CN115981874B - Decentralised AI analysis and data storage method and system based on cloud edge cooperation - Google Patents
Decentralised AI analysis and data storage method and system based on cloud edge cooperation Download PDFInfo
- Publication number
- CN115981874B CN115981874B CN202310264959.5A CN202310264959A CN115981874B CN 115981874 B CN115981874 B CN 115981874B CN 202310264959 A CN202310264959 A CN 202310264959A CN 115981874 B CN115981874 B CN 115981874B
- Authority
- CN
- China
- Prior art keywords
- cloud
- edge
- data storage
- video stream
- analysis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a cloud edge cooperation-based decentralization AI analysis and data storage method and a cloud edge cooperation-based decentralization AI analysis and data storage system, wherein the method comprises a data storage step and an AI analysis step, and the data storage comprises the following steps: acquiring a video stream and uploading the video stream to edge equipment; the edge equipment sends a video stream to the cloud CPU power computing server, realizes data processing through the cloud CPU power computing server and uploads the video stream to the data storage system; the data storage system returns the index table to the edge equipment for visual display; the AI analysis includes: acquiring a video stream, performing frame extraction through edge equipment, and judging whether the video has a change or not; when the video changes, analyzing whether a suspected target exists or not through an edge AI model, and intercepting; and sending the target area picture or video stream to a cloud AI computing server for AI analysis, and feeding back the result to the edge equipment for visual display. The invention can effectively reduce the AI calculation force and the comprehensive cost of data storage.
Description
Technical Field
The invention relates to the technical field of computer networks, in particular to a cloud edge cooperation-based decentralization AI analysis and data storage method and system.
Background
With the development of information technology, the demands for AI computing power and storage are increasing, but the centralized computing power and storage support short plates are prominent, and the contradiction between cost, stability and performance is prominent. With the upgrading of informatization technology, the demands for more complex artificial intelligence systems and computing power systems are more remarkable, but effective computing power solutions are lacking in the current market. Along with the development of the economic society of China and the acceleration of the construction pace of smart cities, the requirements of new capital construction projects and the like are vigorous, and the innovation of digital transformation, intelligent promotion and integration will step into the expressway. Traditional centralized AI computing power and AI storage solutions face the problems of high cost, difficult deployment, difficult capacity expansion, difficult delivery, high delay and difficult maintenance.
Disclosure of Invention
The invention aims to provide a cloud-edge collaboration-based decentralised AI analysis and data storage method and system, which are used for solving the technical problems of high cost, difficult deployment, difficult capacity expansion, difficult delivery, high delay and difficult maintenance caused by the rapid increase of the current AI calculation power and storage requirements.
The invention is realized by adopting the following technical scheme: the cloud edge cooperation-based decentralizing AI analysis and data storage method comprises a data storage step and an AI analysis step, wherein the data storage comprises the following steps:
acquiring a video stream and uploading the video stream to edge equipment;
the edge equipment sends a video stream to the cloud CPU power computing server, realizes data processing through the cloud CPU power computing server and uploads the video stream to the data storage system;
the data storage system returns the index table to the edge equipment for visual display;
the AI analysis includes:
acquiring a video stream, performing frame extraction through edge equipment, and judging whether the video has a change or not;
when the video changes, analyzing whether a suspected target exists or not through an edge AI model, and intercepting;
and sending the target area picture or video stream to a cloud AI computing server for AI analysis, and feeding back the result to the edge equipment for visual display.
Further, the video stream is obtained through a camera or an upper server.
Further, the edge device processor sends the video stream to the cloud CPU power computing server, wherein the video stream comprises two conditions, the first condition is that the edge device processor directly sends the video stream to the cloud CPU power computing server, the second condition is that the edge device processor firstly carries out simple and rough AI identification, after the key region is captured, the key region is sent to the cloud CPU power computing server, and in the sending process, the edge device is responsible for temporary caching of data so as to avoid the cloud CPU power computing server from being disconnected.
Further, the edge device processor includes an edge device CPU, TPU, GPU and an FPGA.
Further, the cloud CPU power calculation server realizes data processing, sends the video stream to the data storage system on the cloud after processing, returns the index table stored by the data storage system to the edge device, and meanwhile, the data storage system backs up the index table, and the edge device retrieves data from the data storage system according to the index table for visual display as required.
Further, the data storage system includes one of an IPFS data storage system, a centralized storage system, or a distributed storage system.
Further, frame extraction is performed through the edge device VPU, CPU, GPU, TPU or the FPGA, and the data after frame extraction is transferred to the NPU of the edge device, and the NPU of the edge device analyzes whether the video changes, so as to determine whether AI analysis is performed.
Further, the cloud AI computing server performs AI analysis on the target area picture or the video stream, sends the AI analysis result to the edge device and the cloud CPU computing server, the cloud CPU computing server realizes data processing, sends the video stream to the data storage system on the cloud after processing, returns the index table stored by the data storage system to the edge device, and meanwhile, the data storage system backs up the index table, and the edge device retrieves data from the data storage system according to the index table for visual display according to the need.
Further, when the video changes, whether a suspected target exists or not is analyzed through an edge AI model, the video stream is intercepted through an NPU of the edge equipment, after the video stream is intercepted, the original image is temporarily stored, after the time of T is delayed, and after all AI analysis is finished, data display is carried out on the edge equipment.
The cloud edge cooperation-based decentralization AI analysis and data storage system comprises a camera, an upper server, edge equipment, a cloud CPU server, a cloud AI server and a storage network, wherein the video stream is acquired through the camera or the upper server and uploaded to the edge equipment, the edge equipment sends the video stream to the cloud CPU server, data processing is realized through the cloud CPU server and uploaded to the storage network, and the storage network returns an index table to the edge equipment for visual display; the edge equipment performs frame extraction on the video stream and judges whether the video has change or not; when the video changes, analyzing whether a suspected target exists or not through an edge AI model, and intercepting; and sending the target area picture or video stream to a cloud AI server for AI analysis, and feeding back the result to the edge equipment for visual display.
The invention has the beneficial effects that: the invention realizes an AI infrastructure system with low cost, high stability and easy deployment, and provides a simple and easy AI computing and storage solution with low cost, high performance and high reliability for intelligent cities, intelligent factories, intelligent agriculture and various AI services.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to the structures shown in these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of the present invention;
fig. 2 is a block diagram of the system of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
It should be noted that: like reference numerals and letters denote like items in the following figures, and thus once an item is defined in one figure, no further definition or explanation thereof is necessary in the following figures.
Some embodiments of the present invention are described in detail below with reference to the accompanying drawings. The following embodiments and features of the embodiments may be combined with each other without conflict.
Example 1
Referring to fig. 1, the cloud-edge collaboration-based decentralizing AI analysis and data storage method includes a data storage step and an AI analysis step, wherein the data storage includes:
s11: and acquiring the video stream through a camera or an upper server.
S12: the edge equipment CPU (TPU, GPU, FPGA and other types of processors can be adopted) sends the video stream to the cloud CPU power computing server, and the edge equipment is responsible for temporary caching of data in the sending process so as to avoid the cloud CPU power computing server from being disconnected.
S13: the cloud CPU power server realizes data processing, processes the video stream, sends the processed video stream to an IPFS cloud storage system (traditional centralized storage, distributed storage and other technologies can be used) on the cloud, and returns an index table (hash value table) stored by the IPFS to the edge equipment (the IPFS cloud storage system can be returned simultaneously for backup).
S14: and the edge equipment retrieves data from the IPFS cloud storage system according to the index table for visual display according to the need.
The AI analysis includes:
s21: and acquiring the video stream through a camera or an upper server.
S22: the frames are decimated by the edge device VPU (which may take CPU, GPU, TPU, FPGA etc.) and the data is transferred to the edge device NPU.
S23: the NPU of the edge device analyzes whether the video has changed to determine whether AI analysis is performed.
S24: and if the change is determined, analyzing whether a suspected target exists according to the pre-deployment edge AI model.
S25: and intercepting the suspected target area and performing definition adjustment. (S22-S25 can be omitted according to the bandwidth condition)
S26: and sending the target area picture or video stream to a cloud AI computing server.
S27: AI analysis is performed according to a pre-deployment AI model.
S28: and sending the AI analysis result to an edge device (used for backup and visual display) and a cloud CPU computing power server.
S29: the cloud CPU power server realizes data processing, sends the video stream to an IPFS cloud storage system on the cloud after processing, and returns an index table stored by the IPFS to the edge device (the IPFS cloud storage system can be returned for backup). And the edge equipment invokes data from the IPFS cloud storage system according to the IPFS storage index table as required for visual display.
It should be noted that, the edge AI model is a rough model, and requires low computational effort, for example, fire identification, so that the edge may be only detected by a similar red shape, and more objects (such as red clothes) may be misjudged, but a suspected object may be subjected to screenshot, so that AI computational effort is reduced, and standard AI flame identification (AI model) is performed after the AI computational effort is transmitted to the cloud.
Referring to fig. 2, a cloud edge collaboration-based decentralization AI analysis and data storage system comprises a camera, an upper server, edge equipment, a cloud CPU server, a cloud AI server and an IPFS storage network (the traditional decentralization storage and the distributed storage can be realized), wherein video streams are acquired through the camera or the upper server and uploaded to the edge equipment, the edge equipment sends the video streams to the cloud CPU server, data processing is realized through the cloud CPU server and uploaded to the IPFS storage network, and the IPFS storage network returns an index table to the edge equipment for visual display; the edge equipment performs frame extraction on the video stream and judges whether the video has change or not; when the video changes, analyzing whether a suspected target exists or not through an edge AI model, and intercepting; and sending the target area picture or video stream to a cloud AI server for AI analysis, and feeding back the result to the edge equipment for visual display.
Furthermore, the edge device VPU needs to temporarily store the original image after intercepting the video stream, delay for N seconds (generally take 3 s), and perform data display on the edge device after all AI analysis results are returned. The edge device can temporarily store the cloud data on the unfinished IPFS split so as to avoid the cloud CPU server from crashing. The edge device can temporarily store the video stream and the ROI picture which are not uploaded as required so as to avoid data loss caused by special conditions such as offline of the edge device. The spare storage space of the edge device should be contributed to the use of the IPFS storage network.
In this embodiment, the camera may be any type of camera on the market; the edge device preferably selects ARM platform microcomputer with AI computing power (such as Rayleigh core micro RK 3588S); the cloud CPU server is a common CPU server (such as Intel X99 server); the cloud AI server is a server (such as Nvidia RTX 3060) comprising AI computing power equipment with high performance GPU, TPU, DPU, FPGA; an IPFS storage network is a computer device that has a large amount of storage space (including but not limited to X86, ARM, RISC-V processor based devices that can connect to various types of storage media).
The invention has at least the following technical effects: 1. the comprehensive cost of AI calculation force is greatly reduced by more than 95 percent. 2. The comprehensive cost of data storage is greatly reduced by more than 70%. 3. The existing infrastructure is fully utilized, and comprehensive AI upgrading can be rapidly realized. 4. The user side can effectively mobilize a large amount of AI calculation force and storage space by only operating one edge device. 5. The deployment cost of the AI model is greatly reduced, and the accuracy and the adaptation degree of the model are improved.
It should be noted that, for simplicity of description, the foregoing embodiments are all described as a series of combinations of actions, but it should be understood by those skilled in the art that the present application is not limited by the order of actions described, as some steps may be performed in other order or simultaneously according to the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are presently preferred embodiments, and that the acts referred to are not necessarily required for the present application.
In the above embodiments, the basic principle and main features of the present invention and advantages of the present invention are described. It will be appreciated by persons skilled in the art that the present invention is not limited by the foregoing embodiments, but rather is shown and described in what is considered to be illustrative of the principles of the invention, and that modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the invention, and therefore, is within the scope of the appended claims.
Claims (5)
1. The cloud edge cooperation-based decentralizing AI analysis and data storage method is characterized by comprising a data storage step and an AI analysis step, wherein the data storage comprises the following steps:
acquiring a video stream and uploading the video stream to edge equipment;
the edge equipment sends a video stream to the cloud CPU power computing server, realizes data processing through the cloud CPU power computing server and uploads the video stream to the data storage system;
the data storage system returns the index table to the edge equipment for visual display; the cloud CPU power calculation server realizes data processing, sends the video stream to a data storage system on the cloud after processing, returns an index table stored by the data storage system to the edge equipment, and meanwhile, the data storage system backs up the index table, and the edge equipment retrieves data from the data storage system according to the index table for visual display as required;
the AI analysis includes:
acquiring a video stream, performing frame extraction through edge equipment, and judging whether the video has a change or not; frame extraction is carried out through edge equipment VPU, CPU, GPU, TPU or FPGA, data after frame extraction is transferred to NPU of the edge equipment, and whether video changes is analyzed by the NPU of the edge equipment, so that whether AI analysis is carried out is determined; the cloud AI computing server performs AI analysis on the target area picture or video stream, sends an AI analysis result to the edge device and the cloud CPU computing server, realizes data processing by the cloud CPU computing server, sends the video stream to a data storage system on the cloud after processing, returns an index table stored by the data storage system to the edge device, and simultaneously, the data storage system backs up the index table, and the edge device retrieves data from the data storage system according to the index table for visual display according to the need;
when the video changes, analyzing whether a suspected target exists or not through an edge AI model, intercepting a video stream through an NPU of edge equipment, temporarily storing an original image after intercepting, delaying for T time, and displaying data at the edge equipment after all AI analysis is finished;
and sending the target area picture or video stream to a cloud AI computing server for AI analysis, and feeding back the result to the edge equipment for visual display.
2. The cloud-edge collaboration-based decentralised AI analysis and data storage method of claim 1, wherein the video stream is acquired through a camera or an upper server.
3. The cloud edge collaboration-based decentralised AI analysis and data storage method of claim 1, wherein the video stream is sent to a cloud CPU power server by an edge device processor, and in the sending process, the edge device is responsible for temporary caching of data to avoid the cloud CPU power server from being dropped.
4. The cloud edge collaboration-based de-centralized AI analysis and data storage method of claim 3, wherein the edge device processor comprises an edge device CPU, TPU, GPU and an FPGA.
5. The cloud edge collaborative-based decentralization AI analysis and data storage system is used for realizing the cloud edge collaborative-based decentralization AI analysis and data storage method according to any one of claims 1-4, and is characterized by comprising a camera, an upper server, edge equipment, a cloud CPU server, a cloud AI server and a storage network, wherein a video stream is acquired through the camera or the upper server and is uploaded to the edge equipment, the edge equipment sends the video stream to the cloud CPU server, data processing is realized through the cloud CPU server and is uploaded to the storage network, and the storage network returns an index table to the edge equipment for visual display; the edge equipment performs frame extraction on the video stream and judges whether the video has change or not; when the video changes, analyzing whether a suspected target exists or not through an edge AI model, and intercepting; and sending the target area picture or video stream to a cloud AI server for AI analysis, and feeding back the result to the edge equipment for visual display.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310264959.5A CN115981874B (en) | 2023-03-20 | 2023-03-20 | Decentralised AI analysis and data storage method and system based on cloud edge cooperation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310264959.5A CN115981874B (en) | 2023-03-20 | 2023-03-20 | Decentralised AI analysis and data storage method and system based on cloud edge cooperation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115981874A CN115981874A (en) | 2023-04-18 |
CN115981874B true CN115981874B (en) | 2023-06-13 |
Family
ID=85970852
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310264959.5A Active CN115981874B (en) | 2023-03-20 | 2023-03-20 | Decentralised AI analysis and data storage method and system based on cloud edge cooperation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115981874B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113239792A (en) * | 2021-05-11 | 2021-08-10 | 深圳市安软科技股份有限公司 | Big data analysis processing system and method |
CN113434729A (en) * | 2021-08-04 | 2021-09-24 | 深圳墨世科技有限公司 | Video related information aggregation obtaining method and device and terminal equipment |
CN114911869A (en) * | 2022-05-13 | 2022-08-16 | 北京航星永志科技有限公司 | Data and file storage system, method and device and electronic equipment |
CN115103157A (en) * | 2022-06-15 | 2022-09-23 | 中科弘云科技(北京)有限公司 | Video analysis method and device based on edge cloud cooperation, electronic equipment and medium |
CN115114236A (en) * | 2021-03-19 | 2022-09-27 | 深圳宏芯宇电子股份有限公司 | Management system and management method for photos and/or videos |
CN115562824A (en) * | 2022-09-22 | 2023-01-03 | 国网冀北电力有限公司张家口供电公司 | Computing resource cooperative scheduling system, method, device and storage medium |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11394929B2 (en) * | 2020-09-11 | 2022-07-19 | Samsung Electronics Co., Ltd. | System and method for language-guided video analytics at the edge |
CN114255432A (en) * | 2021-12-24 | 2022-03-29 | ***数智科技有限公司 | Video stream processing method, device, electronic equipment, storage medium and system |
CN114398449B (en) * | 2021-12-29 | 2023-01-06 | 深圳市海清视讯科技有限公司 | Data processing method, device, video monitoring system and storage medium |
-
2023
- 2023-03-20 CN CN202310264959.5A patent/CN115981874B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115114236A (en) * | 2021-03-19 | 2022-09-27 | 深圳宏芯宇电子股份有限公司 | Management system and management method for photos and/or videos |
CN113239792A (en) * | 2021-05-11 | 2021-08-10 | 深圳市安软科技股份有限公司 | Big data analysis processing system and method |
CN113434729A (en) * | 2021-08-04 | 2021-09-24 | 深圳墨世科技有限公司 | Video related information aggregation obtaining method and device and terminal equipment |
CN114911869A (en) * | 2022-05-13 | 2022-08-16 | 北京航星永志科技有限公司 | Data and file storage system, method and device and electronic equipment |
CN115103157A (en) * | 2022-06-15 | 2022-09-23 | 中科弘云科技(北京)有限公司 | Video analysis method and device based on edge cloud cooperation, electronic equipment and medium |
CN115562824A (en) * | 2022-09-22 | 2023-01-03 | 国网冀北电力有限公司张家口供电公司 | Computing resource cooperative scheduling system, method, device and storage medium |
Non-Patent Citations (2)
Title |
---|
"Distributed intelligence on the Edge-to-Cloud Continuum: A systematic literature review";Daniel Rosendo;《Journal of Parallel and Distributed Computing》;第166卷;第71-94页 * |
"基于深度学习的智能交通视频处理框架的研究";徐亮;《中国优秀硕士学位论文全文数据库 工程科技II辑》(2019年第07期);第C034-363页 * |
Also Published As
Publication number | Publication date |
---|---|
CN115981874A (en) | 2023-04-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109327509B (en) | Low-coupling distributed streaming computing system of master/slave architecture | |
AU2017250159B2 (en) | Video recording method, server, system, and storage medium | |
US9704393B2 (en) | Integrated intelligent server based system and method/systems adapted to facilitate fail-safe integration and/or optimized utilization of various sensory inputs | |
WO2020147331A1 (en) | Micro-service monitoring method and system | |
CA2824330C (en) | An integrated intelligent server based system and method/systems adapted to facilitate fail-safe integration and/or optimized utilization of various sensory inputs | |
WO2017024975A1 (en) | Unmanned aerial vehicle portable ground station processing method and system | |
CN110505312A (en) | It is a kind of can cloud deployment platform of internet of things management system | |
CN114255432A (en) | Video stream processing method, device, electronic equipment, storage medium and system | |
CN110334072A (en) | A kind of distributed file system, file updating method and device | |
CN111091106B (en) | Image clustering method and device, storage medium and electronic device | |
CN111479095B (en) | Service processing control system, method and device | |
CN112785351A (en) | Financial website monitoring method and device | |
CN106844727B (en) | Mass image characteristic data distributed acquisition processing and grading application system and method | |
CN113434573A (en) | Multi-dimensional image retrieval system, method and equipment | |
CN111510680B (en) | Image data processing method, system and storage medium | |
CN110601891B (en) | Alarm processing method and related device | |
CN110505307B (en) | Method and system for exchanging traffic flow data between networks | |
CN112486707A (en) | Redis-based message asynchronous consumption method and device | |
CN115396296A (en) | Service processing method and device, electronic equipment and computer readable storage medium | |
CN115981874B (en) | Decentralised AI analysis and data storage method and system based on cloud edge cooperation | |
CN108134746B (en) | Method and device for processing rail transit data | |
CN111314395A (en) | Image transmission method, terminal and storage medium | |
CN109561346A (en) | A kind of distributed analytic method and system of video | |
CN109241777A (en) | Distribution is ordered approaches to IM, device and network node | |
CN113362090A (en) | User behavior data processing method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |