CN116883717A - Platen checking method and device, computer readable storage medium and computer equipment - Google Patents
Platen checking method and device, computer readable storage medium and computer equipment Download PDFInfo
- Publication number
- CN116883717A CN116883717A CN202310679106.8A CN202310679106A CN116883717A CN 116883717 A CN116883717 A CN 116883717A CN 202310679106 A CN202310679106 A CN 202310679106A CN 116883717 A CN116883717 A CN 116883717A
- Authority
- CN
- China
- Prior art keywords
- target
- checking
- platen
- data
- pressing plate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000003860 storage Methods 0.000 title claims abstract description 17
- 238000003825 pressing Methods 0.000 claims abstract description 68
- 238000001514 detection method Methods 0.000 claims description 27
- 238000012423 maintenance Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 14
- 238000013145 classification model Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000012795 verification Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 description 14
- 230000001681 protective effect Effects 0.000 description 12
- 230000004913 activation Effects 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 9
- 238000004422 calculation algorithm Methods 0.000 description 8
- 238000013527 convolutional neural network Methods 0.000 description 7
- 238000007689 inspection Methods 0.000 description 4
- 230000010354 integration Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000036961 partial effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000011282 treatment Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000009993 protective function Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a platen checking method, a platen checking device, a computer readable storage medium and computer equipment. Wherein the method comprises the following steps: receiving a pressing plate checking instruction; responding to a press plate checking instruction, and shooting a designated area of a target press plate to obtain a target picture; performing image recognition on the target picture to obtain a target checking result; and generating a target checking report corresponding to the pressing plate checking instruction based on the target checking result. The application solves the technical problems of long time consumption, low efficiency and high error rate of the checking mode aiming at the protection pressing plate.
Description
Technical Field
The application relates to the field of operation and maintenance of transformer substations, in particular to a method and a device for checking a pressing plate, a computer readable storage medium and computer equipment.
Background
In the related art, manual checking of the protective platen is generally performed using a paper form, but such checking requires a long time, is low in work efficiency, and has a high checking error rate.
In view of the above problems, no effective solution has been proposed at present.
Disclosure of Invention
The embodiment of the application provides a method and a device for checking a pressing plate, a computer readable storage medium and computer equipment, which are used for at least solving the technical problems of long time consumption, low efficiency and high error rate of a checking mode aiming at a protection pressing plate.
According to an aspect of an embodiment of the present application, there is provided a platen checking method including: receiving a pressing plate checking instruction; responding to a press plate checking instruction, and shooting a designated area of a target press plate to obtain a target picture; performing image recognition on the target picture to obtain a target checking result; and generating a target checking report corresponding to the pressing plate checking instruction based on the target checking result.
Optionally, before receiving the platen check instruction, the method further includes: leading in the pressing plate data and maintaining the pressing plate data; determining check task information and maintaining the check task information; and generating a platen check task based on the platen data and the maintenance result of the check task information, wherein the platen check task corresponds to the platen check instruction.
Optionally, importing platen data and maintaining the platen data includes: and importing new added pressure plate data according to the preset data items, and/or maintaining the pressure plate data according to a preset maintenance template, and importing the maintained pressure plate data.
Optionally, performing image recognition on the target picture to obtain a target checking result, including: judging whether the target picture is consistent with the platen data imported in the database by adopting an image recognition model; and determining a target checking result based on the judging result.
Optionally, before the image recognition model is adopted to determine whether the target image is consistent with the platen data imported in the database, the method further includes: establishing an initial model based on the target detection model and the classification model; acquiring a plurality of groups of sample pictures, and preprocessing the plurality of groups of sample pictures; and training an initial model by adopting the preprocessed multiple groups of sample pictures to obtain an image recognition model.
Optionally, determining the target checking result based on the judging result includes: if the judgment result is inconsistent, performing secondary check on the designated area corresponding to the target picture; and determining a target checking result based on the result of the secondary checking.
Optionally, the method further comprises: and when the target pictures are a plurality of, sequencing the plurality of target pictures.
According to another aspect of the embodiment of the present application, there is also provided a platen check apparatus including: the receiving module is used for receiving the pressing plate checking instruction; the shooting module is used for responding to the press plate checking instruction and shooting a designated area of the target press plate to obtain a target picture; the identification module is used for carrying out image identification on the target picture to obtain a target checking result; and the generating module is used for generating a target checking report corresponding to the pressing plate checking instruction based on the target checking result.
According to another aspect of the embodiment of the present application, there is further provided a computer readable storage medium, where the computer readable storage medium includes a stored program, and when the program runs, the apparatus where the computer readable storage medium is controlled to execute the platen checking method of any one of the above.
According to another aspect of an embodiment of the present application, there is also provided a computer apparatus including: a memory and a processor, the memory storing a computer program; and the processor is used for executing a computer program stored in the memory, and the computer program enables the processor to execute the platen checking method.
In the embodiment of the application, the image recognition mode is adopted, after the press plate checking instruction is received, the region needing to be checked for the target press plate, namely the designated region, is shot to obtain the target image, the image recognition is carried out on the target image, the target checking result is directly determined according to the image recognition result, and the target checking report corresponding to the press plate checking instruction can be generated based on the target checking result, so that the traditional mode of manually checking one by one is replaced, the checking efficiency and accuracy of the target press plate are greatly improved, the technical effects of reducing the time consumption of checking the press plate, improving the checking efficiency and reducing the checking error rate are realized, and the technical problems of long time consumption, low efficiency and high error rate of the checking mode aiming at the protection press plate are further solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:
FIG. 1 is a flow chart of a platen check method provided in accordance with an embodiment of the present application;
FIG. 2 is a check platen workflow diagram provided in accordance with the prior art;
FIG. 3 is a flow chart of a pre-job preparation-model integration implementation of image recognition techniques provided in accordance with an alternative embodiment of the present application;
FIG. 4 is a flowchart of an image recognition technique implementation pre-job preparation-platen library maintenance provided in accordance with an alternative embodiment of the present application;
FIG. 5 is a flowchart of an image recognition technique implementation pre-job preparation-platen library maintenance provided in accordance with an alternative embodiment of the present application;
fig. 6 is a block diagram of a platen check apparatus provided in accordance with an alternative embodiment of the present application.
Detailed Description
In order that those skilled in the art will better understand the present application, a technical solution in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, shall fall within the scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Description of the terms
The protection pressing plate, also called trip pressing plate, is a bridge and a tie of the protection device connected with external wiring, and relates to the protection function and whether the action outlet can normally play a role.
When the operation mode of the power grid system is changed, the switching of the protective pressing plate is often involved, whether the pressing plate state of the protective device is correct or not can directly influence the realization of the protective function, and even the protective action rejection or the false action can be caused when serious, so that a large-range power failure accident is caused, and therefore operation and maintenance personnel should check and check all the names, device models, pressing plate position information and the like of the protective cabinets of the total station regularly.
Fig. 2 is a working flow chart of checking press plates according to the prior art, and as shown in fig. 2, the checking flow mainly comprises five stages of press plate data arrangement, early work preparation, press plate field checking, consistency judgment and record arrangement retention. At present, the checking protection pressing plate can only check through manual paper forms, and the traditional checking mode is long in time consumption, low in working efficiency and high in checking error rate.
In view of the foregoing, embodiments of the present application provide a method embodiment for platen check, it should be noted that the steps illustrated in the flowchart of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flowchart, in some cases, the steps illustrated or described may be performed in an order other than that illustrated herein.
Fig. 1 is a flowchart of a platen checking method according to an embodiment of the present application, as shown in fig. 1, the method includes the following steps:
step S102, receiving a pressing plate checking instruction;
step S104, responding to a press plate checking instruction, and shooting a designated area of a target press plate to obtain a target picture;
step S108, performing image recognition on the target picture to obtain a target checking result;
step S110, generating a target checking report corresponding to the pressing plate checking instruction based on the target checking result.
Through the steps, the image recognition mode is adopted, after the press plate checking instruction is received, the region which needs to be checked for the target press plate, namely the designated region, is shot to obtain a target image, the image recognition is carried out on the target image, the target checking result is directly determined according to the image recognition result, and the target checking report corresponding to the press plate checking instruction can be generated based on the target checking result, so that the traditional mode of manually checking one by one is replaced, the checking efficiency and accuracy of the target press plate are greatly improved, the technical effects of reducing the time consumption of press plate checking, improving the checking efficiency and reducing the checking error rate are realized, and the technical problems of long time consumption, low efficiency and high error rate of the checking mode for protecting the press plate are further solved.
It should be noted that, the method provided by the embodiment not only can carry out high-efficiency check on the pressing plate, but also can carry out unified high-efficiency check on relevant protection devices of the transformer substation such as handles, indicator lamps and the like.
As an alternative embodiment, before receiving the platen check instruction, the method further includes: leading in the pressing plate data and maintaining the pressing plate data; determining check task information and maintaining the check task information; and generating a platen check task based on the platen data and the maintenance result of the check task information, wherein the platen check task corresponds to the platen check instruction.
Wherein the platen data includes at least one of: the names of the pressing plate, the handle and the indicator lamp, the normal operation positions of the pressing plate, the handle and the indicator lamp, remarks, transformer substation names, dispatching numbers, road names, protection device types and checking results. The checking task information includes at least one of the following: the maintenance work responsible person, the work member, the planning start time and the planning end time, whether the pressing plate checking task is automatically generated, checking the patrol content and remarks. And maintaining the platen data and the checking task information, namely updating, debugging, reducing data redundancy and the like, so as to ensure that the platen checking task generated based on the platen data and the checking task information is effective.
Preferably, the above-mentioned platen check task may be displayed on a terminal interface, for example, the terminal may generate a platen check task based on platen data and check task information that have been subjected to data maintenance in the database, and display the platen check task on the display screen, and perform a selected operation, that is, a platen check instruction, on the interface through the user, where after receiving the platen check instruction, the terminal may shoot a designated area of the target platen, and continue to perform the subsequent check operation.
As an alternative embodiment, importing platen data and maintaining the platen data includes: and importing new added pressure plate data according to the preset data items, and/or maintaining the pressure plate data according to a preset maintenance template, and importing the maintained pressure plate data.
When the platen data and the check task information are imported, various modes can be adopted, for example, when new data are needed, the new data can be filled in according to the type, the name, the normal operation position, the remarks and the like of the protection device, the existing data can be maintained according to the name, the scheduling number, the road name, the type of the protection device, the name, the normal operation position, the check result, the remarks and the like of the protection device, and after the data maintenance is finished, the maintained data are imported.
As an optional embodiment, performing image recognition on the target picture to obtain a target checking result, including: judging whether the target picture is consistent with the platen data imported in the database by adopting an image recognition model; and determining a target checking result based on the judging result.
In this embodiment, by adopting the image recognition model to perform image recognition on the shot target picture, it can be efficiently and quickly determined whether the pressing plate (or the indicator light, the handle, etc.) in the current target picture is consistent with the data imported in the database, so as to complete the checking, and obtain the target checking result.
As an alternative embodiment, before using the image recognition model to determine whether the target picture is consistent with the platen data imported in the database, the method further includes: establishing an initial model based on the target detection model and the classification model; acquiring a plurality of groups of sample pictures, and preprocessing the plurality of groups of sample pictures; and training an initial model by adopting the preprocessed multiple groups of sample pictures to obtain an image recognition model.
In order to improve the speed and precision of image recognition, in this embodiment, a mode of determining the position of the target object in the image first, and then classifying the target object is adopted, that is, an initial model is constructed based on a target detection model and a classification model, then platen image data and/or video data with different types, different sizes and different shapes are collected, and the image data with higher quality and definition is obtained by denoising, enhancing contrast, adjusting brightness and the like the data, that is, a plurality of groups of preprocessed sample data are adopted, and training is performed on the initial model based on the plurality of groups of preprocessed sample data, so that the image recognition model in this embodiment can be obtained.
Preferably, the object detection model in this embodiment may be a YOLOv4 detection model, and the classification model may be a ShuffleNet-v2 switch class classification model, where an activation function mich is used as a nonlinear activation unit of the network in the YOLOv4 detection model.
As an alternative embodiment, determining the target checking result based on the determination result includes: if the judgment result is inconsistent, performing secondary check on the designated area corresponding to the target picture; and determining a target checking result based on the result of the secondary checking.
If the determination result is consistent, it is indicated that the current platen state is consistent with the state recorded by the data imported in the database, and if the determination result is inconsistent, it is indicated that the current platen state is inconsistent with the state recorded by the data imported in the database, at this time, the embodiment may perform secondary checking on the designated area corresponding to the target picture, for example, may perform secondary checking by re-shooting the target picture to perform image recognition again, so as to avoid inconsistent determination results caused by shooting problems, or may send a prompt to the user based on the determination result, so as to prompt the user to perform manual checking on the designated area, and so on.
As an alternative embodiment, the method further comprises: and when the target pictures are a plurality of, sequencing the plurality of target pictures. When the target pictures are multiple, the target pictures can be subjected to self-defined sorting according to checking time, names, voltage levels of the pressing plates and the like, and fuzzy inquiry and other operations can be performed on the target pictures.
Based on the foregoing embodiment and the optional embodiments, an optional implementation manner is further provided in the present application, and the following description is provided.
The application provides a protective pressing plate identification method based on an image identification technology.
(1) Image recognition model
In an alternative embodiment of the application, the protection pressing plate identification is realized by adopting an image processing technology and a deep learning model respectively, the switch position distribution of the strip switch pressing plate can be detected through a single frame image, the switching state of each switch is identified, the switch position distribution of the connecting sheet type switch pressing plate can also be detected through a single frame image, and the switching state of each switch is identified.
Aiming at the extracted interested region, a target detection model is firstly adopted to detect the coordinate position of the switch in the image, and then a classification model is utilized to realize fine granularity classification identification of the switch state.
1) Target detection model YOLOv4
There are various target detection schemes based on deep learning, including a Two-stage method represented by fast R-CNN and an One-stage method represented by SSD, YOLO, etc., from the implementation steps, and an Anchor-base method and an Anchor-fre method from the setting of a preset frame. In the aspect of comprehensively considering precision, detection speed and the like, the optional implementation mode of the application adopts a YOLO algorithm as a basic framework for target detection. YOLO is used as an end-to-end single-step target detection model, and compared with the R-CNN series, only a single neural network skeleton is used for completing the whole detection task, so that the time consumption of training and detection of the network can be greatly reduced. Specifically, the target YOLOv4 model is selected as the optional implementation mode of the application, because the model has better model precision and time efficiency.
1. The backbone network selects CSP-Darknet53
The backbone network (feature description network) is a key part of the deep learning model, and its structure has a very large impact on its description ability to extract features. Considering network performance comprehensively, the network is described herein using CSP-DarkNet53 as a feature of the detection model. CSP-DarkNet53 takes along the filter size and overall structure of DarkNet53 and introduces Cross Stage Part ial structure into the network while eliminating the bottleneck structure, thereby reducing network parameters and making it easier to train. Darknet-53 is a deep convolutional neural network, consisting of 53 convolutional layers in total. The filter size of the network adopts conventional small-size filters with the sizes of 1 multiplied by 1 and 3 multiplied by 3, and a large receptive field can be obtained by connecting multiple small-size filter layers in series, and the parameter number of the network is kept low. Cross Stage Part ial (CSP) structure, which initially reduces the computational effort and enhances the performance of gradients. The main idea is: before inputting a block, the input is divided into two parts, wherein one part is calculated through the block, and the other part is directly connected through a shortbutton. Conventionally, the convolutional layer of the network is followed by a BN to aid in network training.
2. Activation function Mish activation unit
The activation function is used for improving the learning ability of the network and improving the gradient transfer efficiency. The activation functions commonly used by CNNs are also continuously developed, and early networks commonly used are ReLU, leakyrlu, softplus, etc., and later Swish, dash, etc. The alternative embodiment of the application uses a Mish function as a nonlinear activation unit of the network, and the calculation formula is as follows:
ζ(x)＝log(1+e x )
Mish(x)＝x·tanh(ζ(x))
mish is free of forward boundary, and gradient saturation can be avoided; secondly the mich function is everywhere smooth and allows some negative values in the negative regions where the absolute value is small.
After the backbone network, a feature fusion network is adopted according to the YOLOv4 idea to realize the reuse and fusion of features. The structure of the feature fusion network incorporates feature pyramid enhancement modules (Feature Pyramid Enhancement Module, FPEM) in the PANNe network for enhancing features. And SPP (Spat ial Pyramid Pool ing) is adopted to construct multi-scale features before different feature layers are stacked, so that the performance of the network on scale invariance is improved. In the training stage, a data set is constructed based on the region of interest extracted from the original image of the acquired image and expanding the image of the region of interest. In training, momentum-SGD is adopted as an optimization function, and MSE and cross entropy are adopted as loss functions. After the training of the target detection model is completed, the position of the switch to be identified in the image and the type of the switch can be obtained from the image of the region of interest.
2) Switch class classification model ShuffleNet-v2
Since the information amount of the protective pressure plate switch is small and the same switch characteristics of the on-off state are similar, the classification of finer granularity is required after the target detection is completed. Because the ShuffleNet-v2 belongs to a lightweight network, the efficiency in terms of reasoning time is high, and the model accuracy is good. The network structure mainly comprises the ideas of packet convolution and channel switching, and can fully help information flow. During training, 8 classification tasks are formulated for the on-off states of the 4 switches, and the output result of the target detection model is utilized to obtain local images of the switches, so that a classification data set is formed. The training process employs Adam optimizers and uses cross entropy as a loss function.
Through the classification model, the algorithm can finally identify the state of the switch on the basis of the output result of the target detection model. Finally, the switch position and the switch state of the whole task can be obtained.
(2) Early work preparation-model integration
FIG. 3 is a flow chart of a pre-job preparation-model integration implementation of image recognition techniques provided in accordance with an alternative embodiment of the present application.
1) Data acquisition and preprocessing
First, in the solution of the protective platen recognition technology, data acquisition is a very important step. Cameras, and the like are commonly employed to acquire image or video data of the protective platen. In order to improve the recognition accuracy, a high-resolution device should be selected, so that errors caused by the device itself are avoided. The alternative implementation mode of the application collects the image data of the pressing plates with different types, different sizes and different shapes and carries out pretreatment, in the pretreatment step, the collected images or videos need to be subjected to denoising, contrast enhancement, brightness adjustment and other treatments so as to improve the image quality and definition, and the subsequent treatment and recognition are convenient to prepare for the subsequent model training.
2) Model training and optimization
Alternative embodiments of the present application will use deep learning algorithms, such as convolutional neural networks (Convolut ional Neural Networks, CNNs), to train the pre-processed platen image data to construct an efficient and accurate platen identification model. The model is continuously optimized, so that the recognition accuracy and the robustness of the model are improved.
3) Platen identification system integration
According to the application, the trained press plate recognition model is integrated into the system in an alternative embodiment, so that automatic press plate recognition is realized. And comparing the result of pattern recognition with a preset standard to judge the state of the protective pressing plate.
4) System testing and debugging
The optional implementation mode of the application carries out comprehensive test and debugging on the integrated protective pressing plate recognition algorithm, and ensures the stability and reliability of the integrated protective pressing plate recognition algorithm in the actual production environment. The optional implementation mode of the application can perform system optimization and improvement according to feedback and requirements so as to meet the actual application requirements.
5) Technical support and maintenance
The optional implementation mode of the application provides continuous technical support and maintenance service for customers, including daily operation monitoring, fault detection, problem solving and the like of the system, and ensures the long-term stability and performance of the system.
(3) Early work preparation-platen library maintenance
FIG. 4 is a flow chart of an image recognition technique implementation pre-job preparation-platen library maintenance provided in accordance with an alternative embodiment of the present application.
1) Importing historical platen data
1. Method 1: the new pressing plate information is maintained, and the pressing plate information is filled with a protecting device, a pressing plate/a handle/an indicating lamp, a pressing plate handle, an indicating lamp name, a normal operation position and remarks.
2. Method 2: and maintaining the template by using the pressing plate library, maintaining the information such as the station name, the dispatching number, the road name, the type of the protection device, the pressing plate/handle/indicator lamp name, the normal operation position, the checking result, the remarks and the like on the template, storing the information to the local after the maintenance is completed, and importing the template subjected to the selective maintenance, so that the successful importing can be achieved.
2) On-site checking
And (3) shooting images on the current pressing plate, the handles and the state of the indicator lamps by using the APP, and automatically checking.
3) Checking whether or not to be manual
Clicking a certain screen cabinet, respectively carrying out image recognition on pictures shot by the indicator lamp, the handle and the pressing plate, judging whether the states of the indicator lamp, the handle and the pressing plate are consistent with the imported data by the APP, and carrying out manual check on some pictures with inaccurate recognition results
4) Accurate history and image recognition verification
If the APP judges that the prompting pressing plate, the handle and the indicator lamp are inconsistent with the imported data, then the APP carries out manual check and correction one by one.
5) Checking and warehousing the pressing plate
And (5) finishing the maintenance work of the pressing plate warehouse.
(4) Platen field check
FIG. 5 is a flow chart of an image recognition technique implementation pre-job preparation-platen library maintenance provided in accordance with an alternative embodiment of the present application.
1) The pressing plate checks the regular patrol task work order dispatch list
After the transformer substation pressing plate library is maintained, the non-periodic inspection task of the transformer substation is newly increased, and the contents of fields such as maintenance work responsible personnel, work members, planning starting time, planning ending time, whether automatic dispatch is performed, inspection contents, remarks and the like are needed. After the information maintenance is completed, a work order is generated by one key, and a patrol work order is automatically generated by the work order. The work order can be dispatched in the substation work order management.
2) Order receiving
And entering an APP order receiving page, clicking the order receiving page, and then entering an on-site page. Arriving at the scene requires taking a photograph of the arriving scene.
3) Shooting indicator lamp and pressing plate, intelligent screenshot checking
Entering the checking page of the transformer substation pressing plate, showing and shooting the checking states of the indicator lamp, the handle and the pressing plate according to the protection screen, and carrying out self-defined sorting according to checking time, names and voltage levels, and further supporting fuzzy inquiry.
4) Checking whether or not to be manual
Clicking a certain screen cabinet, respectively carrying out image recognition on pictures shot by the indicator lamp, the handle and the pressing plate, judging whether the states of the indicator lamp, the handle and the pressing plate are consistent with the imported data by the APP, and carrying out manual check on some pictures with inaccurate recognition results
5) Accurate history and image recognition verification
If the APP judges that the prompting pressing plate, the handle and the indicator lamp are inconsistent with the imported data, then the APP carries out manual check and correction one by one.
6) Work order filing (pressing plate checking normal warehouse)
After checking all the protection screens of all the stations in the work order, submitting the inspection work order and finally forming an inspection report.
In summary, alternative embodiments of the present application have the following advantages:
1. the image recognition technology is utilized to effectively solve the problems of low working efficiency, long working time, high checking error rate and the like of manual checking of the pressing plate. The method is novel in technology in the field of image recognition, has excellent algorithm performance, solves the problems of low working efficiency and long working time of the original manual-checking-based pressing plate method, and improves checking precision. An image recognition method is designed by constructing a target YOLOv4 detection model and a ShuffeNet-v 2 switch class classification model and applying a related algorithm. The image recognition technology has the advantages of high operation speed, high reliability and the like, so that the work of checking the pressing plate is more convenient and simple.
2. The image recognition technology of the optional implementation mode adopts the target YOLOv4 detection model, the algorithm adopts a very excellent optimization strategy, the detection speed is very high, a large amount of network training and consumption of detection time can be saved, and the method has higher model precision.
YOLOv4 chooses to use CSP-dark net53 network architecture as backbone network for extracting target features. The image feature extraction capability is strong, and the CSP structure solves the problem of information repetition in the backbone network, so that the reasoning speed of the final model is improved. The CSP-DarkNet53 network structure improves the capability of the convolution network for feature extraction on the premise of not losing detection precision, and accelerates the detection speed. The calculation loss of the whole model is reduced, and the quick dimension reduction of the feature map is easy to realize.
The YOLOv4 uses an activation function Mish as a nonlinear activation unit of the network, is a novel activation function, and can improve the learning capacity of the network and the gradient transfer efficiency. Compared with the traditional activation function, the method has better stability, faster convergence speed, higher smoothness and stronger generalization performance, can better avoid overfitting, and can prevent the problems of gradient disappearance and gradient explosion.
3. The image recognition technology of the alternative implementation mode of the application adopts a SheffleNet-v 2 switch class classification model, so that good balance is made between speed and accuracy, the performance exceeds the current mainstream network, and a higher level is achieved. First of all, in its network design, the modular unit enables the network to have a greater number of channels and network capacity. Secondly, the special channel directly penetrates through the module unit to connect with the next module, namely, the characteristic is reused. The method has the advantages of higher efficiency in the aspect of reasoning time, good model precision and extremely high recognition accuracy.
According to an embodiment of the present application, there is further provided a platen check apparatus, and fig. 6 is a block diagram of a platen check apparatus according to an alternative embodiment of the present application, as shown in fig. 6, including: the receiving module 61, the photographing module 62, the identifying module 63 and the generating module 64 will be described below.
A receiving module 61, configured to receive a platen check instruction; the shooting module 62 is connected to the receiving module 61 and is used for responding to the checking instruction of the pressing plate and shooting the appointed area of the target pressing plate to obtain a target picture; the identification module 63 is connected to the shooting module 62 and is used for carrying out image identification on the target picture to obtain a target checking result; the generating module 64 is connected to the identifying module 63, and is configured to generate a target checking report corresponding to the platen checking instruction based on the target checking result.
According to an embodiment of the present application, there is further provided a computer readable storage medium, where the computer readable storage medium includes a stored program, and when the program runs, the apparatus where the computer readable storage medium is controlled to execute the platen checking method of any one of the above-mentioned items.
According to an embodiment of the present application, there is also provided a computer apparatus including: a memory and a processor, the memory storing a computer program; and the processor is used for executing a computer program stored in the memory, and the computer program enables the processor to execute the platen checking method.
The foregoing embodiment numbers of the present application are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
In the foregoing embodiments of the present application, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.
In the several embodiments provided in the present application, it should be understood that the disclosed technology may be implemented in other manners. The above-described embodiments of the apparatus are merely exemplary, and the division of the units, for example, may be a logic function division, and may be implemented in another manner, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some interfaces, units or modules, or may be in electrical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The foregoing is merely a preferred embodiment of the present application and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present application, which are intended to be comprehended within the scope of the present application.
Claims (10)
1. A platen check method, comprising:
receiving a pressing plate checking instruction;
responding to the press plate checking instruction, shooting a designated area of a target press plate, and obtaining a target picture;
performing image recognition on the target picture to obtain a target checking result;
and generating a target checking report corresponding to the pressing plate checking instruction based on the target checking result.
2. The method of claim 1, wherein prior to said receiving platen check instructions, the method further comprises:
leading in the pressing plate data and maintaining the pressing plate data;
determining check task information, and maintaining the check task information;
and generating a platen check task based on the platen data and the maintenance result of the check task information, wherein the platen check task corresponds to the platen check instruction.
3. The method of claim 2, wherein the importing platen data and maintaining the platen data comprises:
according to the predetermined data item, new added board data is imported,
and/or the number of the groups of groups,
and maintaining the pressure plate data according to a preset maintenance template, and importing the maintained pressure plate data.
4. The method of claim 1, wherein the performing image recognition on the target picture to obtain a target verification result includes:
judging whether the target picture is consistent with the platen data imported in the database by adopting an image recognition model;
and determining the target checking result based on the judging result.
5. The method of claim 4, further comprising, prior to said determining, using an image recognition model, whether the target picture is consistent with platen data imported in a database:
establishing an initial model based on the target detection model and the classification model;
obtaining a plurality of groups of sample pictures, and preprocessing the plurality of groups of sample pictures;
and training the initial model by adopting the preprocessed multiple groups of sample pictures to obtain the image recognition model.
6. The method of claim 4, wherein determining the target verification result based on the determination result comprises:
if the judging result is inconsistent, performing secondary check on the designated area corresponding to the target picture;
and determining the target checking result based on the result of the secondary checking.
7. The method according to any one of claims 1 to 6, further comprising:
and ordering the plurality of target pictures when the target pictures are a plurality of.
8. A platen check apparatus, comprising:
the receiving module is used for receiving the pressing plate checking instruction;
the shooting module is used for responding to the press plate checking instruction and shooting a designated area of the target press plate to obtain a target picture;
the identification module is used for carrying out image identification on the target picture to obtain a target checking result;
and the generating module is used for generating a target checking report corresponding to the pressing plate checking instruction based on the target checking result.
9. A computer-readable storage medium, characterized in that the computer-readable storage medium includes a stored program, wherein the program, when run, controls a device in which the computer-readable storage medium is located to perform the platen check method according to any one of claims 1 to 7.
10. A computer device, comprising: a memory and a processor, wherein the memory is configured to store,
the memory stores a computer program;
the processor configured to execute a computer program stored in the memory, the computer program when executed causing the processor to perform the platen check method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310679106.8A CN116883717A (en) | 2023-06-08 | 2023-06-08 | Platen checking method and device, computer readable storage medium and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310679106.8A CN116883717A (en) | 2023-06-08 | 2023-06-08 | Platen checking method and device, computer readable storage medium and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116883717A true CN116883717A (en) | 2023-10-13 |
Family
ID=88253805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310679106.8A Pending CN116883717A (en) | 2023-06-08 | 2023-06-08 | Platen checking method and device, computer readable storage medium and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116883717A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541028A (en) * | 2024-01-09 | 2024-02-09 | 国网山东省电力公司菏泽供电公司 | Management system for protecting pressing plate |
-
2023
- 2023-06-08 CN CN202310679106.8A patent/CN116883717A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541028A (en) * | 2024-01-09 | 2024-02-09 | 国网山东省电力公司菏泽供电公司 | Management system for protecting pressing plate |
| CN117541028B (en) * | 2024-01-09 | 2024-04-12 | 国网山东省电力公司菏泽供电公司 | Management system for protecting pressing plate |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105372581B (en) | Flexible circuit board manufacturing process monitors automatically and intelligent analysis system and method | |
| CN111259892B (en) | Inspection method, inspection device, inspection equipment and inspection medium for state of indicator lamp | |
| CN108596277A (en) | A kind of testing vehicle register identification method, apparatus and storage medium | |
| CN110348441A (en) | VAT invoice recognition methods, device, computer equipment and storage medium | |
| CN108520114A (en) | A kind of textile cloth defect detection model and its training method and application | |
| CN111368089A (en) | Service processing method and device based on knowledge graph | |
| CN110097451B (en) | Bank business monitoring method and device | |
| CN116883717A (en) | Platen checking method and device, computer readable storage medium and computer equipment | |
| CN109617233A (en) | Protect pressing plate automatic checking method of calibration | |
| CN111382808A (en) | Vehicle detection processing method and device | |
| CN110379036A (en) | Intelligent substation patrol recognition methods, system, device and storage medium | |
| CN107291863B (en) | A kind of quick check method and device for industrial control equipment information security | |
| CN113762081B (en) | YOLOv5 s-based granary pest detection method | |
| CN112597995B (en) | License plate detection model training method, device, equipment and medium | |
| CN114549812A (en) | Intelligent substation relay protection hard pressing plate checking method based on target detection model | |
| CN110113327A (en) | A kind of method and device detecting DGA domain name | |
| CN109087281A (en) | Display screen peripheral circuit detection method, device, electronic equipment and storage medium | |
| CN111126264A (en) | Image processing method, device, equipment and storage medium | |
| CN113794277A (en) | Method and system for identifying pressing plate state image | |
| CN109726710A (en) | Invoice information acquisition method, electronic device and readable storage medium storing program for executing | |
| CN114858806A (en) | Cigarette quality inspection data analysis system and method | |
| CN110992361A (en) | Engine fastener detection system and detection method based on cost balance | |
| CN113408630A (en) | Transformer substation indicator lamp state identification method | |
| CN113825162B (en) | Method and device for positioning fault reasons of telecommunication network | |
| CN111047731A (en) | AR technology-based telecommunication room inspection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |