Electric power material identification method and system based on visual technology

Abstract

The invention discloses an electric power material identification method and system based on a visual technology, and belongs to the technical field of material supervision, and the method comprises the steps: S1, extracting first key feature information from sensing information based on a preset attribute item, inputting the first key feature information into an electric power material identification model, and carrying out the first analysis, if the first output result meets a preset requirement, obtaining an identification result of the target electric power material based on the first output result, and if the first output result does not meet the preset requirement, executing S2; s2, processing the collected electric power material image to obtain a target electric power material image, and extracting second key feature information from the target electric power material image based on electric power material appearance features; and S3, inputting the second key feature information into the electric power material identification model for secondary analysis, and obtaining an identification result of the target electric power material based on the second output result and the first output result. The problem that the recognition precision is low due to the fact that multi-dimensional recognition cannot be carried out on the electric power materials is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of material supervision, and in particular to a method and system for identifying electric power materials based on vision technology. Background Art

[0002] In the power industry, material management plays a vital role and is directly related to the operational efficiency and safety of power companies. Traditional methods for identifying power materials mainly rely on manual detection or automated recognition systems based on traditional image processing technology. However, these methods face many challenges in practical applications. Manual detection is not only inefficient, but also easily affected by human factors, making it difficult to ensure recognition accuracy. Although traditional image processing technology has achieved a certain degree of automation, it is limited by its simple algorithm and insufficient feature extraction capabilities, making it difficult to maintain high recognition accuracy in complex and changing actual scenarios. In the current market, technical solutions have emerged in the field of power material identification. Barcode recognition technology realizes rapid identification of materials through barcode scanning, such as a storage material scanning and identification method and system with patent number CN116451720A, which includes: obtaining barcode information to be identified, wherein the barcode information includes two-dimensional code information and barcode information; determining the image content deflection angle of the image content in each of the barcode information, and adjusting the angle of the barcode information; identifying the type of the barcode information with adjusted angle, and determining the category of the barcode information; determining the identification mode corresponding to the barcode information according to the category; and outputting the item information corresponding to the barcode information based on the barcode information and the identification mode. The above scheme realizes rapid and accurate identification of barcode information with different inclination angles, but it relies on barcode identification. When the barcode is damaged or lost, it is impossible to identify power materials from other dimensions. Therefore, there is a problem that power materials cannot be identified in multiple dimensions, resulting in low identification accuracy. Summary of the invention

[0003] In view of the problem that the existing technology cannot identify power materials in multiple dimensions, resulting in low identification accuracy, the present invention provides a method and system for identifying power materials based on visual technology. When the information obtained in a certain dimension is input into the power material identification model for a primary analysis and no power material identification result can be obtained, the information in other dimensions is input into the power material identification model for a secondary analysis, and the identification result of the power material is obtained based on the primary analysis result and the secondary analysis result, which solves the problem of low identification accuracy caused by the inability to identify power materials in multiple dimensions. The accuracy of power material identification is improved, and the robustness of power material identification is enhanced.

[0004] In order to solve the above technical problems, the present invention provides a method for identifying electric power materials based on visual technology, comprising: S1: extracting first key feature information from the sensor information based on preset attribute items, inputting the first key feature information into the power material identification model for analysis, and obtaining the identification result of the target power material based on the first output result if the first output result of the power material identification model meets the preset requirements, and executing S2 if the preset requirements are not met; S2: Processing the collected electric power material image to obtain a target electric power material image, and extracting second key feature information from the target electric power material image based on the electric power material appearance feature; S3: inputting the second key characteristic information into the electric power material identification model for secondary analysis, thereby outputting a second output result, and obtaining an identification result of the target electric power material based on the second output result and the first output result.

[0005] After adopting the above technical solution, the present invention has the following advantages: The power materials are identified by obtaining information from power material labels and identifying the appearance characteristics of power materials in two major dimensions, and the appearance characteristics of power materials are obtained from three minor dimensions: shape, color, and texture. When the information obtained from a certain dimension is input into the power material identification model for a primary analysis and no power material identification result can be obtained, the information from other dimensions is input into the power material identification model for a secondary analysis. The identification result of the power materials is obtained based on the primary and secondary analysis results, thereby improving the accuracy and robustness of power material identification. By processing the acquired data and extracting key feature data from the processed data, the complexity and amount of calculation of data processing are reduced, thereby improving the recognition efficiency; The problem of low recognition accuracy caused by the inability to identify power materials in multiple dimensions has been solved.

[0006] Preferably, in S1, the first key feature information is input into the electric power material identification model for analysis. If the first output result of the electric power material identification model meets the preset requirements, the identification result of the target electric power material is obtained based on the first output result. If the preset requirements are not met, S2 is executed, including: Construct an electric power material identification model, and train the constructed electric power material identification model based on historical electric power material images; The first key feature information is input into the trained electric power material identification model for analysis. If the trained electric power material identification model outputs only one first output result, the first output result is the identification result of the target electric power material. If the trained electric power material identification model does not output the first output result or outputs more than one first output result, S2 is executed.

[0007] The first key feature information is information obtained from the label of the target power material. The identification result of the target power material can be obtained through the first key feature information. When the first key feature information is complete and there is no contradiction between the first key feature information, the model will only output one identification result, which is the identification result of the target power material. When the first key feature information is partially missing or there is a contradiction between the first key feature information, the model will output multiple identification results or no identification result. Therefore, S2 is continued to be executed when multiple identification results are output or no identification result is output. By distinguishing the target power material with incomplete information and the target power material with contradictory information, the identification time of the target power material with complete information and the target power material with no contradiction is shortened, the identification efficiency is improved, and at the same time, the target power material with incomplete information or the target power material with contradictory information is re-identified, thereby improving the identification accuracy.

[0008] Preferably, the step of constructing the electric power material identification model and training the constructed electric power material identification model based on historical electric power material images comprises: Construct a power material identification model; Performing image enhancement processing on the historical electric power material image to obtain a second enhanced image, inputting the second enhanced image into the constructed electric power material recognition model for recognition, and if the recognition result is correct, recording information corresponding to the electric power material appearance features in the second enhanced image; The recorded information is bound to the type of electric power material corresponding to the second enhanced image, and an electric power material feature library is established based on the binding information.

[0009] When the recognition result of the electric material is accurate, it indicates that the information corresponding to the appearance characteristics of the electric material can determine the type of the electric material. Therefore, the recorded information is bound to the type of the electric material corresponding to the second enhanced image. In order to ensure the richness of the characteristics of the electric material in the electric material recognition model, an electric material feature library is established. When the electric material is subsequently identified, the features of the image to be identified can be quickly matched with the features in the electric material feature library, thereby improving the recognition speed and the generalization ability and flexibility of the model.

[0010] Preferably, in S2, the step of processing the collected electric power material image to obtain the target electric power material image includes: Grayscale processing is performed on the collected power material image to obtain a grayscale image; The grayscale image is subjected to image enhancement processing to obtain a first enhanced image, and the first enhanced image is segmented based on an edge detection algorithm to obtain a target power material image. The influence of the background image in the target power material identification process is eliminated, thereby improving the identification accuracy and efficiency.

[0011] Preferably, S3 includes: S31: digitizing the second key feature information, and combining the digitized second key feature data into a feature vector based on the appearance features of the power material; S32: Compare the feature vector with the power material feature library, and then output the comparison result, and obtain the recognition result of the target power material based on the comparison result and the first output result. By digitizing the second key feature data, the recognition error caused by the difference in information expression is reduced, and the digitized second key feature data is combined into a feature vector, which can more comprehensively reflect the characteristics of the target power material. The feature vector is compared with the power material feature library to obtain a secondary recognition result, and the final power material recognition result is obtained by combining the result of the first recognition, thereby improving the accuracy of recognition.

[0012] Preferably, in S32, the identification result of the target electric power material obtained based on the comparison result and the first output result includes: If only one comparison result is output, then the comparison result is the identification result of the target power material; If there is more than one comparison result output and more than one first output result, the comparison result is matched with the first output result. If only one comparison result is successfully matched, the successfully matched comparison result is the identification result of the target power material. If it is not successfully matched or more than one comparison result is successfully matched, an error is reported. If no comparison result is output, an error is reported. By analyzing the results of the first identification and the results of the second identification to obtain the identification result of the target power material, the accuracy and stability of the identification are improved, thereby improving the identification performance.

[0013] Preferably, the combining of the numerical second key feature data into a feature vector based on the appearance features of the electric power materials comprises: The appearance characteristics of the power materials include at least color, texture, and shape; The feature vector is: , , , They are the color vector, texture vector and shape vector corresponding to the target power material. The target power material can be identified from multiple small dimensions, which reduces manual intervention and human errors and improves the accuracy and robustness of power material identification.

[0014] Beneficial effects of this program: The power materials are identified by obtaining information from power material labels and identifying the appearance characteristics of power materials in two major dimensions, and the appearance characteristics of power materials are obtained from three minor dimensions: shape, color, and texture. When the information obtained from a certain dimension is input into the power material identification model for a primary analysis and no power material identification result can be obtained, the information from other dimensions is input into the power material identification model for a secondary analysis. The identification result of the power materials is obtained based on the primary and secondary analysis results, thereby improving the accuracy and robustness of power material identification. By processing the acquired data and extracting key feature data from the processed data, the complexity and amount of calculation of data processing are reduced, thereby improving the recognition efficiency; By establishing a power material feature library, when subsequently identifying power materials, the features of the image to be identified can be quickly matched with the features in the power material feature library, thereby improving the generalization ability and flexibility of the model; The problem of low recognition accuracy caused by the inability to identify power materials in multiple dimensions has been solved.

[0015] The present invention also provides a power material identification system based on visual technology, including a data acquisition module, a feature extraction module, and a power material identification module; The data acquisition module is used to acquire sensor information and collect images of power materials when the first output result of the power material identification model does not meet the preset requirements; The feature extraction module is used to extract first key feature information from the sensor information, and when the first output result of the power material identification model does not meet the preset requirements, extract second key feature information from the target power material image; The electric power material identification module is used to perform a primary analysis on the extracted first key feature information, and when the first output result of the electric power material identification model does not meet the preset requirements, perform a secondary analysis on the extracted second key feature information.

[0016] Preferably, it also includes a data processing module, and the data processing module is used to process the collected power material image to obtain the target power material image.

[0017] Preferably, the data acquisition module at least includes a high-resolution camera unit and a sensor unit, wherein the high-resolution camera unit is used to collect images of electric power materials, and the sensor unit is used to acquire sensor information.

[0018] Beneficial effects of this program: The data processing module eliminates the influence of the background image in the process of power material identification. The feature extraction module extracts the first key feature data and the second key feature data from the sensor information and the target power material image respectively. The combination of the data processing module and the feature extraction module reduces the complexity and calculation amount of data processing, thereby improving the recognition efficiency. The power material identification module identifies the target power materials from two large dimensions and three small dimensions, and then comprehensively analyzes the primary identification results and the secondary identification results to obtain the final identification results of the target power materials, thereby improving the accuracy and robustness of power material identification; The problem of low recognition accuracy caused by the inability to identify power materials in multiple dimensions has been solved.

[0019] After adopting the above technical solution, the present invention has the following advantages: The power materials are identified by obtaining information from power material labels and identifying the appearance characteristics of power materials in two major dimensions, and the appearance characteristics of power materials are obtained from three minor dimensions: shape, color, and texture. When the information obtained from a certain dimension is input into the power material identification model for a primary analysis and no power material identification result can be obtained, the information from other dimensions is input into the power material identification model for a secondary analysis. The identification result of the power materials is obtained based on the primary and secondary analysis results, thereby improving the accuracy and robustness of power material identification. By combining the data processing module with the feature extraction module, the complexity and computational complexity of data processing are reduced, thereby improving recognition efficiency; By establishing a power material feature library, when subsequently identifying power materials, the features of the image to be identified can be quickly matched with the features in the power material feature library, thereby improving the generalization ability and flexibility of the model; The problem of low recognition accuracy caused by the inability to identify power materials in multiple dimensions has been solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0021] Figure 1 The present invention is a method flow chart of a method for identifying electric power materials based on visual technology. DETAILED DESCRIPTION

[0022] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0024] Embodiment 1: like Figure 1 As shown, a method for identifying electric power materials based on visual technology includes the following steps: S1: extracting first key feature information from the sensor information based on preset attribute items, inputting the first key feature information into the electric power material identification model for analysis, and if the first output result of the electric power material identification model meets the preset requirements, obtaining the identification result of the target electric power material based on the first output result; if it does not meet the preset requirements, executing S2.

[0025] In S1, the first key feature information is input into the electric power material identification model for analysis. If the first output result of the electric power material identification model meets the preset requirements, the identification result of the target electric power material is obtained based on the first output result. If it does not meet the preset requirements, S2 is executed, including: Construct an electric power material identification model, and train the constructed electric power material identification model based on historical electric power material images; The first key feature information is input into the trained electric power material identification model for analysis. If the trained electric power material identification model outputs only one first output result, the first output result is the identification result of the target electric power material. If the trained electric power material identification model does not output the first output result or outputs more than one first output result, S2 is executed.

[0026] The step of constructing the electric power material identification model and training the constructed electric power material identification model based on the historical electric power material images includes: Construct a power material identification model; Performing image enhancement processing on the historical electric power material image to obtain a second enhanced image, inputting the second enhanced image into the constructed electric power material recognition model for recognition, and if the recognition result is correct, recording information corresponding to the electric power material appearance features in the second enhanced image; The recorded information is bound to the type of electric power material corresponding to the second enhanced image, and an electric power material feature library is established based on the binding information.

[0027] The preset attribute items include material name, model, manufacturer, etc. If the sensor information obtained by the sensor includes the material name, the sensor information is compared with the preset attribute items. If the material name is transformer, the first key feature information is transformer. The transformer is input into the power material identification model, and the transformer can be directly output as the first output result. At this time, the identification result of the target power material is the transformer. In order to ensure the accuracy of the identification result, if the sensor information obtained by the sensor includes the material name, manufacturer, and weight, the sensor information is compared with the preset attribute items. If the material name is transformer and the manufacturer is the first manufacturer, the first key feature information is transformer and the first manufacturer. The transformer and the first manufacturer are input into the power material identification model to obtain the type of power materials produced by the first manufacturer. If the type of power materials produced includes transformers, the first output result of the model is transformer, and then the identification result of the target power material is transformer. If the type of power materials produced does not include transformers, the model does not output the first output result, and continues to execute S2. When multiple data corresponding to the attribute items are obtained, the multiple data are input into the model for verification one by one, and then the final recognition result is obtained. If the final recognition result is not obtained, a secondary recognition is performed, which improves the accuracy and robustness of the identification of power materials.

[0028] The historical power material image is processed by histogram equalization and other methods to obtain the second enhanced image. If the power material in the second enhanced image is a transformer, the output result when the second enhanced image is input into the constructed power material recognition model for recognition is also a transformer, indicating that the recognition result is correct. The shape, color, and texture features of the transformer are recorded, and the shape, color, and texture features are bound to the transformer type to obtain a power material feature library, which includes the shape, color, and texture features of the power material. When the power material is subsequently recognized, the features of the image to be recognized can be quickly matched with the features in the power material feature library, which improves the recognition speed and the generalization ability and flexibility of the model.

[0029] S2: Processing the collected electric power material image to obtain a target electric power material image, and extracting second key feature information from the target electric power material image based on the electric power material appearance feature.

[0030] In S2, the processing of the collected power material image to obtain the target power material image includes: Grayscale processing is performed on the collected power material image to obtain a grayscale image; The grayscale image is subjected to image enhancement processing to obtain a first enhanced image, and the first enhanced image is segmented based on an edge detection algorithm to obtain a target power material image.

[0031] The brightness values ​​of the three RGB components of each pixel in the target power material image are averaged to obtain a grayscale image. The grayscale image is processed by methods such as histogram equalization to obtain a first enhanced image. The image is segmented using the Canny algorithm, which is a classic edge detection algorithm. The Canny algorithm searches for the maximum value of the gradient in the image to detect the edge, and improves the accuracy and stability of the detection results through maximum value suppression and double threshold processing. The second key feature information is extracted from the target power material image obtained through processing. The second key feature information is the shape, color and texture information of the target power material, which eliminates the influence of the background image in the target power material recognition process, thereby improving the recognition accuracy and efficiency.

[0032] S3: inputting the second key characteristic information into the electric power material identification model for secondary analysis, thereby outputting a second output result, and obtaining an identification result of the target electric power material based on the second output result and the first output result.

[0033] The S3 includes: S31: digitizing the second key feature information, and combining the digitized second key feature data into a feature vector based on the appearance features of the power material; S32: Compare the feature vector with the electric power material feature library, and then output the comparison result, and obtain the recognition result of the target electric power material based on the comparison result and the first output result.

[0034] In S32, the identification result of the target electric power material obtained based on the comparison result and the first output result includes: If only one comparison result is output, then the comparison result is the identification result of the target power material; If there is more than one comparison result output and more than one first output result, the comparison result will be matched with the first output result. If only one comparison result is successfully matched, the successfully matched comparison result is the identification result of the target power material. If the match is not successful or more than one comparison result is successfully matched, an error is reported. If no comparison result is output, an error is reported.

[0035] The combining of the numerical second key feature data into a feature vector based on the appearance features of the electric power materials comprises: The appearance characteristics of the power materials include at least color, texture, and shape; The feature vector is: , , , They are the color vector, texture vector and shape vector corresponding to the target power material respectively.

[0036] The second key feature data is the shape, color and texture data of the target power material image. The appearance features of power materials include shape, color and texture. The numerical second key feature data is combined into a feature vector , compare the feature vector with the power material feature library. After the comparison, if the comparison result output by the power material identification model is only the transformer, the identification result of the target power material is the transformer. If the comparison result output by the power material identification model is the transformer and the generator, and the first output result output by the power material identification model is the transformer and the turbine, the identification result of the target power material is the transformer. If the comparison result output by the power material identification model is the transformer and the generator, and the first output result output by the power material identification model is the transformer and the generator, the identification result of the target power material cannot be obtained, and an error is reported. It can more comprehensively reflect the characteristics of power materials, perform secondary identification of power materials from multiple small dimensions, and then combine the first identification results, reduce manual intervention and human errors, and improve the accuracy and robustness of power material identification.

[0037] Embodiment 2: An electric power material identification system based on visual technology, comprising a data acquisition module, a feature extraction module, and an electric power material identification module; The data acquisition module is used to acquire sensor information and collect images of power materials when the first output result of the power material identification model does not meet the preset requirements; The feature extraction module is used to extract first key feature information from the sensor information, and when the first output result of the power material identification model does not meet the preset requirements, extract second key feature information from the target power material image; The electric power material identification module is used to perform a primary analysis on the extracted first key feature information, and when the first output result of the electric power material identification model does not meet the preset requirements, perform a secondary analysis on the extracted second key feature information.

[0038] It also includes a data processing module, which is used to process the collected power material image to obtain a target power material image.

[0039] The data acquisition module at least includes a high-resolution camera unit and a sensor unit. The high-resolution camera unit is used to collect images of power materials, and the sensor unit is used to obtain sensor information.

[0040] The image information of power materials is collected in real time through the high-resolution camera unit, and the sensor information is obtained through the sensor unit. The data processing module uses image processing algorithms such as Gaussian filtering and histogram equalization to optimize the image. The feature extraction module uses deep learning technologies such as convolutional neural network (CNN) to automatically learn and extract features such as edges, textures, and shapes in the image. The power material identification module builds and trains the power material identification model. Through the learning of a large amount of labeled data, the power material identification model has the ability to accurately identify power materials. At the same time, the extracted first key feature information and second key feature information are analyzed, and then the identification results of the target power materials are output. The system also includes an adaptive optimization module and a real-time feedback and visualization display module. The adaptive optimization module uses online learning, transfer learning and other technologies to enable the power material identification model to continuously adapt to new data distribution and feature changes, improve recognition accuracy and robustness, and develop a user-friendly interactive interface. Combined with data visualization technology, it provides intuitive and easy-to-understand recognition result display and data analysis functions. Identifying power materials from multiple dimensions reduces manual intervention and human errors, and improves the accuracy and robustness of power material identification.

[0041] In this embodiment, high-definition cameras and infrared sensors are deployed in the power material storage area, and system software and power material recognition models are installed. After the data acquisition module obtains the image information of the power material in real time, it transmits it to the data processing module. The data processing module uses advanced image processing technology to perform denoising and enhancement on the image. The feature extraction module uses the CNN model in deep learning to extract the first key feature information and the second key feature information from the sensor information and the preprocessed image respectively. The power material recognition model recognizes and classifies the power material based on the extracted key feature information, and transmits the recognition result to the real-time feedback and visualization display module. The real-time feedback and visualization display feed back the recognition result to the management personnel in real time, and displays the location, status and recognition time of the material through an intuitive graphical interface. At the same time, it provides data analysis and statistical functions to help the management personnel fully grasp the dynamic changes of the storage materials. In addition, the adaptive optimization module has a strong adaptive learning ability, which can automatically adjust and optimize the power material recognition model according to the changes in the actual application scenario. In addition, by regularly maintaining and upgrading the system, introducing the latest deep learning algorithms and technologies, the recognition accuracy and robustness of the system are continuously improved to ensure continuous optimization of user experience. The system is closely integrated with the warehouse management system and supply chain system to realize the automatic entry, update and full-process management of material information. This not only reduces labor costs and error rates, but also significantly improves the accuracy and efficiency of warehouse operations. Through continuous integration and optimization, the intelligent level of the entire power material management is further promoted.

[0042] The specific implementation described above is a preferred implementation of a method and system for identifying electric materials based on visual technology of the present invention, and is not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A method for identifying electric power materials based on visual technology, characterized in that: The following steps are involved: S1: extracting first key feature information from the sensor information based on preset attribute items, inputting the first key feature information into the power material identification model for analysis, and obtaining the identification result of the target power material based on the first output result if the first output result of the power material identification model meets the preset requirements, and executing S2 if the preset requirements are not met; S2: Processing the collected electric power material image to obtain a target electric power material image, and extracting second key feature information from the target electric power material image based on the electric power material appearance feature; S3: inputting the second key characteristic information into the electric power material identification model for secondary analysis, thereby outputting a second output result, and obtaining an identification result of the target electric power material based on the second output result and the first output result.

2. According to the method of electric power material identification based on visual technology in claim 1, it is characterized in that: In S1, the first key feature information is input into the electric power material identification model for analysis. If the first output result of the electric power material identification model meets the preset requirements, the identification result of the target electric power material is obtained based on the first output result. If it does not meet the preset requirements, S2 is executed, including: Construct an electric power material identification model, and train the constructed electric power material identification model based on historical electric power material images; The first key feature information is input into the trained electric power material identification model for analysis. If the trained electric power material identification model outputs only one first output result, the first output result is the identification result of the target electric power material. If the trained electric power material identification model does not output the first output result or outputs more than one first output result, S2 is executed.

3. The method for identifying electric power materials based on visual technology according to claim 2 is characterized in that: The step of constructing the electric power material identification model and training the constructed electric power material identification model based on the historical electric power material images includes: Construct a power material identification model; Performing image enhancement processing on the historical electric power material image to obtain a second enhanced image, inputting the second enhanced image into the constructed electric power material recognition model for recognition, and if the recognition result is correct, recording information corresponding to the electric power material appearance features in the second enhanced image; The recorded information is bound to the type of electric power material corresponding to the second enhanced image, and an electric power material feature library is established based on the binding information.

4. The method for identifying electric power materials based on visual technology according to claim 1 is characterized in that: In S2, the processing of the collected power material image to obtain the target power material image includes: Grayscale processing is performed on the collected power material image to obtain a grayscale image; The grayscale image is subjected to image enhancement processing to obtain a first enhanced image, and the first enhanced image is segmented based on an edge detection algorithm to obtain a target power material image.

5. The method for identifying electric power materials based on visual technology according to claim 3 is characterized in that: The S3 includes: S31: digitizing the second key feature information, and combining the digitized second key feature data into a feature vector based on the appearance features of the power material; S32: Compare the feature vector with the electric power material feature library, and then output the comparison result, and obtain the recognition result of the target electric power material based on the comparison result and the first output result.

6. The method for identifying electric power materials based on visual technology according to claim 5 is characterized in that: In S32, the identification result of the target electric power material obtained based on the comparison result and the first output result includes: If only one comparison result is output, then the comparison result is the identification result of the target power material; If there is more than one comparison result output and more than one first output result, the comparison result will be matched with the first output result. If only one comparison result is successfully matched, the successfully matched comparison result is the identification result of the target power material. If the match is not successful or more than one comparison result is successfully matched, an error is reported. If no comparison result is output, an error is reported.

7. The method for identifying electric power materials based on visual technology according to claim 5 is characterized in that: The combining of the numerical second key feature data into a feature vector based on the appearance features of the electric power materials comprises: The appearance characteristics of the power materials include at least color, texture, and shape; The feature vector is: , , , They are the color vector, texture vector and shape vector corresponding to the target power material respectively.

8. A power material identification system based on visual technology, applicable to a power material identification method based on visual technology as claimed in any one of claims 1 to 7, characterized in that: It includes data acquisition module, feature extraction module and power material identification module; The data acquisition module is used to acquire sensor information and collect images of power materials when the first output result of the power material identification model does not meet the preset requirements; The feature extraction module is used to extract first key feature information from the sensor information, and when the first output result of the power material identification model does not meet the preset requirements, extract second key feature information from the target power material image; The electric power material identification module is used to perform a primary analysis on the extracted first key feature information, and when the first output result of the electric power material identification model does not meet the preset requirements, perform a secondary analysis on the extracted second key feature information.

9. The electric power material identification system based on visual technology according to claim 8 is characterized in that: It also includes a data processing module, which is used to process the collected power material image to obtain a target power material image.

10. The electric power material identification system based on visual technology according to claim 8, characterized in that: The data acquisition module at least includes a high-resolution camera unit and a sensor unit. The high-resolution camera unit is used to collect images of power materials, and the sensor unit is used to obtain sensor information.

Images