Detection and identification method and system for electric power material image

Abstract

The invention discloses an electric power material image detection and identification method and system, and relates to the technical field of electric power material detection, and the method comprises the steps: S1, obtaining an electric power material image when a material moves, sorting the electric power material image according to the obtained flow and position, and obtaining a material image set related to the flow position; s2, performing image detection on the material image set, obtaining a feature set corresponding to the material image set, performing linear regression analysis on the feature set, clustering the feature set according to a linear regression analysis result to obtain a feature clustering result, and setting an electric power material label of the current electric power material based on the feature clustering result; s3, identifying the relative position of a feature vector corresponding to the feature clustering result on the electric power material image, and obtaining the feature coverage and the overlapping area ratio of the feature clustering result on each surface of the adjacent electric power material image; the accuracy and efficiency of electric power material identification can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power material detection, and in particular to a method and system for detecting and identifying electric power material images. Background Art

[0002] Power material storage management is an important part of power enterprise management. Power material storage management involves storage planning, storage management, storage environment adjustment and other links. Among them, power material storage management is a necessary condition to ensure the normal operation of power production and power dispatching. Through optimized management, the operating efficiency and profitability of power production enterprises can be improved.

[0003] For example, Chinese patent publication number CN114971501A discloses a monitoring guidance and analysis system for the entry and exit of electric power materials based on feature analysis. Before the handling robot performs the warehousing operation on each incoming electric power material, the stacking state image is collected, and the stacking height corresponding to each incoming electric power material is identified thereby. At the same time, the entry distance corresponding to each incoming electronic material is identified based on the designated sub-storage area corresponding to each incoming electric power material. Thus, the warehousing sequence is comprehensively planned according to the entry distance and stacking height corresponding to each incoming electric power material. This takes into account the principle of nearest warehousing and avoids damage to electric power materials before use, thereby achieving dual guarantees for warehousing efficiency and warehousing safety. At the same time, when the handling robot transports the incoming electric power materials, the integrated operation of drop monitoring, early warning and processing during the warehousing and transportation of the electric power materials is achieved.

[0004] For example, Chinese patent publication number CN116308047A discloses an electric power material in-and-out warehousing management system based on RFID technology, which relates to the technical field of material management and solves the technical problem that a large amount of electric power materials are stored at the same time, and some electric power material packages are disassembled for unknown reasons, resulting in internal materials being swapped, thereby causing logistics losses. According to the overall image of the surface where the RFID tag is located, the area ratio parameters of the corresponding area are obtained, and the same method is used for processing in the later stage during the outbound process. Under different turning conditions, the obtained area ratio parameters are in a consistent state, and there will be no numerical errors due to the different directions of the obtained overall image. At the same time, by processing in this way, it is possible to quickly identify whether the material package is lost or unpacked during the outbound process of electric power materials.

[0005] In the prior art, the identification of electric power materials is completed by the parameters such as the storage order and length value of the electric power materials. However, when multiple electric power materials with similar characteristics are piled in one location, the differences between these electric power materials are easily ignored, resulting in identification errors when tracking and identifying the electric power materials, resulting in reduced efficiency in the management of electric power materials. Summary of the invention

[0006] The present application solves the technical problem of accuracy in tracking electric power materials in the prior art by providing a method and system for detecting and identifying electric power material images, thereby improving the accuracy of electric power material identification.

[0007] The present application provides a method for detecting and identifying an image of electric power materials, comprising:

[0008] S1, acquiring power material images when the materials are moved, sorting the power material images according to the acquired processes and positions, and obtaining a material image set related to the process positions.

[0009] S2, performing image detection on the material image set, obtaining a feature set corresponding to the material image set, performing linear regression analysis on the feature set, clustering the feature set according to the linear regression analysis result to obtain a feature clustering result, and setting an electric power material label for the current electric power material based on the feature clustering result.

[0010] S3, identifying the relative position of the feature vector corresponding to the feature clustering result on the power material image, and obtaining the feature coverage and overlapping area ratio of the feature clustering result on each surface of the adjacent power material image, matching the feature coverage and overlapping area ratio with the power material label, and obtaining the comprehensive processing label of the current power material at different priorities.

[0011] S4, based on the comprehensive processing tags, tracks the power material image, determines the current distribution position and distribution probability of the power material, and obtains the distribution feedback result of the power material.

[0012] The present invention provides a detection and recognition system for power material images, comprising:

[0013] The image acquisition module is used to acquire images of electric power materials during movement, and sort the images according to the acquired process and position to form a material image set related to the process position.

[0014] The image detection module is used to perform image detection on the material image set and extract the feature set; construct a linear regression model and perform linear regression analysis on the feature set; cluster the feature set according to the linear regression analysis results to obtain the feature clustering results; and set the power material label of the current power material based on the feature clustering results.

[0015] The feature processing module is used to identify the relative position of the feature vector corresponding to the feature clustering result on the power material image; calculate the feature coverage and overlapping area ratio, and match them with the power material label; combine the posture angle of the power material to generate the comprehensive matching result of the power material; according to the comprehensive matching result, obtain the comprehensive processing label of the current power material at different priorities.

[0016] The location verification module is used to obtain the current distribution location of power materials and extract the tracking displacement value; verify whether the power materials are at the optimal location point and calculate the displacement evaluation coefficient; and output the distribution feedback result.

[0017] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0018] Through image detection and feature clustering technology, the appearance, transportation requirements and batch processing characteristics of power materials can be identified quickly and accurately; by calculating the feature coverage and overlapping area ratio, combined with the posture angle of the power materials, the feature vector can be accurately matched; based on the matching results, comprehensive processing labels are generated to provide a reliable basis for the subsequent processing of power materials; by calculating the displacement evaluation coefficient, it is verified whether the power materials are in the optimal position; and the distribution and management of power materials are optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The figure is a flow chart of a method for detecting and identifying an image of electric power materials according to the present invention.

[0020] Figure 2 The figure is a flow chart of step S2 of a method for detecting and identifying an image of electric power materials according to the present invention.

[0021] Figure 3 A schematic flow chart of step S3 of a method for detecting and identifying an image of electric power materials according to the present invention.

[0022] Figure 4 A schematic flow chart of step S35 of a method for detecting and identifying an image of electric power materials according to the present invention.

[0023] Figure 5 A schematic flow chart of step S4 of a method for detecting and identifying an image of electric power materials according to the present invention.

[0024] Figure 6 A system framework diagram of a power material image detection and recognition system of the present invention. DETAILED DESCRIPTION

[0025] To facilitate the understanding of the present invention, the present application will be described more comprehensively below with reference to the relevant drawings; the drawings show preferred embodiments of the present invention, but the present invention can be implemented in many different forms and is not limited to the embodiments described herein; on the contrary, the purpose of providing these embodiments is to enable a more thorough and comprehensive understanding of the disclosed content of the present invention.

[0026] It should be noted that the terms “vertical”, “horizontal”, “up”, “down”, “left”, “right” and similar expressions used in this document are only for illustrative purposes and do not represent the only implementation method.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which the present invention belongs; the terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention; the term "and / or" used herein includes any and all combinations of one or more related listed items.

[0028] In the present invention, when the images of electric power materials are sorted, if the electric power materials are identified according to the length, width, height and area of ​​the electric power materials identified at this time, it is easy to ignore the condition of the electric power materials during transportation. For example, if the electric power materials currently need to be transported smoothly, it is necessary to verify the smoothness of the electric power materials during transportation. If the materials need to be transported quickly, it is necessary to pay attention to the batches and priorities of the materials. Therefore, when identifying the electric power materials at this time, it is necessary to give priority to obtaining the order and characteristics corresponding to the materials, and represent the usage characteristics of the electric power materials to complete the rapid identification of the materials.

[0029] Therefore, when the present invention processes electric power materials, it first constructs a linear regression model of electric power materials, performs regression analysis on the features extracted from electric power materials, determines the linear relationship between the current electric power materials in sequence, and the feature set corresponding to the electric power materials in sequence according to the feature set, and clusters the data in the electric power materials in accordance with the current processing method, and determines the additional features of the current materials according to the position of the image acquisition surface and the six-sided map where the electric power materials are located, and matches these features obtained from different surfaces with the labels of the materials themselves, obtains the matching degree and priority association degree of the materials under specific priorities, thereby obtaining the comprehensive processing label of the materials, and processes the current electric power materials in and out of the warehouse according to this label, thereby completing the rapid identification and processing of the electric power materials. The matching degree indicates the matching situation between these additional features and the labels of the materials themselves, and the priority association degree indicates the influence and priority representation of the features of the current electric power materials on the six-sided map relative to the current electric power materials.

[0030] At the same time, these processing methods tend to recognize the posture of electronic power materials themselves to classify power materials with similar shapes or sizes together, which helps to optimize the use of storage space and reduce the risk of collision and damage between power materials. At the same time, according to the posture of the power materials themselves, problems related to power materials can be located more quickly, and improperly transported power materials can be discovered in time, which can reduce the problem of power material collapse or handling errors when the power materials are placed.

[0031] like Figure 1 As shown, the present application provides a method for detecting and identifying an image of electric power materials, including:

[0032] S1, acquiring power material images when the materials are moved, sorting the power material images according to the acquired processes and positions, and obtaining a material image set related to the process positions.

[0033] S2, performing image detection on the material image set, obtaining a feature set corresponding to the material image set, performing linear regression analysis on the feature set, clustering the feature set according to the linear regression analysis result, obtaining a feature clustering result, and setting an electric power material label for the current electric power material based on the feature clustering result.

[0034] like Figure 2 As shown, step S2 includes: S21, obtaining appearance recognition features for the appearance of the current power material, demand features for the transportation of the power material, and batch processing features in the feature set; constructing a linear regression model related to the appearance recognition features, demand features, and batch processing features, the appearance features represent the feature information related to the appearance and texture of the current power material, the demand features display the feature information of the current power material during placement, storage, transportation, and use, and the batch processing features represent the priority of the current material processing and the logical order of the material during processing.

[0035] For example, appearance recognition features refer to visual features extracted from power material images that can be directly used to identify and classify objects; these features are usually related to the physical properties of the object, such as shape, color, texture, etc.

[0036] Appearance recognition features can be in the following forms:

[0037] Color histogram: describes the color distribution on the surface of electrical materials and can be used to distinguish different types of equipment or components.

[0038] Edge detection: Detect the outline of objects through the Canny operator or other methods to help determine the shape structure of electrical materials.

[0039] Texture analysis: Quantify surface texture characteristics using techniques such as LBP (Local Binary Pattern), suitable for identifying specific materials or manufacturing processes.

[0040] Brand logo / barcode: Locating and reading the identification information on power supplies helps to quickly confirm their identity.

[0041] Shape descriptors: such as Hu moments and Fourier descriptors, can be used to describe complex shapes for accurate matching and classification.

[0042] For example, demand characteristics reflect the functional requirements and technical specifications of power materials in specific application scenarios; these characteristics usually involve information on performance parameters, installation conditions, maintenance standards, etc.

[0043] Requirement characteristics can be the following:

[0044] Power level: Indicates the maximum electrical energy output that electrical materials can provide, which is particularly important for key equipment such as transformers and generators.

[0045] Voltage range: specifies the applicable working voltage range of electrical materials to ensure correct selection and safe use.

[0046] Protection level: describes the protection ability of power materials against external environmental factors (such as dust and water), which affects the choice of their deployment location.

[0047] Size and weight: Determine whether electrical supplies are suitable for specific space constraints and handling methods.

[0048] Connection interface type: such as bolt connection, plug-in connector, etc., affects compatibility with other systems and ease of installation.

[0049] Service Life and Warranty: Provides information on the long-term reliability of electrical supplies to guide purchasing decisions and service plans.

[0050] Batch processing characteristics refer to attributes related to the same production batch, including but not limited to production date, supplier information, quality inspection results, etc. Such characteristics are mainly used for supply chain management and quality traceability.

[0051] For example, batch processing features can be in the following form:

[0052] Production batch number: Each batch has a unique number to facilitate tracking the flow and history of all products in the batch.

[0053] Production date: Marks the manufacturing time of electrical supplies, which helps to assess their freshness or shelf life.

[0054] Supplier information: records the name and address of the company that provides raw materials or finished products, supporting transparent supply chain management.

[0055] Quality inspection report: contains various test data and certificates of conformity to ensure that product quality meets established standards.

[0056] Logistics information: Track the transportation path of power materials from the factory to the end user to ensure timely delivery and reduce the risk of loss.

[0057] Inventory status: Real-time update of quantity changes of power materials to assist in warehouse management and allocation decisions.

[0058] These features can be identified using image recognition algorithms, and power inspection robots can be used to obtain in real time the power materials currently being transported or stored, and these identified features can be sent to an external processing terminal, which will then perform comprehensive processing on these identified power material images.

[0059] S22, analyzing the appearance recognition feature, the demand feature, and the batch processing feature in sequence, and obtaining a first correlation coefficient corresponding to the appearance recognition feature, a second correlation coefficient corresponding to the demand feature, and a third correlation coefficient corresponding to the batch processing feature in sequence.

[0060] The calculation method of the first correlation coefficient is expressed as follows: the appearance recognition features in the material image set are taken as input, and the cosine similarities between the appearance recognition features in adjacent power material images are judged in turn. The calculation method of the cosine similarity is to represent the appearance recognition features in the form of vectors to obtain the cosine similarity; the calculated cosine similarity is used as the first correlation coefficient; the calculation methods of the second correlation coefficient and the third correlation coefficient are the same, and both use the cosine similarities of the corresponding demand features and batch processing features in adjacent power material images to judge the similarity between related images in adjacent power material images.

[0061] S23, performing linear regression on the first correlation coefficient, the second correlation coefficient, and the third correlation coefficient to obtain regression coefficients corresponding to the first correlation coefficient, the second correlation coefficient, and the third correlation coefficient.

[0062] At this time, the regression coefficient method is to use the second correlation coefficient and the third correlation coefficient as independent variables and the first correlation coefficient as the dependent variable to perform linear regression analysis to determine the corresponding changes of the actually stored and transported materials in the adjacent power material images when the current priority of material transportation and transmission changes.

[0063] When performing linear regression analysis, the main focus is on judging the probability of occurrence of the values ​​corresponding to the first correlation coefficient, the second correlation coefficient and the third correlation coefficient in continuous power material images, and comparing the relative positions of the image features existing therein relative to the continuous image recognition, so as to complete the linear regression analysis of the first correlation coefficient, the second correlation coefficient and the third correlation coefficient.

[0064] For example, a first correlation coefficient C1, a second correlation coefficient C2, and a third correlation coefficient C3 of corresponding features are extracted from the feature set; and a linear regression model is formed by the first correlation coefficient, the second correlation coefficient, and the third correlation coefficient.

[0065] For the appearance recognition features, demand features and batch processing features in the feature set, these features are matched one by one, and the first correlation coefficient, second correlation coefficient and third correlation coefficient related to these features are obtained. These coefficients are calculated to obtain a linear regression model to obtain the linear regression coefficient.

[0066] C1=β0+β1C2+β2C3+ε;

[0067] Among them, β0 represents the intercept term, β1 and β2 represent the regression coefficients, and ε represents the error term.

[0068] The obtained regression coefficients are estimated using the least squares method to obtain the minimized residual sum of squares RSS (β) of the regression coefficients.

[0069]

[0070] And use the matrix to get the optimal regression coefficient value

[0071]

[0072] Among them, X represents the design matrix, Y contains all independent variables C2 and C3; represents the vector of dependent variable C1, X T Represents the transpose operation of the design matrix; the optimal regression coefficient value calculated at this time represents the optimal regression coefficient value corresponding to these data under the corresponding number of independent variables and dependent variables, and this optimal regression coefficient value is substituted into the intercept term that minimizes the residual sum of squares to obtain a set of regression coefficients that are most relevant to the current independent variables and dependent variables.

[0073] Specifically Wherein, X and Y contain n rows of values, and each row of values ​​represents a set of corresponding first correlation coefficients, second correlation coefficients, and third correlation coefficients.

[0074] When the minimized residual sum of squares of the regression coefficient and the optimal regression coefficient value are both satisfied, the corresponding regression coefficient is output, thereby completing the combined analysis of the first correlation coefficient, the second correlation coefficient and the third correlation coefficient.

[0075] When the output regression coefficient is obtained, it is also necessary to determine the first correlation coefficient, the second correlation coefficient and the third correlation coefficient under the corresponding regression coefficient, and compare the relative positions of the first correlation coefficient, the second correlation coefficient and the third correlation coefficient with respect to the adjacent power material images, and use the set of first correlation coefficients, second correlation coefficients and third correlation coefficients with the smallest relative position distance and the highest probability of occurrence as the initial clustering center for subsequent clustering calculations.

[0076] S24, according to the range of values ​​of the regression coefficient, set multiple clusters, the initial cluster center of each cluster is represented by the weighted average of the first correlation coefficient, the second correlation coefficient and the third correlation coefficient, calculate the distance between each element in the feature set and the initial cluster center, that is, calculate the distance value between the feature vector in the feature set and the initial cluster center, calculate the first correlation coefficient, the second correlation coefficient and the third correlation coefficient of all members in each cluster, and use the weighted average of these correlation coefficients to recalculate the new center position of the cluster; iterate the cluster center continuously until the cluster center meets the convergence condition; the convergence condition is to reach a preset maximum number of iterations, and the maximum number of iterations can be set by selecting the average value for power material image processing in historical data; output the iterated cluster cluster as the feature clustering result.

[0077] The role of the regression coefficient is to guide the setting of the initial clustering clusters to ensure that there is a reasonable degree of distinction between different clusters; by considering the influence of multiple correlation coefficients, the cluster center can comprehensively reflect the correlation between members.

[0078] After the clustering process is completed, an electric power material label is set for each element in the feature clustering result according to the feature clustering result.

[0079] S3, identifying the relative position of the feature vector corresponding to the feature clustering result on the power material image, and obtaining the feature coverage and overlapping area ratio of the feature clustering result on each surface of the adjacent power material image, matching the feature coverage and overlapping area ratio with the power material label, and obtaining the comprehensive processing label of the current power material at different priorities.

[0080] When obtaining the feature coverage and overlapping area ratio, the feature coverage and overlapping area ratio at the current posture angle of the power material will be calculated to identify the specific placement of the current power material, and based on these identified placement conditions, the power material will be comprehensively identified to complete the identification processing of the power material.

[0081] At the same time, the adjacent power material images will be compared with the surfaces where the current power materials are placed, that is, the relative position of the position corresponding to the power material image in the six surfaces of the current power materials, and the distribution of the corresponding features in the feature clustering results on these six surfaces will be identified to identify which surface the current power materials mainly rely on for features during centralized identification.

[0082] like Figure 3 As shown, the implementation of step S3 includes:

[0083] S31, obtaining the feature clustering result of the current electric power material, and mapping the feature clustering result to the six planes corresponding to the electric power material, and determining the relative position of each feature vector in the feature clustering result in the electric power material.

[0084] S32, calculating the feature coverage of the feature vector on each surface, where the feature coverage represents the ratio of the pixel ratio occupied by the current feature vector to the sum of the pixel ratios occupied by all feature vectors on the surface.

[0085] S33, calculating the overlapping area ratio between the feature vector and the adjacent power material image at the corresponding feature coverage, where the overlapping area ratio represents the intersection-and-union ratio of contours between multiple faces in the power material image.

[0086] S34, obtaining the attitude angle of the current electric power material, comparing the feature coverage and overlapping area ratio of the feature vector with the attitude angle of the current electric power material, and determining a matching result between the feature vector and the attitude angle.

[0087] S35, combining the matching results on the six surfaces corresponding to the electric power material according to the matching results of the feature vector and the attitude angle to generate a comprehensive matching result of the electric power material; and obtaining a comprehensive processing label according to the comprehensive matching result of the electric power material.

[0088] In this step, the relative position of the current power material in the six planes is identified by comparing the feature distribution in the adjacent power material images; the distribution of the corresponding features in the feature clustering results on the six planes is analyzed to find the most representative plane and the features on it.

[0089] The above feature coverage can be expressed as: Among them, C f represents the feature coverage on the fth surface, Ft f Indicates the pixel ratio of the feature vector on the fth face, Ft f,total It represents the sum of the pixel ratios of all feature vectors on the fth surface. The feature coverage calculated at this time is for a single feature vector to identify which of these feature vectors is outstanding in the current surface, so that the existing electric power materials can be quickly identified based on the differences in the six surfaces corresponding to the goods.

[0090] The overlapping area ratio is the ratio of the contours between two faces. It is used to measure the overlap between two different faces or regions, especially when evaluating the feature matching between adjacent power material images, and is used to evaluate the feature consistency of the same object under different viewing angles. For example, when comparing different faces of the same power material in two adjacent images, it is recognized that these faces belong to the same object, and their relative positions remain consistent, and when there are important features on multiple faces, the intersection-union ratio can indicate which face can provide more information.

[0091] When determining the comparison result between the feature vector and the attitude angle, the absolute value of the difference between the feature coverage and overlapping area ratio of the feature vector in the absence of the attitude angle and the feature coverage and overlapping area ratio at the current attitude angle is taken as the matching result of the feature vector and the attitude angle at this time; then, the comprehensive matching result of the power materials is obtained based on the matching result of the feature vector and the attitude angle.

[0092] Then the matching result can be expressed as the matching difference value of the feature coverage and overlapping area ratio of the feature vector at the corresponding attitude angle. Based on the obtained matching difference value, the matching score of the current feature vector is calculated, and the matching score is output as the matching result of the feature vector and the attitude angle.

[0093] The matching difference value of feature coverage is expressed as, ΔC f =|C f _C f,θ |; where ΔC f represents the matching difference value of the feature coverage on the fth surface, θ represents the posture angle, C f,θ It represents the feature coverage when there is an attitude angle θ on the f-th surface.

[0094] At this time, the relevant attitude angles of the power materials are compared to identify the actual placement of these power materials and the corresponding situations of the power material images taken by the current inspection equipment. Because the inspection equipment will not adopt a completely parallel form with the power materials when taking images of power materials, it is necessary to identify the specific characteristics of the corresponding power materials when the placement is offset or the viewing angle is offset during shooting, so as to quickly identify these related power materials.

[0095] The matching difference value of the overlap area ratio is expressed as, ΔI ff′ =|IoU ff′ _IoU ff′,θ |; where ΔI ff′ Represents the matching difference value of the overlapping area ratio of the fth face and the f′th face, IoU ff′ represents the fth face and the fth face ′ The overlapping area ratio of the faces, IoU ff′,θ represents the fth face and the fth face ′ The overlapping area ratio of the faces with attitude angle θ, f′ represents the faces other than face f.

[0096] At this time, the matching score is expressed as, S f represents the matching score of the fth face, F represents the number of faces in the power material image, e represents the exponential constant, α represents the adjustment coefficient, and the adjustment coefficient is used to control the impact of the difference on the matching score; W f represents the weight of the f-th face, Wf′ Represents the weight of the f′th face; the weights of different faces on the power material can be set according to the ratio of the eigenvector on the corresponding face to the standard eigenvector, and the standard eigenvector is the average value of the eigenvector set in the historical data; the value range of f and f′ is 1 to F.

[0097] The comprehensive matching result of electric power materials can be expressed as: comprehensively calculating the matching results of the feature vector and the attitude angle to obtain a comprehensive matching score, and outputting the comprehensive matching score as the comprehensive matching result.

[0098]

[0099] Among them, S total It represents the comprehensive matching score. The comprehensive matching score obtained at this time can reflect the comprehensive situation of the electric power materials identified in the current electric power material image on multiple surfaces, and output these identified comprehensive situations to represent the corresponding comprehensive characteristics of the current electric power materials; according to the value corresponding to the comprehensive matching result, a comprehensive processing label is set for the currently identified electric power materials to represent the corresponding characteristics of the current electric power materials.

[0100] When setting the comprehensive processing tag, it is also necessary to match the comprehensive matching results of the power materials with the power material tags to select the corresponding comprehensive processing tag.

[0101] like Figure 4 As shown, step S35 also includes S351, according to the comprehensive matching result of the electric power material, matching the comprehensive matching result with the electric power material tag to obtain at least one matching point.

[0102] First, the matching points between the comprehensive matching results and the power material labels are obtained. These matching points represent a certain association or similarity between the comprehensive matching results and the power material labels; then, the system will allocate these matching points, that is, assign a unique identifier to each matching point and record their corresponding information, such as power material labels, comprehensive matching results, etc.

[0103] In this step, the matching method of the power material label and the comprehensive matching result is restricted according to the common occurrence probability of the power material label and the comprehensive matching result, and the matching points are set according to the common occurrence probability of the power material label and the comprehensive matching result, and each matching point corresponds to a value of the common occurrence probability of the power material label and the comprehensive matching result.

[0104] S352, performing allocation processing on the matching points, identifying first matching thresholds corresponding to the corresponding matching points in turn; and screening out a plurality of target points from the matching points according to the first matching thresholds.

[0105] For each matching point, the system will identify its corresponding first matching threshold. This threshold may be a preset value used to measure whether the matching point meets a certain condition or standard. Based on the first matching threshold, the system will filter out multiple target points that meet the conditions from all matching points. These target points represent the power material tags that are closest or similar to the comprehensive matching results.

[0106] In this step, a threshold of the probability of co-occurrence of power material labels and comprehensive matching results is set. This threshold is the first matching threshold. The first matching threshold can be set by selecting the average value of the probability of co-occurrence of power material labels and comprehensive matching results in historical data; and the matching points greater than the first matching threshold are set as target points.

[0107] S353, performing difference analysis on the target points, and taking the power material label corresponding to the point with the largest difference value after the difference analysis as the output comprehensive processing label.

[0108] Variance analysis is a statistical method used to compare data differences between different samples or groups. For power material labels, variance analysis may involve comparing the differences in quantity, quality, price, supply capacity, etc. of power materials corresponding to different labels. Calculate the difference between each target point and other target points, and find the point with the largest difference. This point represents the power material label with the largest difference from the comprehensive matching result.

[0109] In this step, the corresponding values ​​at different target points are calculated. The values ​​set at the target points will select the values ​​of the eigenvectors corresponding to the comprehensive processing results and the power material labels to select the power material labels when the difference value is the largest, thereby obtaining the output comprehensive processing labels.

[0110] S4, based on the comprehensive processing tags, tracks the power material image, determines the current distribution position and distribution probability of the power material, and obtains the distribution feedback result of the power material.

[0111] For the allocation position, the specific coordinates of the power materials in the corresponding spatial position are verified to determine whether the position of the power materials will change after changes or the corresponding deployment settings are completed, and the displacement values ​​of these changes are regarded as tracking displacement values; the allocation probability is to determine the probability of the power materials being allocated at different positions under the condition of the characteristic vector of the current power materials. Combined with this probability value, the location where the current power materials need to be placed can be identified in time.

[0112] like Figure 5 As shown, the implementation method of this step includes, S41, obtaining the current distribution position of the power material, and extracting the tracking displacement value related to the distribution position;

[0113] S42, combining the tracking displacement value with the distribution probability of the electric power material to verify whether the electric power material is at the optimal position point; and calculating the displacement evaluation coefficient of the electric power material at the optimal position point, and outputting the data corresponding to the displacement evaluation coefficient as the distribution feedback result.

[0114] In this step, the method of verifying whether the power material is at the optimal position point is to obtain the displacement deviation of the current power material, combine the displacement deviation with the allocation probability, and calculate the displacement evaluation coefficient.

[0115]

[0116] Among them, δ represents the displacement assessment coefficient, D represents the tracking displacement value, ΔDI represents the displacement deviation, ΔDI represents the standard value of the displacement deviation, at this time the displacement deviation represents the difference between the currently obtained tracking displacement value and the predicted tracking displacement value, and the standard value of the displacement deviation is the average value of the displacement deviation in the historical data; P(DI) represents the distribution probability of power materials, α1, α2 and α3 represent the weight coefficients respectively, and the values ​​of the three weight coefficients α1, α2, α3 are set to 0.25, 0.35, and 0.4 in sequence.

[0117] At this time, the actual displacement of the power material and its distribution probability at different locations are taken into account. Whenever the position of the power material changes, the new coordinates are immediately written into the database; the identified new coordinates are tracked, the distribution probability of the power material at different locations is evaluated, the best placement location is selected and real-time feedback is generated. Finally, all position changes will be recorded in the database, and any abnormal situation will trigger the corresponding alarm mechanism to ensure the safety and reliability of the entire process.

[0118] The obtained displacement assessment coefficient will be reflected in the relative distribution of these electric power materials during tracking when using the above identification of the six surfaces of the electric power materials, so as to update the position coordinates of the electric power materials in real time and promptly verify whether the currently identified electric power materials are accurate. The output distribution feedback result will include the currently identified displacement assessment coefficient and the feature vectors of the corresponding electric power materials on different surfaces, thereby assisting the external terminal in identifying the electric power materials.

[0119] like Figure 6 As shown, the present invention also provides a detection and recognition system for power material images, comprising:

[0120] The image acquisition module is used to acquire images of electric power materials during movement, and sort the images according to the acquired process and position to form a material image set related to the process position.

[0121] The image detection module is used to perform image detection on the material image set and extract the feature set; construct a linear regression model and perform linear regression analysis on the feature set; cluster the feature set according to the linear regression analysis results to obtain the feature clustering results; and set the power material label of the current power material based on the feature clustering results.

[0122] The feature processing module is used to identify the relative position of the feature vector corresponding to the feature clustering result on the power material image; calculate the feature coverage and overlapping area ratio, and match them with the power material label; combine the posture angle of the power material to generate the comprehensive matching result of the power material; according to the comprehensive matching result, obtain the comprehensive processing label of the current power material at different priorities.

[0123] The location verification module is used to obtain the current distribution location of power materials and extract the tracking displacement value; verify whether the power materials are at the optimal location point and calculate the displacement evaluation coefficient; and output the distribution feedback result.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for detecting and identifying an image of electric power materials, characterized in that: include: S1, acquiring images of electric power materials when materials are moved, sorting the images of electric power materials according to the acquired processes and positions, and obtaining a set of material images related to the process positions; S2, performing image detection on the material image set, obtaining a feature set corresponding to the material image set, performing linear regression analysis on the feature set, clustering the feature set according to the linear regression analysis result, obtaining a feature clustering result, and setting an electric power material label for the current electric power material based on the feature clustering result; S3, identifying the relative position of the feature vector corresponding to the feature clustering result on the power material image, and obtaining the feature coverage and overlapping area ratio of the feature clustering result on each surface of the adjacent power material image, matching the feature coverage and overlapping area ratio with the power material label, and obtaining the comprehensive processing label of the current power material at different priorities; S4, based on the comprehensive processing tags, tracks the power material image, determines the current distribution position and distribution probability of the power material, and obtains the distribution feedback result of the power material.

2. The method for detecting and identifying an electric power material image according to claim 1, characterized in that: Step S2 includes: S21, obtaining appearance recognition features for the current appearance of electric power materials, demand features for the transportation of electric power materials, and batch processing features in the feature set; and constructing a linear regression model related to the appearance recognition features, demand features, and batch processing features; S22, analyzing the appearance recognition feature, the demand feature, and the batch processing feature in sequence, and obtaining a first correlation coefficient corresponding to the appearance recognition feature, a second correlation coefficient corresponding to the demand feature, and a third correlation coefficient corresponding to the batch processing feature in sequence; S23, performing linear regression on the first correlation coefficient, the second correlation coefficient, and the third correlation coefficient to obtain regression coefficients corresponding to the first correlation coefficient, the second correlation coefficient, and the third correlation coefficient; S24, according to the range of values ​​of the regression coefficient, multiple clusters are set, the initial cluster center of each cluster is represented by the weighted average of the first correlation coefficient, the second correlation coefficient and the third correlation coefficient, the distance between each element in the feature set and the initial cluster center is calculated, the cluster center is continuously iterated, and the iterated cluster cluster is output as the feature clustering result.

3. The method for detecting and identifying an electric power material image according to claim 1, characterized in that: The implementation of step S3 includes: S31, obtaining the feature clustering results of the current electric power materials, and mapping the feature clustering results to the six planes corresponding to the electric power materials, and determining the relative position of each feature vector in the feature clustering results in the electric power materials; S32, calculating the feature coverage of the feature vector on each surface, where the feature coverage represents the ratio of the pixel ratio occupied by the current feature vector to the sum of the pixel ratios occupied by all feature vectors on the surface; S33, calculating the overlapping area ratio between the feature vector and the adjacent power material image at the corresponding feature coverage, where the overlapping area ratio represents the intersection-and-union ratio of contours between multiple faces in the power material image; S34, obtaining the attitude angle of the current electric power material, comparing the feature coverage and overlapping area ratio of the feature vector with the attitude angle of the current electric power material, and determining the matching result between the feature vector and the attitude angle; S35, combining the matching results on the six surfaces corresponding to the electric power material according to the matching results of the feature vector and the attitude angle to generate a comprehensive matching result of the electric power material; and obtaining a comprehensive processing label according to the comprehensive matching result of the electric power material.

4. A method for detecting and identifying an electric power material image as claimed in claim 3, characterized in that: Feature coverage can be expressed as: Among them, C f represents the feature coverage on the fth surface, Ft f Indicates the pixel ratio of the feature vector on the fth face, Ft f,total Represents the sum of the pixel ratios occupied by all eigenvectors on the fth face.

5. A method for detecting and identifying an electric power material image as claimed in claim 4, characterized in that: The matching result in step S34 is expressed as: The matching difference values ​​of the feature coverage and the overlapping area ratio of the feature vector at the corresponding attitude angle are calculated, and the matching score of the current feature vector is calculated according to the obtained matching difference values, and the matching score is output as the matching result of the feature vector and the attitude angle; The matching difference value of feature coverage is expressed as: ΔC f =|C f _C f,θ |; Where, ΔC f represents the matching difference value of the feature coverage on the fth surface, θ represents the posture angle, C f,θ It represents the feature coverage when there is an attitude angle θ on the f-th surface; The matching difference value of the overlap area ratio is expressed as: I ff′ =|IoU ff′ _IoU ff′,θ |; Among them, ΔI ff′ Represents the matching difference value of the overlapping area ratio of the fth face and the f′th face, IoU ff′ Represents the overlapping area ratio of the fth face and the f′th face, IoU ff′,θ It represents the overlapping area ratio between the f-th surface and the f′-th surface having the attitude angle θ.

6. A method for detecting and identifying an electric power material image as claimed in claim 5, characterized in that: The matching score is expressed as: S f represents the matching score of the fth face, F represents the number of faces in the power material image, e represents the exponential constant, and α represents the adjustment coefficient; W f represents the weight of the f-th face, W f′ represents the weight of the f′th face; The comprehensive matching result of the power materials is expressed as: the matching result of the feature vector and the attitude angle is comprehensively calculated to obtain a comprehensive matching score, and the comprehensive matching score is output as the comprehensive matching result; Among them, S total Represents the comprehensive matching score.

7. A method for detecting and identifying an electric power material image as claimed in claim 3, characterized in that: Step S35 also includes: S351, according to the comprehensive matching result of the electric power material, matching the comprehensive matching result with the electric power material tag to obtain at least one matching point; S352, performing allocation processing on the matching points, identifying first matching thresholds corresponding to the corresponding matching points in sequence; and selecting a plurality of target points from the matching points according to the first matching thresholds; S353, performing difference analysis on the target point, and taking the power material label corresponding to the point with the largest difference value after the difference analysis as the output comprehensive processing label.

8. The method for detecting and identifying an electric power material image according to claim 1, characterized in that: Step S4 includes: S41, obtaining the current distribution location of the electric power materials and extracting the tracking displacement value related to the distribution location; S42, combining the tracking displacement value with the distribution probability of the electric power material to verify whether the electric power material is at the optimal position point; and calculating the displacement evaluation coefficient of the electric power material at the optimal position point, and outputting the data corresponding to the displacement evaluation coefficient as the distribution feedback result.

9. A method for detecting and identifying an electric power material image as claimed in claim 8, characterized in that: The method of verifying whether the power material is at the optimal position is to obtain the displacement deviation of the current power material, combine the displacement deviation with the allocation probability, and calculate the displacement evaluation coefficient; Among them, δ represents the displacement evaluation coefficient, D represents the tracking displacement value, and ΔDI represents the displacement deviation. represents the standard value of displacement deviation, P(DI) represents the distribution probability of power materials, and α1, α2 and α3 represent weight coefficients respectively.

10. A detection and recognition system for electric power material images, using the detection and recognition method for electric power material images as claimed in any one of claims 1 to 9, characterized in that: include: An image acquisition module is used to acquire images of power materials during movement, and to sort the images according to the acquired process and position to form a material image set related to the process position; Image detection module, used to perform image detection on material image sets and extract feature sets; Construct a linear regression model and perform linear regression analysis on the feature set; cluster the feature set according to the linear regression analysis results to obtain feature clustering results; Based on the feature clustering results, the power material label of the current power material is set; A feature processing module, used to identify the relative position of the feature vector corresponding to the feature clustering result on the power material image; Calculate the feature coverage and overlapping area ratio, and match them with the power material labels; combine the posture angles of the power materials to generate comprehensive matching results of the power materials; based on the comprehensive matching results, obtain the comprehensive processing labels of the current power materials at different priorities; The location verification module is used to obtain the current distribution location of power materials and extract the tracking displacement value; verify whether the power materials are at the optimal location point and calculate the displacement evaluation coefficient; Output distribution feedback results.

Images