Oss-based power equipment inspection image automatic collection method

Abstract

The application discloses an OSS-based power equipment inspection image automatic collection method and relates to the technical field of power equipment inspection, which comprises the following steps: collecting on-site power equipment images through a camera, analyzing based on an artificial intelligence algorithm, judging whether the equipment has defects and faults, and then storing the image analysis result in a description file in a standard format; packing the power equipment inspection images and the corresponding description files of the same day, and uploading the corresponding data packet to an OSS platform through a station-end image automatic uploading program the next day; the method has the characteristics of fast reading and writing speed and is beneficial to sharing, can quickly collect massive power equipment inspection images, and the OSS platform receives the uploaded data packet, performs sample collection according to preset collection rules, judges the collection level according to an inspection coefficient, collects the data packets of different collection levels to corresponding cloud storage nodes respectively, is convenient for managers to classify and count, and improves the inspection efficiency and data storage safety.

Description

Technical Field

[0001] This invention relates to the field of power equipment inspection technology, specifically to an automatic image collection method for power equipment inspection based on OSS. Background Technology

[0002] As power inspection operations become increasingly intelligent, remote intelligent inspections are gradually replacing human on-site inspections. Remote intelligent inspections rely on artificial intelligence image recognition technology based on massive amounts of inspection samples. Because power transmission and transformation equipment covers a wide area, has many types, and exhibits complex and diverse defects, current inspection methods still primarily rely on human observation. This results in high workload and risk for frontline inspection personnel. Furthermore, the samples taken by inspection personnel cannot be transmitted back to the back-end system in a timely manner, and the sample data is primarily uploaded and managed manually, leading to low efficiency.

[0003] OSS, also known as cloud object storage service, is a massive, secure, low-cost, and highly reliable cloud storage service. Compared to host servers, it features fast read and write speeds and is easy to share. It can quickly collect massive amounts of power equipment inspection images and store them securely and reliably. Therefore, in order to efficiently collect, manage, and utilize power equipment inspection image data, we propose an automatic collection method for power equipment inspection images based on OSS. Summary of the Invention

[0004] To address the problems of the above solutions, this invention provides an automatic image collection method for power equipment inspection based on OSS.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] An automatic image collection method for power equipment inspection based on OSS includes:

[0007] Step 1: Capture images of the power equipment on site using cameras, analyze them based on artificial intelligence algorithms to determine if there are any defects or faults in the equipment, and then store the image analysis results in a description file according to a standardized format;

[0008] Step 2: Package the power equipment inspection images and corresponding description files from the same day, and upload the corresponding data package to the OSS platform the next day using the station-side image automatic upload program;

[0009] Step 3: The OSS platform receives the uploaded data packets and performs sample aggregation according to preset aggregation rules. Sample aggregation refers to aggregating the data packets to the corresponding cloud storage nodes; specifically:

[0010] The inspection coefficient XG of the data packets is used to evaluate the aggregation level. When XG≤X1, the corresponding data packets are aggregated to the low-level cloud storage node; when X1<XG≤X2, the corresponding data packets are aggregated to the mid-level cloud storage node; when XG>X2, the corresponding data packets are aggregated to the high-level cloud storage node. X1 and X2 are preset fixed values; X1<X2.

[0011] Furthermore, the method for calculating the inspection coefficient XG is as follows:

[0012] Mark the power equipment corresponding to the data packet as the target equipment, mark the total length of the power supply line of the target equipment as NL, the total number of households supplied as ND, and the operating years as NX; obtain the defect and fault information of the target equipment within the sixty days prior to the current system time;

[0013] The number of failures of the target equipment is counted and marked as the failure frequency C1; the failure value QWi of each failure is evaluated according to the failure duration and failure level; the number of times QWi ≥ the failure threshold is counted as the failure exceedance frequency P3, the difference between the corresponding failure value and the failure threshold is calculated, and the difference is summed to obtain the failure exceedance total value CJ; the failure exceedance coefficient QJ is calculated using the formula QJ=P3×k3+CJ×k4, where k3 and k4 are coefficient factors; the data packet inspection coefficient XG is calculated using the formula XG=NL×a1+ND×a2+NX×a3+C1×a4+QJ×a5, where a1, a2, a3, a4, and a5 are coefficient factors.

[0014] Furthermore, the analysis is based on artificial intelligence algorithms to determine whether the equipment has defects or malfunctions, specifically:

[0015] S11: First, the acquired images of power equipment undergo rigorous geometric registration preprocessing;

[0016] S12: Read the preprocessed image using artificial intelligence algorithms, extract the image features, and use them to build a three-dimensional model; the artificial intelligence algorithm is an image recognition algorithm.

[0017] S13: Based on the established three-dimensional model, determine whether the equipment has defects or faults, and obtain image analysis results, which include defect or fault information.

[0018] Furthermore, the steps for creating a three-dimensional model are as follows:

[0019] The edge contour data of the image is analyzed and extracted. Each edge contour data is stored in a vector, and the elements of the vector are the coordinates of the edge contour pixels.

[0020] The direction of the edge contour is obtained based on the inclusion relationship of the edge contour data; the direction of the edge contour is represented as follows: starting from the outermost circle, odd-numbered layers are clockwise and even-numbered layers are counterclockwise.

[0021] The points of the edge contour are compressed using the APPROX_TC89_KCOS algorithm;

[0022] A cross section of the image's 3D model is drawn based on the edge contour, and a 3D solid model is created by stretching the cross section along a given curve direction; 3D feature effects are added to the created 3D solid model; the added 3D feature effects include chamfering, twisting, and scaling.

[0023] Furthermore, in the step of compressing the points of the edge contour, the compression of the edge contour can be completed by simplifying the points selected for the edge contour, reducing redundant data, so as to fully define the curve of the contour edge using the fewest points.

[0024] Furthermore, the simplification of the points on the edge contour includes: selecting two endpoints from several consecutive points on the same straight line to define the line; and using several triangles to approximate a spatial surface, thereby forming a three-dimensional model;

[0025] Furthermore, the description file is an XML format file, and its content includes the site ID, image name, algorithm ID, defect code, and defect coordinates.

[0026] Furthermore, the automatic image upload program at the station connects to the OSS platform via a RESTful API interface and uploads the corresponding data packets to the OSS platform according to the JAVA language interface specifications supported by OSS. The upload quantity and upload time of the automatic image upload program at the station can be flexibly configured.

[0027] Furthermore, it also includes: monitoring cloud storage nodes within the OSS platform for worm viruses, and classifying cloud storage nodes according to their threat rating (CS), specifically:

[0028] The threat score CS is compared with a preset threshold, which includes Y1 and Y2, and Y1 < Y2.

[0029] When CS≤Y1, the corresponding cloud storage node is classified as a high-level cloud storage node; when Y1<CS≤Y2, the corresponding cloud storage node is classified as a mid-level cloud storage node; when CS>Y2, the corresponding cloud storage node is classified as a low-level cloud storage node.

[0030] Furthermore, the threat assessment CS is calculated as follows:

[0031] When a virus is detected, record the virus attack information; count the number of virus attacks on cloud storage nodes within a preset time period as P1; mark the attack duration of each virus attack as Gi and the number of viruses as Li; calculate the attack value GQi using the formula GQi=Gi×d1+Li×d2; where d1 and d2 are coefficient factors.

[0032] The number of times GQi ≥ the attack threshold is counted as the over-attack frequency P2. The difference between the corresponding attack value and the attack threshold is calculated, and the difference is summed to obtain the total over-attack value TQ. The over-attack coefficient CF is calculated using the formula CF = P2 × d3 + TQ × d4, where d3 and d4 are coefficient factors. The threat rating CS of the cloud storage node is evaluated based on the number of virus attacks and the over-attack coefficient.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] 1. In this invention, images of on-site power equipment are collected through sensing terminals such as cameras and robots. Based on artificial intelligence algorithms, the images are analyzed to determine whether there are defects or faults in the equipment. The image analysis results are then stored in a description file according to a standardized format. The power equipment inspection images and corresponding description files of the same day are packaged together and uploaded to the OSS platform the next day through an automatic image upload program at the station. This method features fast read and write speeds, easy sharing, and the ability to quickly collect massive amounts of power equipment inspection images while ensuring secure and reliable storage.

[0035] 2. In this invention, after receiving the uploaded data packet, the OSS platform performs inspection coefficient analysis on the data packet, and then evaluates the aggregation level based on the inspection coefficient. Data packets of different aggregation levels are aggregated to the corresponding cloud storage nodes, which facilitates classification and statistics by management personnel and improves inspection efficiency. At the same time, worm virus monitoring is performed on the cloud storage nodes within the OSS platform. The cloud storage nodes are classified according to their threat rating (CS), and the inspection coefficient of the data packet is matched with the cloud storage node level to improve data storage security. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram illustrating the principle of the automatic image collection method for power equipment inspection based on OSS according to the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] like Figure 1 As shown, the automatic collection method for power equipment inspection images based on OSS includes the following steps:

[0040] Step 1: Collect images of on-site power equipment through sensing terminals such as cameras and robots, analyze them based on artificial intelligence algorithms to determine whether there are defects or faults in the equipment, and then store the image analysis results in a description file in a standardized format;

[0041] The analysis, based on artificial intelligence algorithms, determines whether the equipment has defects or malfunctions. Specifically:

[0042] S11: First, the acquired images of power equipment undergo rigorous geometric registration preprocessing;

[0043] S12: The preprocessed image is read using an artificial intelligence algorithm, image features are extracted, and a 3D model is built based on these features; the artificial intelligence algorithm can be an image recognition algorithm or other AI algorithms; specifically:

[0044] The edge contour data of the image is analyzed and extracted. The data of each edge contour can be stored in a vector, and the elements of the vector are the coordinates of the edge contour pixels.

[0045] The direction of the edge contour is obtained based on the inclusion relationship of the edge contour data. The direction of the edge contour is represented as follows: starting from the outermost circle, odd-numbered layers are clockwise and even-numbered layers are counterclockwise.

[0046] To compress the points of the edge contour, this invention chooses to use the APPROX_TC89_KCOS algorithm to compress the edge contour.

[0047] The cross section of the image is drawn based on the edge contour, and the cross section is stretched along the given curve direction to create a three-dimensional solid model; three-dimensional feature effects are added to the created three-dimensional solid model; the added three-dimensional feature effects can be chamfering, twisting, scaling, etc.

[0048] S13: Determine whether there are defects or faults in the equipment based on the established three-dimensional model, and obtain image analysis results, which include defect and fault information;

[0049] In the step of compressing the points of the edge contour, the compression of the edge contour can be achieved by simplifying the points selected for the edge contour, reducing redundant data, so as to fully define the curve of the contour edge using the fewest points.

[0050] Optionally, simplifying the points of the edge contour includes: selecting two endpoints from several consecutive points on the same straight line to define the straight line; and using several triangles to approximate a spatial surface to form a three-dimensional model.

[0051] The description file is in XML format and includes information such as site ID, image name, algorithm ID, defect code, and defect coordinates. The standard format for the description file is as follows:

[0052]

[0053]

[0054]

[0055] Step 2: Add timestamps to the collected power equipment inspection images and corresponding description files. Based on the timestamps, package the power equipment inspection images and corresponding description files of the same day, and upload the corresponding data package to the OSS platform the next day through the station-side image automatic upload program.

[0056] The station-side image automatic upload program connects to the OSS platform via a RESTful API interface. It uploads the corresponding data packets to the OSS platform according to the JAVA language interface specifications supported by OSS. The upload quantity and upload time of the station-side image automatic upload program can be flexibly configured. It features fast read and write speeds and is easy to share. It can quickly collect massive amounts of power equipment inspection images and store them securely and reliably.

[0057] Step 3: The OSS platform receives the uploaded data packets and performs sample aggregation according to preset aggregation rules. Sample aggregation refers to aggregating data packets to the corresponding cloud storage nodes to facilitate classification and statistics by management personnel and improve inspection efficiency. The preset aggregation rules are as follows:

[0058] The inspection coefficient XG of the data packets is used to evaluate the aggregation level; when XG≤X1, the corresponding data packets are aggregated to the low-level cloud storage node; when X1<XG≤X2, the corresponding data packets are aggregated to the mid-level cloud storage node; when XG>X2, the corresponding data packets are aggregated to the high-level cloud storage node; where X1 and X2 are preset fixed values; X1<X2.

[0059] The method for calculating the inspection coefficient XG is as follows:

[0060] S31: Mark the power equipment corresponding to the data packet as the target device, mark the total length of the power supply line of the target device as NL, mark the total number of households supplied by the target device as ND, and mark the operating years of the target device as NX;

[0061] S32: Obtain defect and fault information of the target device within the previous sixty days of the current system time; the defect and fault include the fault duration and fault level; the fault level is uploaded to the OSS platform after maintenance personnel have completed the repair; the higher the fault level, the more serious the fault problem; count the number of faults of the target device and mark it as the fault frequency C1;

[0062] The duration of each failure of the target equipment is denoted as Ti, and the failure level is denoted as Di.

[0063] The fault value QWi for each fault is evaluated based on the fault duration and fault level, specifically as follows:

[0064]

[0065] In the formula, g1 and g2 are preset coefficient factors;

[0066] Compare the fault value QWi with the fault threshold; count the number of times QWi≥fault threshold as the fault frequency P3; calculate the difference between the corresponding fault value and the fault threshold, and sum the differences to obtain the total fault value CJ; calculate the fault coefficient QJ using the formula QJ=P3×k3+CJ×k4, where k3 and k4 are coefficient factors.

[0067] S33: Normalize the total length of the power supply line, the total number of households supplied, the years of operation, the frequency of faults, and the failure coefficient, and take their values; calculate the inspection coefficient XG of the data packet using the formula XG=NL×a1+ND×a2+NX×a3+C1×a4+QJ×a5, where a1, a2, a3, a4, and a5 are all coefficient factors;

[0068] In this embodiment, the method further includes: monitoring cloud storage nodes within the OSS platform for worm viruses, and classifying cloud storage nodes according to their threat rating (CS); wherein the lower the threat rating (CS), the higher the level of the cloud storage node and the better its security; specifically:

[0069] Existing worm virus monitoring methods are used to monitor cloud storage nodes for worm viruses. When a virus is detected, the virus attack information is recorded, including the attack duration and the number of viruses.

[0070] The number of virus attacks on cloud storage nodes within a preset time period is P1; the attack duration of each virus attack is marked as Gi, and the number of viruses is marked as Li; the attack value GQi is calculated using the formula GQi=Gi×d1+Li×d2; where d1 and d2 are coefficient factors.

[0071] Compare the attack value GQi with the attack threshold, and count the number of times GQi≥ the attack threshold as the over-attack frequency P2. Calculate the difference between the corresponding attack value and the attack threshold, and sum the differences to obtain the total over-attack value TQ. Use the formula CF=P2×d3+TQ×d4 to calculate the over-attack coefficient CF, where d3 and d4 are coefficient factors.

[0072] The threat rating (CS) of cloud storage nodes is assessed based on the number of virus attacks and the over-attack coefficient, specifically as follows: In the formula, g3 and g4 are preset coefficient factors;

[0073] The threat score CS is compared with a preset threshold, which includes Y1 and Y2, and Y1 < Y2.

[0074] When CS≤Y1, the corresponding cloud storage node will be classified as an advanced cloud storage node;

[0075] When Y1 < CS ≤ Y2, the corresponding cloud storage node is classified as a mid-level cloud storage node;

[0076] When CS > Y2, the corresponding cloud storage node will be classified as a low-level cloud storage node.

[0077] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0078] Working principle of the invention:

[0079] The OSS-based automatic image collection method for power equipment inspection first collects images of on-site power equipment through sensing terminals such as cameras and robots. Then, it analyzes the images based on artificial intelligence algorithms to determine whether there are defects or faults in the equipment. The image analysis results are then stored in a description file according to a standardized format. The power equipment inspection images and corresponding description files of the same day are then packaged together and uploaded to the OSS platform the next day through the station-side image automatic upload program.

[0080] Meanwhile, after receiving the uploaded data packets, the OSS platform performs inspection coefficient analysis on the data packets, and then judges the aggregation level based on the inspection coefficient. Data packets of different aggregation levels are aggregated to the corresponding cloud storage nodes, which facilitates classification and statistics by management personnel and improves inspection efficiency. At the same time, the OSS platform monitors cloud storage nodes for worm viruses, classifies cloud storage nodes according to their threat rating (CS), and matches the inspection coefficient of data packets with the cloud storage node level to improve data storage security.

[0081] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0082] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An automatic image collection method for power equipment inspection based on OSS, characterized in that, include: Step 1: Capture images of the power equipment on site using cameras, analyze them based on artificial intelligence algorithms to determine if there are any defects or faults in the equipment, and then store the image analysis results in a description file according to a standardized format; The analysis, based on artificial intelligence algorithms, determines whether the equipment has defects or malfunctions. Specifically: S11: First, perform geometric registration preprocessing on the acquired power equipment images; S12: Read the preprocessed image using artificial intelligence algorithms, extract the image features, and use them to build a three-dimensional model; the artificial intelligence algorithm is an image recognition algorithm. The steps for creating a three-dimensional model are as follows: The edge contour data of the image is analyzed and extracted. Each edge contour data is stored in a vector, and the elements of the vector are the coordinates of the edge contour pixels. The direction of the edge contour is obtained based on the inclusion relationship of the edge contour data; the direction of the edge contour is represented as follows: starting from the outermost circle, odd-numbered layers are clockwise and even-numbered layers are counterclockwise. The points of the edge contour are compressed using the APPROX_TC89_KCOS algorithm; Based on the edge contour, a cross section of the image 3D model is drawn, and the cross section is stretched along a given curve direction to create a 3D solid model; 3D feature effects are added to the created 3D solid model. The added 3D feature effects include chamfering, twisting, and scaling; S13: Determine whether the equipment has defects or faults based on the established three-dimensional model, and obtain image analysis results, which include defect or fault information; Step 2: Package the power equipment inspection images and corresponding description files from the same day, and upload the corresponding data package to the OSS platform the next day using the station-side image automatic upload program; Step 3: The OSS platform receives the uploaded data packets and performs sample aggregation according to preset aggregation rules. Sample aggregation refers to aggregating the data packets to the corresponding cloud storage nodes; specifically: The inspection coefficient XG of the data packets is used to evaluate the level of collection; The method for calculating the inspection coefficient XG is as follows: Mark the power equipment corresponding to the data packet as the target device, mark the total length of the power supply line of the target device as NL, the total number of households supplied as ND, and the operating years as NX; Obtain defect and fault information of the target device within the sixty days prior to the current system time; The number of failures of the target equipment is counted and marked as the failure frequency C1; the failure duration of each failure of the target equipment is marked as Ti, and the failure level is marked as Di. The fault value QWi for each fault is evaluated based on the fault duration and fault level, specifically as follows: In the formula, g1 and g2 are preset coefficient factors; The number of times QWi ≥ the fault threshold is counted as the fault frequency P3. The difference between the corresponding fault value and the fault threshold is calculated, and the difference is summed to obtain the total fault value CJ. The obstacle-crossing coefficient QJ is calculated using the formula QJ=P3×k3+CJ×k4, where k3 and k4 are coefficient factors. The inspection coefficient XG of the data packet is calculated using the formula XG=NL×a1+ND×a2+NX×a3+C1×a4+QJ×a5, where a1, a2, a3, a4, and a5 are all coefficient factors. When XG≤X1, the corresponding data packets will be aggregated to the lower-level cloud storage node; When X1 < XG ≤ X2, the corresponding data packets will be aggregated to the intermediate cloud storage node; When XG > X2, the corresponding data packets will be aggregated to the advanced cloud storage node; where X1 and X2 are preset fixed values; X1 < X2; The method also includes: monitoring cloud storage nodes within the OSS platform for worm viruses, and classifying cloud storage nodes according to their threat rating (CS), specifically: The threat score CS is compared with a preset threshold, which includes Y1 and Y2, and Y1 < Y2. When CS≤Y1, the corresponding cloud storage node will be classified as an advanced cloud storage node; When Y1 < CS ≤ Y2, the corresponding cloud storage node is classified as a mid-level cloud storage node; When CS > Y2, the corresponding cloud storage node will be classified as a low-level cloud storage node; The threat assessment (CS) is calculated as follows: When a virus is detected, record the virus attack information; count the number of virus attacks on cloud storage nodes within a preset time period as P1; mark the attack duration of each virus attack as Gi and the number of viruses as Li; calculate the attack value GQi using the formula GQi=Gi×d1+Li×d2; where d1 and d2 are coefficient factors. The number of times GQi ≥ the attack threshold is called the over-attack frequency P2. The difference between the corresponding attack value and the attack threshold is calculated, and the difference is summed to obtain the total over-attack value TQ. The over-attack coefficient CF is calculated using the formula CF=P2×d3+TQ×d4, where d3 and d4 are coefficient factors. The threat rating (CS) of cloud storage nodes is assessed based on the number of virus attacks and the over-attack coefficient, specifically as follows: In the formula, g3 and g4 are preset coefficient factors.

2. The method for automatic collection of power equipment inspection images based on OSS according to claim 1, characterized in that, The description file is in XML format and includes the site ID, image name, algorithm ID, defect code, and defect coordinates.

3. The method for automatic collection of power equipment inspection images based on OSS according to claim 1, characterized in that, The automatic image upload program at the station connects to the OSS platform via a RESTful API interface. It uploads the corresponding data packets to the OSS platform according to the JAVA language interface specifications supported by OSS. The upload quantity and upload time of the automatic image upload program at the station can be flexibly configured.

Images