time monitoring and early warning: Use the multimodal large and small models on the edge computing platform to analyze the surveillance video data in real time. Once a violation is identified, the early warning mechanism is immediately triggered and the early warning information is sent to the management personnel through the preset communication method for timely response and processing; e) Data transmission and storage: The key data and warning information processed on the edge computing platform are uploaded to the cloud platform for storage and analysis. At the same time, the cloud platform also provides remote access and management functions to support managers to remotely configure and monitor the monitoring system.

2. The optimization method of edge computing equipment in reservoir intelligent measurement based on multimodal large model according to claim 1 is characterized by: The multimodal large model includes but is not limited to convolutional neural networks, recurrent neural networks and their variants, and deep learning models based on attention mechanisms, which are used to extract key features in images and videos and improve recognition accuracy and generalization capabilities.

3. The optimization method of edge computing equipment in reservoir intelligent measurement based on multimodal large model according to claim 1 is characterized by: The small model is designed based on a lightweight neural network architecture, such as MobileNet and EfficientNet, to reduce computational complexity and resource consumption while ensuring recognition accuracy.

4. The optimization method of edge computing equipment in reservoir intelligent measurement based on multimodal large model according to claim 1 is characterized by: The edge computing platform also includes an IoT device access function, which can access meteorological monitoring equipment, pump station soundprint monitoring equipment, and water quality monitoring equipment to achieve comprehensive monitoring of the reservoir environment.

5. The optimization method of edge computing equipment in reservoir intelligent measurement based on multimodal large model according to claim 1 is characterized by: The early warning mechanism includes but is not limited to sound alarm, SMS notification, email reminder and APP push to ensure that managers can receive early warning information in a timely manner.

6. The optimization method of edge computing equipment in reservoir intelligent measurement based on multimodal large model according to claim 1 is characterized by: The method also includes continuous training and optimization of multimodal large and small models to adapt to changes in the reservoir monitoring environment and updates of violation behavior patterns.

Images