CN-121982274-A - Method and system for detecting instrument reading and hidden danger based on AI (advanced technology attachment) identification

CN121982274ACN 121982274 ACN121982274 ACN 121982274ACN-121982274-A

Abstract

The invention provides a meter reading and hidden danger detection method and system based on AI identification, which relate to the technical field of gas fault detection, and are used for logging in a gas inspection terminal APP, executing identity verification and household verification to generate metering inspection function information, executing image acquisition frame acquisition, adjustment and index analysis to generate shooting positioning index information, generating standard acquisition images according to the shooting positioning index information, executing image scanning, extraction and identification on the standard acquisition images, generating gas meter reading acquisition data, executing fault analysis determination on the gas meter reading acquisition data, generating check fault determination information, acquiring fault equipment position parameter data, executing fault degree and fault correlation analysis on the fault equipment position parameter data, generating a fault correlation map and executing fault early warning.

Inventors

WU PENG
ZHAO JUNWEI
FENG JUNJIE
DU WEI
LIAO MIN
Song Jiangeng
Lai jiaxing
QIU ZHENDONG

Assignees

深圳市赛易特信息技术有限公司
深圳市燃气投资有限公司

Dates

Publication Date: 20260505
Application Date: 20260130

Claims (10)

1.A method for detecting meter readings and hidden danger based on AI identification, the method comprising: S1, logging in a gas inspection terminal APP, performing identity verification, collecting meter reading task account information, and further performing household verification to generate metering inspection function information; S2, image acquisition frame acquisition, adjustment and index analysis are performed through a metering inspection function, shooting positioning index information is generated, and standard acquisition images are generated according to the shooting positioning index information; S3, performing image scanning, extraction and recognition on the standard acquired images through a visual target detection model to generate gas meter reading acquisition data, performing fault analysis and determination on the gas meter reading acquisition data, and generating check fault determination information; s4, acquiring fault equipment position parameter data according to the check fault determination information, performing fault degree and fault association analysis on the fault equipment position parameter data, generating a fault association map and performing fault early warning.
2. The AI-identification-based meter reading and hidden danger detection method of claim 1, wherein S1 comprises: logging in a gas inspection terminal APP, and executing identity verification according to the gas inspection terminal APP to generate identity verification information; collecting meter reading task account information according to the identity verification information; And performing field identity verification according to the meter reading task account information, generating identity verification information, activating a metering inspection function according to the identity verification information, and generating metering inspection function information.
3. The AI-identification-based meter reading and hidden danger detection method of claim 2, wherein the performing on-site identity verification according to meter reading task ledger information, generating identity verification information, activating a metering patrol function according to the identity verification information, and generating a metering patrol function comprises: collecting a standardized task list according to meter reading task standing account information; Three matching checks of satellite positioning coordinates, a near field communication access control address and a terminal user address are executed on the standardized task list, and matching check information is generated; the gas inspection terminal APP starts a bidirectional authentication function of inspector identity and terminal user identity, and generates an on-site operation electronic certificate according to the bidirectional authentication function; Uploading the field operation electronic certificate to a distributed account book certificate storage system, and activating a metering and inspection function.
4. The AI-identification-based meter reading and hidden danger detection method of claim 1, wherein S2 comprises: acquiring image acquisition frame information through a metering and inspection function of a gas inspection terminal APP; performing image acquisition frame acquisition according to the image acquisition frame information to generate image acquisition frame acquisition information; executing image acquisition frame acquisition analysis according to the image acquisition frame acquisition information to generate image acquisition frame acquisition analysis data; executing image acquisition adjustment analysis according to the image acquisition frame acquisition analysis data to generate image acquisition adjustment analysis data; generating a framing image acquisition index according to the image acquisition adjustment analysis data, and generating shooting positioning index information; And acquiring image acquisition feedback information according to the shooting positioning index information, and executing image acquisition frame acquisition analysis on the image acquisition feedback information until a standard acquisition image is generated.
5. The AI-recognition-based meter reading and hidden danger detection method of claim 4, wherein the performing image acquisition frame acquisition parsing based on the image acquisition frame acquisition information to generate image acquisition frame acquisition parsing data comprises: comparing the image acquisition frame acquisition information with image acquisition frame preset information to generate image acquisition frame comparison information; determining fault characteristic information of image acquisition frame acquisition information in image acquisition frame preset information according to the image acquisition frame comparison information; Performing data division of a plurality of preset fault types on the fault characteristic information to generate characteristic type fault data; Performing fault feature labeling on the image acquisition frame acquisition information according to the feature type fault data to generate fault feature labeling information; the fault characteristic labeling information is acquired analysis data of the image acquisition frame.
6. The AI-recognition-based meter reading and hidden danger detection method of claim 4, wherein the performing image acquisition adjustment resolution based on the image acquisition frame acquisition resolution data to generate image acquisition adjustment resolution data comprises: Acquiring analysis data according to an image acquisition frame to generate corresponding fault correction data; executing characteristic automatic adjustment according to the fault correction data to generate characteristic adjustment data; Triggering a meter detection list flow according to the characteristic adjustment data, and collecting equipment to be collected and environmental item combinations; executing mode switching on different image acquisition equipment and environment items to generate combined mode switching information; and acquiring characteristic adjustment data of the combined mode switching information, namely image acquisition adjustment analysis data.
7. The AI-identification-based meter reading and hidden danger detection method of claim 1, wherein S3 comprises: Acquiring a visual target detection model, and executing pixel-by-pixel image recognition on the standard acquired images according to the visual target detection model to generate image scanning information; performing instrument display area positioning on the image scanning information to generate information display positioning data; Performing registration information extraction according to the information display positioning data to generate registration information extraction data; Performing the reading character analysis on the reading information extraction data according to the fusion recognition model to generate reading character analysis data; combining and checking the reading character analysis data to generate gas meter reading acquisition data; performing fault checking on the gas meter reading acquisition data to generate fault checking data; And performing check fault determination on the gas meter reading acquisition data according to the fault check data to generate check fault determination information.
8. The AI-identification-based meter reading and hidden danger detection method of claim 1, wherein S4 comprises: performing device classification, location classification, and parameter classification on the verification failure determination information, generating failure device data, failure location data, and failure parameter data; Performing association matching on the fault equipment data, the fault location data and the fault parameter data to generate fault equipment bit parameter data; performing fault degree evaluation on the fault equipment bit parameter data to generate fault degree evaluation data; Performing fault correlation analysis of the fault degree evaluation data on the plurality of fault equipment bit parameters to generate fault correlation analysis data; comparing the fault association analysis data with a reference association analysis threshold value to generate a fault association comparison result; and generating a fault association map according to the fault association comparison result, and executing fault early warning.
9. The AI-identification-based meter reading and hidden danger detection method of claim 8, wherein performing a fault level assessment on the fault device bit parameter data generates fault level assessment data comprising: collecting the duty ratio of the fault equipment bit parameter data in all the fault equipment bit parameter data, and generating bit reference duty ratio weights; performing sequencing from large to small on bit reference weight of the plurality of fault equipment bit parameter data to generate an equipment position parameter fault sequence; generating difference data of adjacent ranking bit reference ratio weights of the equipment position parameter fault sequence, and generating adjacent difference coefficients; comparing the adjacent difference coefficient with a preset adjacent difference threshold value to generate an adjacent difference comparison result; And performing grading on the equipment position parameter fault sequence according to the adjacent difference value comparison result to generate fault degree evaluation data.
10. The system is characterized by comprising an authentication activation module, a verification module and a verification module, wherein the authentication activation module is used for logging in a gas inspection terminal APP, performing identity verification, collecting meter reading task account information, further performing household verification and generating metering inspection function information; The system comprises an image acquisition analysis module, an image analysis module, a gas meter reading acquisition module, a gas meter analysis module and a control module, wherein the image acquisition analysis module is used for executing image acquisition frame acquisition, adjustment and index analysis through a metering inspection function to generate shooting positioning index information and generating standard acquisition images according to the shooting positioning index information; The fault early warning module is used for acquiring fault equipment position parameter data according to the check fault determination information, performing fault degree and fault association analysis on the fault equipment position parameter data, generating a fault association map and performing fault early warning.

Description

Method and system for detecting instrument reading and hidden danger based on AI (advanced technology attachment) identification Technical Field The invention provides a method and a system for detecting meter readings and hidden dangers based on AI (advanced technology) identification, relates to the technical field of gas hidden dangers identification, and particularly relates to the technical field of meter readings and hidden dangers detection based on AI identification. Background The gas meter reading acquisition and hidden danger detection are the key links of gas safety management and metering charging. At present, the operation is mainly performed manually, and the problems of nonstandard checking of the identity of the user, difficult traceability of the operation, counterfeit, shoveling, missed reading and the like exist, when the on-site shooting instrument image is used for AI identification, the identification accuracy is low due to the fact that the on-site shooting instrument image is easily affected by the environment, and hidden danger detection only focuses on a single instrument, and the correlation analysis is lacking, so that the early warning pertinence is poor. Disclosure of Invention The invention provides a method and a system for detecting instrument readings and hidden dangers based on AI identification, which are used for solving the problems: the invention provides a method and a system for detecting instrument readings and hidden danger based on AI identification, wherein the method comprises the following steps: S1, logging in a gas inspection terminal APP, performing identity verification, collecting meter reading task account information, and further performing household verification to generate metering inspection function information; S2, image acquisition frame acquisition, adjustment and index analysis are performed through a metering inspection function, shooting positioning index information is generated, and standard acquisition images are generated according to the shooting positioning index information; S3, performing image scanning, extraction and recognition on the standard acquired images through a visual target detection model to generate gas meter reading acquisition data, performing fault analysis and determination on the gas meter reading acquisition data, and generating check fault determination information; s4, acquiring fault equipment position parameter data according to the check fault determination information, performing fault degree and fault association analysis on the fault equipment position parameter data, generating a fault association map and performing fault early warning. Further, the step S1 includes: logging in a gas inspection terminal APP, and executing identity verification according to the gas inspection terminal APP to generate identity verification information; collecting meter reading task account information according to the identity verification information; And performing field identity verification according to the meter reading task account information, generating identity verification information, activating a metering inspection function according to the identity verification information, and generating metering inspection function information. Further, the step of performing field identity verification according to the meter reading task ledger information to generate identity verification information, and activating the metering inspection function according to the identity verification information to generate the metering inspection function comprises the following steps: collecting a standardized task list according to meter reading task standing account information; Three matching checks of satellite positioning coordinates, a near field communication access control address and a terminal user address are executed on the standardized task list, and matching check information is generated; the gas inspection terminal APP starts a bidirectional authentication function of inspector identity and terminal user identity, and generates an on-site operation electronic certificate according to the bidirectional authentication function; Uploading the field operation electronic certificate to a distributed account book certificate storage system, and activating a metering and inspection function. Further, the step S2 includes: acquiring image acquisition frame information through a metering and inspection function of a gas inspection terminal APP; performing image acquisition frame acquisition according to the image acquisition frame information to generate image acquisition frame acquisition information; executing image acquisition frame acquisition analysis according to the image acquisition frame acquisition information to generate image acquisition frame acquisition analysis data; executing image acquisition adjustment analysis according to the image acquisition frame acquisition analysis data to generate image acquisition adjus