CN-121982628-A - Method, system, equipment and medium for intelligent recognition of secondary safety measure area image based on area positioning geometric comparison

CN121982628ACN 121982628 ACN121982628 ACN 121982628ACN-121982628-A

Abstract

The invention discloses a method, a system, equipment and a medium for intelligently identifying a secondary safety measure area image based on area positioning geometric comparison, which belong to the technical field of power inspection automation and comprise the following steps: the method comprises the steps of collecting field device images of secondary safety measures, preprocessing the field device images of the secondary safety measures, extracting salient areas in the images, dividing the salient areas into blocks through area positioning and module, obtaining image sub-modules, performing area positioning and feature extraction, performing text recognition, processing recognition results, outputting text information, performing target part recognition on the image sub-modules, extracting part state information, forming structural recognition results, checking the consistency of the recognition results based on preset secondary safety measure rules, and outputting check conclusion.

Inventors

CHEN FEIJIAN
YAO YU
LUO ZHONGKUI
LI CHENGCAI
HUANG LI
SUN JIANGSHAN
GU YOUFANG
WANG ZHIZHAO
ZHOU DACHENG
LIANG LINLIN
YANG JINCHENG
LI SHICAI
LU WAN
WANG RUISONG
YANG TIANYU
LI BO
ZHANG HENG

Assignees

贵州电网有限责任公司

Dates

Publication Date: 20260505
Application Date: 20251219

Claims (10)

1. A method for intelligently identifying a secondary safety measure area image based on area positioning geometric comparison is characterized by comprising the steps of, Acquiring a field device image of the secondary safety measure, preprocessing the field device image of the secondary safety measure, extracting a salient region in the image, and partitioning the region with a module through region positioning to obtain an image sub-module; Based on the image sub-module, performing region positioning and feature extraction, performing text recognition, processing recognition results, and outputting text information; And carrying out target component identification on the image submodule based on the text information, extracting component state information to form a structural identification result, checking the consistency of the identification result based on a preset secondary safety measure rule, and outputting a verification conclusion.
2. The method for intelligently identifying the secondary safety measure area image based on the area positioning geometric comparison of claim 1, wherein the steps of collecting the field device image of the secondary safety measure, preprocessing the field device image of the secondary safety measure, extracting the salient area in the image, and obtaining the image submodule comprise the steps of, Performing space conversion and noise reduction treatment on the secondary safety measure operation image to obtain a treated image; Performing self-adaptive binarization and geometric correction on the processed image to obtain a standardized binary image; and extracting a salient region in the binary image, performing region positioning and module blocking on the binary image, and outputting an image sub-module.
3. The method for intelligently identifying the secondary safety measure regional image based on regional positioning and geometric comparison according to claim 2, wherein the regional positioning and characteristic extraction is carried out based on the image sub-module, the character recognition is carried out, the recognition result is processed, and the output of the character information comprises, Based on the image sub-module, performing region positioning to obtain a white pixel region, and calculating to obtain geometric features; Dividing a white pixel area according to the size relation between the geometric features and a preset geometric feature threshold value, and preprocessing the white pixel area; and carrying out character recognition based on the preprocessed white pixel area part, processing a recognition result and outputting character information.
4. The method for intelligently identifying the secondary safety measure area image based on the area positioning geometric comparison of claim 3, wherein the step of carrying out target component identification on the image submodule based on the text information, extracting component state information to form a structural identification result, checking the consistency of the identification result based on a preset secondary safety measure rule, and outputting a check conclusion comprises the steps of, Based on the text information, carrying out target component identification on the image submodule and outputting a target detection result; based on the target detection result, performing spatial position analysis, extracting component state information, and forming a structural identification result; And according to the structural identification result, checking the consistency of the identification result based on a preset secondary safety measure rule, and outputting a verification conclusion.
5. The method for intelligently identifying the secondary safety measure regional image based on regional positioning geometric comparison of claim 4, wherein the performing spatial conversion and noise reduction on the secondary safety measure work image to obtain a processed image comprises, Converting the input RGB color image into HSV color space, and the expression is: Wherein, the For the purpose of lightness, the brightness is, 、、 Respectively the red, green and blue channel components of the input image, In order to be of saturation level, Is a tone; the RGB image is converted into a gray image by adopting a human eye perception weighting method based on CRT display standard, and the expression is: Wherein, the For the output of a gray-scale image, 、、 Respectively the red, green and blue channel components of the input image, 、、 The weight coefficients of the red, green and blue channel components of the input image are respectively; Morphological filtering of selected single channel images using structural elements The isolated white noise point, burr and tiny non-target bulge in the image are eliminated through the operation of corrosion and expansion, and the expression is: Wherein, the In order to open the output image after the operation processing, In order to input an image of the subject, As a structural element of the structure of the metal-insulator-metal composite, In order to be able to etch the substrate, Is expansion.
6. The method for intelligently identifying the secondary measure area image based on the area positioning geometric comparison of claim 5, wherein the adaptive binarization and the geometric correction are carried out on the processed image to obtain a standardized binary image, Each pixel point in the image adopts a local self-adaptive threshold method Threshold of (2) From the neighborhood The mean and standard differential state calculations of (2) are expressed as: Wherein, the And Respectively, are neighborhoods Is defined as the mean and standard deviation of (c), In order to correct the coefficient of the coefficient, For the dynamic range of the device, As a result of the binary image being obtained, The output image is obtained after the on operation processing; applying probability Hough transform to binary image to detect long straight line segment, and setting the detected straight line set as The included angle between each straight line and the horizontal axis is Calculating the mean as the overall tilt angle of the document Geometric correction is carried out through an image rotation matrix, and the expression is as follows: Wherein, the Is the coordinates of the original image, and the coordinate of the original image is the coordinates of the original image, In order to correct the post-coordinates, In order to be able to translate the quantity, Is the overall tilt angle of the document.
7. The method for intelligently identifying the secondary safety measure area image based on the area positioning geometric comparison of claim 6, wherein the area positioning is performed based on the image sub-module to obtain a white pixel area, the geometric feature is obtained through calculation, Traversing the image by adopting a self-adaptive line scanning algorithm, and defining the pixel density of the local window W as follows: Wherein, the For the pixel density of the local window W, For the local image region currently being scanned, Is the position Binary pixel values at; The scanning step length L is based on Dynamic adjustment, the expression is: Wherein, the For a dynamically adjusted scanning step size, As a base step size of the steps, In order to adjust the coefficient of the power supply, Pixel density for the local window W; Detecting continuous white pixel area, and recording the width of the circumscribed rectangle as High as The area is Constructing a four-dimensional feature vector, wherein the expression is as follows: Wherein, the In the case of a four-dimensional feature vector, 、、、 Respectively shape characteristics, scale characteristics, internal filling density and uniformity of horizontal distribution, For the purpose of the transposition, Is the width of the external rectangle, Is of a height of an external rectangle, Is the area of the external rectangle, and the external rectangle is the area of the external rectangle, 、 The total width of the original image and the total height of the original image respectively, Is the position A binary pixel value at which, For a region-level projection histogram, 、 Standard deviation and mean of the region horizontal projection histogram, respectively.
8. A system for intelligently identifying a secondary safety measure regional image based on regional positioning geometric comparison, which is characterized by comprising an image acquisition and feature extraction module, a character identification module and a picture identification and reasoning module, The image acquisition and feature extraction module acquires field device images of secondary safety measures, performs preprocessing on the field device images of the secondary safety measures, extracts salient regions in the images, and obtains image sub-modules through region positioning and module blocking; the character recognition module is based on the image sub-module, performs region positioning and feature extraction, performs character recognition, processes recognition results and outputs character information; the picture recognition and reasoning module is used for carrying out target component recognition on the image submodule based on the text information, extracting component state information to form a structural recognition result, checking the consistency of the recognition result based on a preset secondary safety measure rule and outputting a check conclusion.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, performs the steps of a method for intelligent recognition of a secondary safety measure area image based on area location geometric comparison as defined in any one of claims 1 to 7.
10. A computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the steps of a method for intelligent identification of a secondary safety measure area image based on area location geometric comparison as defined in any one of claims 1 to 7.

Description

Method, system, equipment and medium for intelligent recognition of secondary safety measure area image based on area positioning geometric comparison Technical Field The invention relates to the technical field of power inspection automation, in particular to a method, a system, equipment and a medium for intelligent identification of a secondary safety measure regional image based on regional positioning geometric comparison. Background In the operation and maintenance of a power system, a secondary safety measure is a core link for guaranteeing the maintenance operation safety of places such as a transformer substation, a power plant and the like, and relates to the operations such as isolation, blocking, identification and the like of secondary equipment such as a protection device, a control loop and the like. The implementation and the auditing of the traditional secondary safety measure are highly dependent on manual experience, an operator needs to execute the secondary safety measure one by paper measure, and a professional auditing operator needs to compare text records by site or remotely, so that the problems of low efficiency, tension of human resources, easiness in safety accidents caused by human errors and the like exist. In recent years, with the development of intelligent inspection and image recognition technologies, the power industry starts to explore and utilize visual means to assist safety measure verification, for example, by shooting a state image of a field device and comparing the state image with a standard template, digital supervision of the operation process is realized. However, the existing method focuses on the identification of a single device or a fixed scene, and lacks effective means for judging the overall consistency of multi-region and multi-type safety measure projects in a complex working environment, and has obvious defects in the aspects of cross-screen cross-view image contrast, dynamic environment adaptability, misoperation real-time early warning and the like. Disclosure of Invention The present invention has been made in view of the above-described problems. The invention solves the technical problems that the verification and execution of the secondary safety measures mainly depend on a manual mode, the manual verification has strong dependence, the measures are executed for non-image verification, and the traditional mode only uses paper measure single comparison, so that the traditional mode can not be used for verifying whether the field measures are executed correctly or not, the risk instantiation is insufficient and the error early warning is absent. In order to solve the technical problems, the invention provides a method for intelligently identifying a secondary safety measure area image based on area positioning geometric comparison, which comprises the following steps, Acquiring a field device image of the secondary safety measure, preprocessing the field device image of the secondary safety measure, extracting a salient region in the image, and partitioning the region with a module through region positioning to obtain an image sub-module; Based on the image sub-module, performing region positioning and feature extraction, performing text recognition, processing recognition results, and outputting text information; And carrying out target component identification on the image submodule based on the text information, extracting component state information to form a structural identification result, checking the consistency of the identification result based on a preset secondary safety measure rule, and outputting a verification conclusion. The invention is a preferable scheme of a secondary safety measure area image intelligent identification method based on area positioning and geometric comparison, wherein the field device image of the secondary safety measure is collected, the field device image of the secondary safety measure is preprocessed, the salient area in the image is extracted, and the image submodule is obtained through area positioning and module blocking, Performing space conversion and noise reduction treatment on the secondary safety measure operation image to obtain a treated image; Performing self-adaptive binarization and geometric correction on the processed image to obtain a standardized binary image; and extracting a salient region in the binary image, performing region positioning and module blocking on the binary image, and outputting an image sub-module. The invention relates to a method for intelligently identifying a secondary safety measure regional image based on regional positioning and geometric comparison, which is a preferable scheme, wherein the regional positioning and characteristic extraction are carried out based on an image sub-module, the character recognition is carried out, the recognition result is processed, the output character information comprises, Based on the image sub-module, performing region pos