Disassembling method and system for cover plate of distribution transformer

By using optical and infrared image acquisition equipment to generate dismantling plans during the dismantling of distribution transformers and controlling the operation of robotic arms, the problems of low efficiency and poor safety of manual dismantling have been solved, achieving a more efficient and safer dismantling process.

CN119006452AActive Publication Date: 2024-11-22STATE GRID ZHEJIANG ELECTRIC POWER CO LTD +1
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411473089.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-22
Publication Date
2024-11-22
Estimated Expiration
2044-10-22

AI Technical Summary

Technical Problem

Manually operating a robotic arm to dismantle the cover plate of a distribution transformer is inefficient and makes it difficult to ensure that the parameters set during the dismantling process are reasonable, which can easily cause equipment damage and safety hazards.

Method used

Optical and infrared images of the distribution transformer are acquired using image acquisition equipment. A disassembly scheme is generated using a disassembly model, and the disassembly operation of the robotic arm is controlled, including training the model to generate and adjust the disassembly scheme.

Benefits of technology

This improves the efficiency and safety of disassembling distribution transformer cover plates and reduces the risk of equipment damage.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006452A_ABST
    Figure CN119006452A_ABST
Patent Text Reader

Abstract

The invention discloses a disassembling method and system for a cover plate of a distribution transformer, and the method comprises the steps: obtaining an optical cover plate image and an infrared cover plate image, which are obtained by shooting the cover plate of the distribution transformer, after a manipulator starts to carry out the disassembling operation of the cover plate of the distribution transformer; based on the optical cover plate image and the infrared cover plate image, a disassembling scheme is generated by using a disassembling model so as to send a control instruction to the manipulator; the training process of the disassembly model comprises the following steps: acquiring a sample optical image and a sample infrared image corresponding to a sample distribution transformer; encoding the sample optical image and the sample infrared image by using the to-be-trained model, and obtaining optical features and infrared features generated in the encoding process; generating a first disassembly scheme by using a large model based on the optical features, the infrared features and the information of the sample distribution transformer; and training the to-be-trained model based on the optical features, the infrared features and the first disassembly scheme to obtain the disassembly model, thereby improving the disassembly efficiency and disassembly safety of the distribution transformer cover plate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of automatic control technology, and in particular to a method and system for disassembling a distribution transformer cover. Background Art

[0002] The distribution transformer cover is one of the important components of the distribution transformer. It is mainly used to protect the electrical components inside the transformer from the influence of the external environment, and at the same time provide safety protection to prevent people from contacting the live parts. Before inspecting and repairing the internal components of the distribution transformer, it is usually necessary to disassemble the distribution transformer cover by manually operating a manipulator throughout the process.

[0003] However, manual operation of the robot disassembly process is not only inefficient, but also heavily dependent on the operator's experience. It is difficult to ensure the rationality of the robot's disassembly speed, angle, force and other parameter settings during the disassembly process, which can easily cause damage to the distribution transformer and threaten the safety of equipment and personnel. Summary of the invention

[0004] In order to solve the above technical problems, the embodiments of the present application propose a method and system for disassembling a distribution transformer cover, which can improve the disassembly efficiency and disassembly safety of the distribution transformer cover.

[0005] In a first aspect, an embodiment of the present application provides a method for disassembling a distribution transformer cover, comprising: After detecting that the manipulator starts to perform the cover disassembly operation on the distribution transformer, the cover of the distribution transformer is photographed by an image acquisition device to obtain an optical cover image and an infrared cover image; Based on the optical cover plate image and the infrared cover plate image, a disassembly model is used to generate a disassembly plan, and a control instruction is sent to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover plate disassembly operation of the manipulator; The training process of the disassembly model includes: Acquire a sample optical image and a sample infrared image corresponding to a sample distribution transformer; Using the model to be trained, respectively encoding the sample optical image and the sample infrared image, and respectively obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process; Based on the optical features, the infrared features and the information of the sample distribution transformer, a first disassembly plan is generated using a large model; Based on the optical features, the infrared features and the first disassembly scheme, the model to be trained is trained to obtain the disassembly model.

[0006] Optionally, the generating a first disassembly solution using a large model based on the optical feature, the infrared feature and the information of the sample distribution transformer includes: Performing inner product processing on the optical feature and the infrared feature to obtain an inner product feature; The inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model.

[0007] Optionally, the information of the sample distribution transformer includes a standard screw profile template diagram and a standard temperature distribution diagram of the sample distribution transformer, and the inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model, including: Inputting the inner product feature into the large model, so that the large model generates a simulated screw contour template diagram and a simulated temperature distribution diagram according to the inner product feature; The standard screw contour template image and the standard temperature distribution image are input into the large model, so that the large model generates and outputs the first disassembly plan based on the difference between the simulated screw contour template image and the standard screw contour template image, and the difference between the simulated temperature distribution image and the standard temperature distribution image.

[0008] Optionally, the training the model to be trained based on the optical feature, the infrared feature and the first disassembly scheme to obtain the disassembly model includes: Performing inner product processing on the optical feature and the infrared feature to obtain an inner product feature; Inputting the inner product feature into the model to be trained to generate a second disassembly solution; Extracting a first text feature of the first disassembly solution, and extracting a second text feature of the second disassembly solution; Determining a text feature difference between the first text feature and the second text feature using a preset loss function; Based on the text feature differences, the model to be trained is trained to obtain the disassembly model.

[0009] Optionally, the sample optical image is cropped into a plurality of optical partial images, and the sample infrared image is cropped into a plurality of infrared partial images, N optical partial images among the plurality of optical partial images correspond one-to-one to N infrared partial images among the plurality of infrared partial images, N is a positive integer, the corresponding optical partial images are the same as the partial areas indicated by the infrared partial images, the optical features at least include the optical partial features corresponding to the N optical partial images, and the infrared features at least include the infrared partial features corresponding to the N infrared partial images; The method of using the model to be trained to encode the sample optical image and the sample infrared image respectively, and obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process respectively, includes: Using the model to be trained, encoding the sample optical image, and obtaining the optical local features corresponding to each of the N optical local images generated during the encoding process, The sample infrared image is encoded using the model to be trained, and infrared local features corresponding to each of the N infrared local images generated during the encoding process are obtained.

[0010] Optionally, each optical partial image in the N optical partial images and its corresponding infrared partial image form a partial image pair, and performing inner product processing on the optical feature and the infrared feature to obtain an inner product feature includes: For each of the local image pairs, inner product processing is performed on the optical local features and the infrared local features corresponding to the local image pair to obtain the inner product local features corresponding to the local image pair; The inner product feature is determined based on the inner product local features corresponding to all the local image pairs.

[0011] Optionally, the training process further includes: Determining a weight corresponding to each of the partial image pairs according to the partial area indicated by each of the partial image pairs; The determining the inner product feature based on the inner product local features corresponding to all the local image pairs respectively comprises: Based on the weight corresponding to each of the local image pairs, weighted processing is performed on the inner product local features corresponding to each of the local image pairs; The weighted inner product local features corresponding to all the local image pairs are concatenated into the inner product feature.

[0012] Optionally, before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the method further includes: Obtaining information of the distribution transformer; The step of generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image includes: A disassembly solution is generated using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image.

[0013] Optionally, before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the method further includes: Preprocessing the optical cover plate image and the infrared cover plate image; The generating a disassembly solution using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image includes: The information of the distribution transformer, the preprocessed optical cover plate image and the preprocessed infrared cover plate image are input into the disassembly model to generate a disassembly plan.

[0014] In a second aspect, an embodiment of the present application provides a disassembly system for a distribution transformer cover, comprising: Manipulator; image acquisition equipment; and, A control device is communicatively connected to the manipulator and the image acquisition device, and the control device is configured as follows: After detecting that the manipulator starts to perform the cover disassembly operation on the distribution transformer, the cover of the distribution transformer is photographed by the image acquisition device to obtain an optical cover image and an infrared cover image; Based on the optical cover plate image and the infrared cover plate image, a disassembly model is used to generate a disassembly plan, and a control instruction is sent to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover plate disassembly operation of the manipulator; The training process of the disassembly model includes: Acquire a sample optical image and a sample infrared image corresponding to a sample distribution transformer; Using the model to be trained, respectively encoding the sample optical image and the sample infrared image, and respectively obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process; Based on the optical features, the infrared features and the information of the sample distribution transformer, a first disassembly plan is generated using a large model; Based on the optical features, the infrared features and the first disassembly scheme, the model to be trained is trained to obtain the disassembly model.

[0015] In summary, the embodiments of the present application have at least the following beneficial effects: According to an embodiment of the present application, after detecting that a manipulator starts to perform a cover disassembly operation on a distribution transformer, an optical cover image and an infrared cover image are obtained by photographing the cover of the distribution transformer through an image acquisition device; based on the optical cover image and the infrared cover image, a disassembly model is used to generate a disassembly plan, and a control instruction is sent to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover disassembly operation of the manipulator; wherein the training process of the disassembly model includes: obtaining a sample optical image and a sample infrared image corresponding to a sample distribution transformer; using a model to be trained, respectively encoding the sample optical image and the sample infrared image, and respectively obtaining an optical feature corresponding to the sample optical image and an infrared feature corresponding to the sample infrared image generated during the encoding process; based on the optical feature, the infrared feature and the information of the sample distribution transformer, a first disassembly plan is generated using a large model; based on the optical feature, the infrared feature and the first disassembly plan, the model to be trained is trained to obtain the disassembly model, thereby improving the disassembly efficiency and safety of the cover of the distribution transformer. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic flow chart of a method for disassembling a distribution transformer cover provided in an embodiment of the present application; Figure 2 It is a flowchart of the training process of the disassembly model provided in the embodiment of the present application; Figure 3 It is a structural schematic diagram of a disassembly system for a distribution transformer cover provided in an embodiment of the present application. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0018] In the description of the present application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, features defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise specified, the meaning of "multiple" is two or more. In the description of the present application, the term "including" and its variations are open inclusions, i.e., "including but not limited to". The term "based on" means "at least partially based on". The term "according to" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments".

[0019] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0020] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0021] First, see Figure 1 , shows a schematic flow chart of a method for disassembling a distribution transformer cover provided in an embodiment of the present application, the method comprising steps S101-S102, which are specifically as follows: S101, after detecting that the robot arm starts to perform a cover disassembly operation on the distribution transformer, the cover of the distribution transformer is photographed by an image acquisition device to obtain an optical cover image and an infrared cover image.

[0022] It should be noted that the optical cover image and the infrared cover image described in this embodiment can be taken at the same time, so as to characterize the situation of the cover of the distribution transformer at the same time point. Here, the same time can be the moment after a preset time has passed after the manipulator starts to perform the cover disassembly operation on the distribution transformer, but it should be understood that the shooting time of the optical cover image and the infrared cover image can also be different, as long as the difference between the shooting time of the optical cover image and the infrared cover image is within a reasonable error range.

[0023] It can be understood that the optical cover plate image in this embodiment can be used to indicate the appearance of the cover plate of the distribution transformer, so as to further characterize the disassembly speed, disassembly angle, etc. of the current cover plate disassembly operation of the manipulator.

[0024] It can be understood that the infrared cover plate image in this embodiment can be used to indicate the temperature distribution of the cover plate of the distribution transformer, thereby further characterizing whether the materials at various locations of the cover plate of the distribution transformer are overheated due to improper disassembly.

[0025] In one example, the image acquisition device may include an optical imaging device and an infrared imaging device, wherein the optical imaging device may include a camera of a general mobile device (such as a terminal device, a digital camera, etc.), an industrial camera dedicated to disassembling a distribution transformer cover, etc., and the infrared imaging device may include an infrared imager, an infrared sensor, etc. In addition, all or part of the devices in the image acquisition device in this embodiment may also be installed on a manipulator.

[0026] S102, generating a disassembly plan using a disassembly model based on the optical cover plate image and the infrared cover plate image, and sending a control instruction to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover plate disassembly operation of the manipulator.

[0027] In one example, the disassembly scheme is used to indicate at least one of the following control parameters of the manipulator: disassembly force, disassembly speed, and disassembly angle. In this embodiment, after receiving the control instruction, the manipulator can adjust the parameters of the manipulator accordingly according to the various control parameters indicated by the disassembly scheme carried in the control instruction, so that the manipulator can subsequently operate according to the disassembly scheme, thereby achieving real-time adjustment of the working parameters of the manipulator, facilitating adaptation to real-time disassembly conditions, and avoiding damage to the distribution transformer and its cover plate due to improper disassembly.

[0028] Among them, see Figure 2 , shows a flow chart of the training process of the disassembly model provided in an embodiment of the present application, the training process includes steps S201-S204, which are as follows: S201, obtaining a sample optical image and a sample infrared image corresponding to a sample distribution transformer.

[0029] It should be noted that the sample optical image and the sample infrared image in this embodiment are both taken when the cover of the sample distribution transformer is disassembled, and can be used to indicate the situation when the cover of the sample distribution transformer is disassembled.

[0030] S202, using the model to be trained, respectively encode the sample optical image and the sample infrared image, and respectively obtain the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process.

[0031] In one example, the model to be trained may include a universal encoding model, which can be used to encode optical images and infrared images. In this case, the corresponding optical features and infrared features can be used to characterize the conditions of the sample distribution transformers captured by the corresponding images.

[0032] S203, generating a first disassembly plan using a large model based on the optical features, the infrared features and the information of the sample distribution transformer; wherein the information of the sample distribution transformer may include detailed information such as the model, size, structure, and material of the transformer.

[0033] In one example, based on the optical features, the infrared features, and the information of the sample distribution transformer, generating a first disassembly solution using a large model may include: inputting the optical features, the infrared features, and the information of the sample distribution transformer into the large model, so that the large model generates an initial disassembly solution based on the information of the sample distribution transformer and the optical features, and adjusting the initial disassembly solution using the infrared features to generate a first disassembly solution. In this embodiment, the large model may be a complex deep learning model, which may be a convolutional neural network, a recurrent neural network, a Transformer, or other types of neural networks, which is pre-trained with a large amount of relevant data, so that it has the ability to generate an initial disassembly solution based on the information and optical features of the sample distribution transformer, and can adjust the generated solution using infrared features.

[0034] S204: Based on the optical features, the infrared features and the first disassembly scheme, train the model to be trained to obtain the disassembly model.

[0035] In one example, based on the optical feature, the infrared feature and the first disassembly scheme, training the model to be trained to obtain the disassembly model may include: inputting the optical feature and the infrared feature into the model to be trained so that the model to be trained generates a fourth disassembly scheme according to the optical feature and the infrared feature; extracting the first text feature of the first disassembly scheme, and extracting the fourth text feature of the fourth disassembly scheme; using a preset loss function, determining the text feature difference between the first text feature and the fourth text feature; based on the text feature difference, training the model to be trained to obtain the disassembly model. In this embodiment, the model to be trained may include a convolutional neural network, a recurrent neural network, a Transformer or other types of neural networks, which have the ability to output the desired content according to specific feature inputs, so that a corresponding disassembly scheme can be generated.

[0036] It should be noted that in each embodiment associated with model training in the present application, the corresponding model training can be performed through back propagation and gradient descent or other common model training methods, which will not be repeated later.

[0037] In an optional implementation, the generating a first disassembly solution using a large model based on the optical feature, the infrared feature and the information of the sample distribution transformer includes: Inner product processing is performed on the optical feature and the infrared feature to obtain an inner product feature.

[0038] The inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model.

[0039] Generally speaking, large models are usually deployed in the cloud, that is, information transmission is usually involved when calling the large model. In this embodiment, the inner product can be used to calculate the projection of a vector on another vector, so that the inner product feature can have both optical feature and infrared feature information. In this way, the amount of information that needs to be transmitted when interacting with the large model can be reduced.

[0040] In an optional implementation, the information of the sample distribution transformer includes a standard screw profile template diagram and a standard temperature distribution diagram of the sample distribution transformer, and the inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model, including: The inner product feature is input into the large model, so that the large model generates a simulated screw contour template diagram and a simulated temperature distribution diagram according to the inner product feature.

[0041] The standard screw contour template image and the standard temperature distribution image are input into the large model, so that the large model generates and outputs the first disassembly plan based on the difference between the simulated screw contour template image and the standard screw contour template image, and the difference between the simulated temperature distribution image and the standard temperature distribution image.

[0042] In one example, the difference between the simulated screw outline template image and the standard screw outline template image, and the difference between the simulated temperature distribution map and the standard temperature distribution map can be obtained by the following steps: extracting the image features of the simulated screw outline template image, the image features of the standard screw outline template image, the image features of the simulated temperature distribution map, and the image features of the standard temperature distribution map, respectively. Calculating the cosine similarity between the image features of the simulated screw outline template image and the image features of the standard screw outline template image to characterize the difference between the simulated screw outline template image and the standard screw outline template image; calculating the cosine similarity between the image features of the simulated temperature distribution map and the image features of the standard temperature distribution map to characterize the difference between the simulated temperature distribution map and the standard temperature distribution map.

[0043] In an optional implementation, the training of the model to be trained based on the optical feature, the infrared feature and the first disassembly scheme to obtain the disassembly model includes: Inner product processing is performed on the optical feature and the infrared feature to obtain an inner product feature.

[0044] The inner product feature is input into the model to be trained to generate a second disassembly solution.

[0045] A first text feature of the first disassembly solution is extracted, and a second text feature of the second disassembly solution is extracted.

[0046] The text feature difference between the first text feature and the second text feature is determined using a preset loss function.

[0047] Based on the text feature differences, the model to be trained is trained to obtain the disassembly model.

[0048] In this embodiment, it is equivalent to borrowing the powerful computing power of the large model to simulate an optimal first disassembly solution, so as to adjust the model parameters of the model to be trained accordingly according to the difference between the solution generated by the model to be trained and the first disassembly solution, so that the output of the model to be trained can be close to the output of the large model.

[0049] In one example, the loss function may include mean square error, root mean square error, mean absolute error, cross entropy loss, cosine similarity loss, L1 regularization, L2 regularization, etc.

[0050] In one example, after obtaining the inner product feature, training the model to be trained to obtain the disassembly model may further include: According to the information of the sample distribution transformer, a corresponding solution database is determined.

[0051] A third disassembly solution matching the inner product feature is selected from the solution database.

[0052] A third text feature of the third disassembly solution is extracted.

[0053] The text feature difference between the first text feature and the third text feature is determined using a preset loss function.

[0054] Based on the text feature difference between the first text feature and the third text feature, the model to be trained is trained to obtain the disassembly model.

[0055] In an optional embodiment, the sample optical image is cropped into multiple optical local images, and the sample infrared image is cropped into multiple infrared local images, N optical local images among the multiple optical local images correspond one-to-one to N infrared local images among the multiple infrared local images, N is a positive integer, the corresponding optical local images are the same as the local areas indicated by their respective infrared local images, the optical features include at least optical local features corresponding to each of the N optical local images, and the infrared features include at least infrared local features corresponding to each of the N infrared local images.

[0056] The method of using the model to be trained to encode the sample optical image and the sample infrared image respectively, and obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process respectively, includes: Using the model to be trained, encoding the sample optical image, and obtaining the optical local features corresponding to each of the N optical local images generated during the encoding process, The sample infrared image is encoded using the model to be trained, and infrared local features corresponding to each of the N infrared local images generated during the encoding process are obtained.

[0057] It can be understood that in this embodiment, the sample infrared image can be cropped into multiple infrared local images according to a certain cropping method. Correspondingly, the sample infrared image can also be cropped into multiple infrared local images according to the corresponding cropping method. At this time, at least N of the cropped infrared local images and the infrared local images can correspond to each other, that is, the indicated local areas are the same.

[0058] In an optional implementation, each optical partial image in the N optical partial images and its corresponding infrared partial image constitute a partial image pair, and the inner product processing is performed on the optical feature and the infrared feature to obtain the inner product feature, including: For each of the local image pairs, inner product processing is performed on the optical local features and the infrared local features corresponding to the local image pair to obtain the inner product local features corresponding to the local image pair.

[0059] The inner product feature is determined based on the inner product local features corresponding to all the local image pairs.

[0060] In one example, determining the inner product feature based on the inner product local features corresponding to all the partial image pairs may include: concatenating the inner product local features corresponding to all the partial image pairs into the inner product feature.

[0061] In an optional implementation, the training process further includes: A weight corresponding to each of the local image pairs is determined according to the local area indicated by each of the local image pairs.

[0062] The determining the inner product feature based on the inner product local features corresponding to all the local image pairs respectively comprises: Based on the weight corresponding to each of the local image pairs, weighted processing is performed on the inner product local features corresponding to each of the local image pairs.

[0063] The weighted inner product local features corresponding to all the local image pairs are concatenated into the inner product feature.

[0064] In an optional implementation, before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the method further includes: The information of the distribution transformer is obtained.

[0065] The step of generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image includes: A disassembly solution is generated using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image.

[0066] In an optional implementation, before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the method further includes: The optical cover plate image and the infrared cover plate image are preprocessed.

[0067] The generating a disassembly solution using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image includes: The information of the distribution transformer, the preprocessed optical cover plate image and the preprocessed infrared cover plate image are input into the disassembly model to generate a disassembly plan.

[0068] In an example, the preprocessing may include normalization processing, cropping processing, scaling processing, rotation processing, translation processing, denoising processing, etc., which may be selected as needed.

[0069] In the second aspect, accordingly, the embodiments of the present application further provide a system for disassembling a distribution transformer cover, which can implement all processes of the method for disassembling a distribution transformer cover provided in the above embodiments.

[0070] See also Figure 3 , shows a schematic structural diagram of a system for disassembling a distribution transformer cover provided in an embodiment of the present application, the system 300 for disassembling a distribution transformer cover comprises: Robot 301.

[0071] an image acquisition device 302; and The control device 303 is in communication with the manipulator 301 and the image acquisition device 302, and the control device is configured as follows: After it is detected that the manipulator 301 starts to perform the cover disassembly operation on the distribution transformer, the cover of the distribution transformer is photographed by the image acquisition device 302 to obtain an optical cover image and an infrared cover image.

[0072] Based on the optical cover plate image and the infrared cover plate image, a disassembly plan is generated using a disassembly model, and a control instruction is sent to the manipulator 301 according to the disassembly plan, wherein the control instruction is used to regulate the cover plate disassembly operation of the manipulator 301 .

[0073] The training process of the disassembly model includes: A sample optical image and a sample infrared image corresponding to the sample distribution transformer are obtained.

[0074] The sample optical image and the sample infrared image are encoded respectively using the model to be trained, and the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process are obtained respectively.

[0075] Based on the optical features, the infrared features and the information of the sample distribution transformer, a first disassembly plan is generated using a large model.

[0076] Based on the optical features, the infrared features and the first disassembly scheme, the model to be trained is trained to obtain the disassembly model.

[0077] In an optional implementation, the generating a first disassembly solution using a large model based on the optical feature, the infrared feature and the information of the sample distribution transformer includes: Inner product processing is performed on the optical feature and the infrared feature to obtain an inner product feature.

[0078] The inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model.

[0079] In an optional implementation, the information of the sample distribution transformer includes a standard screw profile template diagram and a standard temperature distribution diagram of the sample distribution transformer, and the inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model, including: The inner product feature is input into the large model, so that the large model generates a simulated screw contour template diagram and a simulated temperature distribution diagram according to the inner product feature.

[0080] The standard screw contour template image and the standard temperature distribution image are input into the large model, so that the large model generates and outputs the first disassembly plan based on the difference between the simulated screw contour template image and the standard screw contour template image, and the difference between the simulated temperature distribution image and the standard temperature distribution image.

[0081] In an optional implementation, the training of the model to be trained based on the optical feature, the infrared feature and the first disassembly scheme to obtain the disassembly model includes: Inner product processing is performed on the optical feature and the infrared feature to obtain an inner product feature.

[0082] The inner product feature is input into the model to be trained to generate a second disassembly solution.

[0083] A first text feature of the first disassembly solution is extracted, and a second text feature of the second disassembly solution is extracted.

[0084] The text feature difference between the first text feature and the second text feature is determined using a preset loss function.

[0085] Based on the text feature differences, the model to be trained is trained to obtain the disassembly model.

[0086] In an optional embodiment, the sample optical image is cropped into multiple optical local images, and the sample infrared image is cropped into multiple infrared local images, N optical local images among the multiple optical local images correspond one-to-one to N infrared local images among the multiple infrared local images, N is a positive integer, the corresponding optical local images are the same as the local areas indicated by their respective infrared local images, the optical features include at least optical local features corresponding to each of the N optical local images, and the infrared features include at least infrared local features corresponding to each of the N infrared local images.

[0087] The method of using the model to be trained to encode the sample optical image and the sample infrared image respectively, and obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process respectively, includes: Using the model to be trained, encoding the sample optical image, and obtaining the optical local features corresponding to each of the N optical local images generated during the encoding process, The sample infrared image is encoded using the model to be trained, and infrared local features corresponding to each of the N infrared local images generated during the encoding process are obtained.

[0088] In an optional implementation, each optical partial image in the N optical partial images and its corresponding infrared partial image constitute a partial image pair, and the inner product processing is performed on the optical feature and the infrared feature to obtain the inner product feature, including: For each of the local image pairs, inner product processing is performed on the optical local features and the infrared local features corresponding to the local image pair to obtain the inner product local features corresponding to the local image pair.

[0089] The inner product feature is determined based on the inner product local features corresponding to all the local image pairs.

[0090] In an optional implementation, the training process further includes: A weight corresponding to each of the local image pairs is determined according to the local area indicated by each of the local image pairs.

[0091] The determining the inner product feature based on the inner product local features corresponding to all the local image pairs respectively comprises: Based on the weight corresponding to each of the local image pairs, weighted processing is performed on the inner product local features corresponding to each of the local image pairs.

[0092] The weighted inner product local features corresponding to all the local image pairs are concatenated into the inner product feature.

[0093] In an optional implementation, before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the control device is further configured to: The information of the distribution transformer is obtained.

[0094] The step of generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image includes: A disassembly solution is generated using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image.

[0095] In an optional implementation, before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the control device is further configured to: The optical cover plate image and the infrared cover plate image are preprocessed.

[0096] The generating a disassembly solution using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image includes: The information of the distribution transformer, the preprocessed optical cover plate image and the preprocessed infrared cover plate image are input into the disassembly model to generate a disassembly plan.

[0097] In summary, the embodiments of the present application have at least the following beneficial effects: According to an embodiment of the present application, after detecting that a manipulator starts to perform a cover disassembly operation on a distribution transformer, an optical cover image and an infrared cover image are obtained by photographing the cover of the distribution transformer through an image acquisition device; based on the optical cover image and the infrared cover image, a disassembly model is used to generate a disassembly plan, and a control instruction is sent to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover disassembly operation of the manipulator; wherein the training process of the disassembly model includes: obtaining a sample optical image and a sample infrared image corresponding to a sample distribution transformer; using a model to be trained, respectively encoding the sample optical image and the sample infrared image, and respectively obtaining an optical feature corresponding to the sample optical image and an infrared feature corresponding to the sample infrared image generated during the encoding process; based on the optical feature, the infrared feature and the information of the sample distribution transformer, a first disassembly plan is generated using a large model; based on the optical feature, the infrared feature and the first disassembly plan, the model to be trained is trained to obtain the disassembly model, thereby improving the disassembly efficiency and safety of the cover of the distribution transformer.

[0098] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary hardware platform, and of course, it can also be implemented entirely by hardware. Based on such an understanding, all or part of the contribution of the technical solution of the present application to the background technology can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application or some parts of the embodiments.

[0099] The above is a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications are also considered to be within the scope of protection of the present application.

Claims

1. A method for disassembling a distribution transformer cover, characterized in that: include: After detecting that the manipulator starts to perform the cover disassembly operation on the distribution transformer, the cover of the distribution transformer is photographed by an image acquisition device to obtain an optical cover image and an infrared cover image; Based on the optical cover plate image and the infrared cover plate image, a disassembly model is used to generate a disassembly plan, and a control instruction is sent to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover plate disassembly operation of the manipulator; The training process of the disassembly model includes: Acquire a sample optical image and a sample infrared image corresponding to a sample distribution transformer; Using the model to be trained, respectively encoding the sample optical image and the sample infrared image, and respectively obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process; Based on the optical features, the infrared features and the information of the sample distribution transformer, a first disassembly plan is generated using a large model; Based on the optical features, the infrared features and the first disassembly scheme, the model to be trained is trained to obtain the disassembly model.

2. The method for disassembling a distribution transformer cover plate according to claim 1, characterized in that: The method of generating a first disassembly solution using a large model based on the optical features, the infrared features and the information of the sample distribution transformer includes: Performing inner product processing on the optical feature and the infrared feature to obtain an inner product feature; The inner product feature and the information of the sample distribution transformer are input into a large model to obtain a first disassembly solution output by the large model.

3. The method for disassembling a distribution transformer cover plate according to claim 2, characterized in that: The information of the sample distribution transformer includes a standard screw profile template diagram and a standard temperature distribution diagram of the sample distribution transformer, and inputting the inner product feature and the information of the sample distribution transformer into a large model to obtain a first disassembly solution output by the large model, including: Inputting the inner product feature into the large model, so that the large model generates a simulated screw contour template diagram and a simulated temperature distribution diagram according to the inner product feature; The standard screw contour template image and the standard temperature distribution image are input into the large model, so that the large model generates and outputs the first disassembly plan based on the difference between the simulated screw contour template image and the standard screw contour template image, and the difference between the simulated temperature distribution image and the standard temperature distribution image.

4. The method for disassembling a distribution transformer cover plate according to claim 1, characterized in that: The step of training the model to be trained based on the optical feature, the infrared feature and the first disassembly scheme to obtain the disassembly model includes: Performing inner product processing on the optical feature and the infrared feature to obtain an inner product feature; Inputting the inner product feature into the model to be trained to generate a second disassembly solution; Extracting a first text feature of the first disassembly solution, and extracting a second text feature of the second disassembly solution; Determining a text feature difference between the first text feature and the second text feature using a preset loss function; Based on the text feature differences, the model to be trained is trained to obtain the disassembly model.

5. The method for disassembling a distribution transformer cover plate according to any one of claims 2 to 4, characterized in that: The sample optical image is cropped into a plurality of optical partial images, and the sample infrared image is cropped into a plurality of infrared partial images, N optical partial images among the plurality of optical partial images correspond one to one with N infrared partial images among the plurality of infrared partial images, N is a positive integer, the corresponding optical partial images are the same as the partial areas indicated by the infrared partial images, the optical features at least include the optical partial features corresponding to the N optical partial images, and the infrared features at least include the infrared partial features corresponding to the N infrared partial images; The method of using the model to be trained to encode the sample optical image and the sample infrared image respectively, and obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process respectively, includes: Using the model to be trained, encoding the sample optical image, and obtaining the optical local features corresponding to each of the N optical local images generated during the encoding process, The sample infrared image is encoded using the model to be trained, and infrared local features corresponding to each of the N infrared local images generated during the encoding process are obtained.

6. The method for disassembling a distribution transformer cover plate according to claim 5, characterized in that: Each optical partial image in the N optical partial images and its corresponding infrared partial image constitute a partial image pair, and the inner product processing is performed on the optical feature and the infrared feature to obtain the inner product feature, including: For each of the local image pairs, inner product processing is performed on the optical local features and the infrared local features corresponding to the local image pair to obtain the inner product local features corresponding to the local image pair; The inner product feature is determined based on the inner product local features corresponding to all the local image pairs.

7. The method for disassembling a distribution transformer cover plate according to claim 6, characterized in that: The training process also includes: Determining a weight corresponding to each of the partial image pairs according to the partial area indicated by each of the partial image pairs; The determining the inner product feature based on the inner product local features corresponding to all the local image pairs respectively comprises: Based on the weight corresponding to each of the local image pairs, weighted processing is performed on the inner product local features corresponding to each of the local image pairs; The weighted inner product local features corresponding to all the local image pairs are concatenated into the inner product feature.

8. The method for disassembling a distribution transformer cover plate according to any one of claims 1 to 4, characterized in that: Before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the method further includes: Obtaining information of the distribution transformer; The step of generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image includes: A disassembly solution is generated using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image.

9. The method for disassembling a distribution transformer cover plate according to claim 8, characterized in that: Before generating a disassembly solution using a disassembly model based on the optical cover plate image and the infrared cover plate image, the method further includes: Preprocessing the optical cover plate image and the infrared cover plate image; The generating a disassembly solution using a disassembly model based on the information of the distribution transformer, the optical cover plate image, and the infrared cover plate image includes: The information of the distribution transformer, the preprocessed optical cover plate image and the preprocessed infrared cover plate image are input into the disassembly model to generate a disassembly plan.

10. A disassembly system for a distribution transformer cover, characterized in that: include: Manipulator; Image acquisition equipment; as well as, A control device is communicatively connected with the manipulator and the image acquisition device, and the control device is configured as follows: After detecting that the manipulator starts to perform the cover disassembly operation on the distribution transformer, the cover of the distribution transformer is photographed by the image acquisition device to obtain an optical cover image and an infrared cover image; Based on the optical cover plate image and the infrared cover plate image, a disassembly model is used to generate a disassembly plan, and a control instruction is sent to the manipulator according to the disassembly plan, wherein the control instruction is used to regulate the cover plate disassembly operation of the manipulator; The training process of the disassembly model includes: Acquire a sample optical image and a sample infrared image corresponding to a sample distribution transformer; Using the model to be trained, respectively encoding the sample optical image and the sample infrared image, and respectively obtaining the optical features corresponding to the sample optical image and the infrared features corresponding to the sample infrared image generated during the encoding process; Based on the optical features, the infrared features and the information of the sample distribution transformer, a first disassembly plan is generated using a large model; Based on the optical features, the infrared features and the first disassembly scheme, the model to be trained is trained to obtain the disassembly model.

Citation Information

Patent Citations

  • Carrier roller monitoring method and device based on three-light fusion and inspection robot

    CN115239652A

  • Battery pack disassembling method and system, electronic equipment and computer medium

    CN116433618A

  • Insulator defect detection method and system and computer readable storage medium

    CN116468714A

  • Hyperspectral face fusion and recognition method, electronic equipment and storage medium

    CN117576755A

  • Waste mobile phone part recycling and disassembling method, computer equipment and medium

    CN118506365A