Nuclear power valve drawing information recognition and collection method and system

Abstract

The present application provides a kind of nuclear power valve drawing information recognition acquisition method and system, acquisition method includes: obtaining target valve drawing, and the target valve drawing is carried out geometric requirement and pre-processing on format, obtain first target valve drawing;Valve drawing information recognition model of CRNN is used to the information of first target valve drawing is pre-identified, after identification is completed, in turn, drawing image information recognition removal, drawing pixel channel is merged, drawing information filtering processing and threshold processing, obtain second target valve drawing;Valve drawing information recognition model of CRNN is used to the information of the to-be-identified area of second target valve drawing is identified;Valve drawing information recognition model of CRNN is used to the information of the identification result in step 3 is extracted and collected, and the information obtained by collection is classified and integrated;The drawing of classified integration is evaluated correctly, then correct and effective information is operated into valve drawing information database.

Description

Technical Field

[0001] This invention relates to the field of nuclear power valve design technology, specifically to a method and system for identifying and collecting information from nuclear power valve drawings. Background Technology

[0002] Nuclear power plant valves are widely used in the reactor primary loop system. Currently, there are problems such as a large number of valves, a large number of design drawings submitted by suppliers, and a heavy review task, which makes it impossible to save the relevant information in a timely manner.

[0003] Moreover, nuclear power valves involve many types, such as gate valves, globe valves, ball valves, butterfly valves, check valves, and safety valves. The design information of different valve drawing types varies. In addition, different nuclear power projects have different valve suppliers, and their design drawings also differ greatly in terms of information expression style, text font, and legend annotation.

[0004] Valves, as a crucial component of nuclear power plant piping systems, can serve as critical equipment in the system's functional execution. Therefore, the optimization of nuclear power plant valve design directly impacts the overall design optimization of the nuclear power system. Important design parameters contained in nuclear power plant valve drawings must be properly preserved; otherwise, data resources for valve design will be wasted.

[0005] It can be seen that the information contained in nuclear power plant valve drawings can be roughly divided into: general assembly drawing information, design requirement information, bill of materials information, technical parameter information, detailed design information, and other information. Among them, the general assembly drawing and detailed design information are graphic information, while the rest are parameter information. Parameter information is important for nuclear power plant valve design and can be further divided into textual information (such as design requirements and standards), numerical information (such as the number of components), and character information (such as the material standard name).

[0006] Current drawing information acquisition systems or methods can identify design parameter information in drawings, but they still have the following problems or shortcomings:

[0007] I. Existing solutions are mainly based on industrial valve drawings and lack specificity and applicability for identifying information parameters in nuclear power valve drawings. They are difficult to solve practical engineering problems of nuclear power valve equipment and cannot meet the special requirements for identifying parameter information in nuclear power valve drawings.

[0008] Second, the existing solutions only focus on using novel and cutting-edge drawing parameter recognition methods to identify the model's modeling capabilities, thereby improving the model's recognition ability. However, in reality, the model needs to be trained based on the drawings, so the drawings used for model training will affect the model's actual capabilities. It is evident that the existing solutions lack self-learning and self-enhancement methods, and the actual capabilities of the model cannot be fundamentally updated, enhanced, and iteratively upgraded.

[0009] Third, existing solutions mainly focus on the collection and identification of drawing information, and stop at this process. They lack the work of classifying, storing and managing data information, which makes it impossible to truly make effective use of valve design parameter resources and wastes drawing parameter data resources. Furthermore, they fail to judge the correctness of important information in the drawings and cannot assess the effectiveness and credibility of the drawing information.

[0010] Fourth, the existing solutions employ relatively outdated machine learning models, whose computational analysis and recognition capabilities are insufficient, and they lack effectiveness in identifying key elements in nuclear power plant valve drawings. Traditional OCR (Optical Character Recognition) methods mostly segment text into individual characters before performing character recognition, which has poor recognition capabilities for text sequences of variable length, thus limiting their practicality for image sequence recognition problems.

[0011] Based on this, the inventors of this application propose a method and system for identifying and collecting information from nuclear power plant valve drawings, in order to solve one or more of the aforementioned technical problems. Summary of the Invention

[0012] The technical problem to be solved by the present invention is to overcome the shortcomings of low accuracy and utilization rate of valve drawing parameter identification in the prior art, and to provide a method and system for identifying and collecting nuclear power valve drawing information.

[0013] The present invention solves the above-mentioned technical problems through the following technical solution:

[0014] This invention provides a method for identifying and collecting nuclear power plant valve drawing information through a first aspect, comprising:

[0015] Step 1: Obtain the target valve drawing and preprocess the target valve drawing according to geometric requirements and format to obtain the first target valve drawing;

[0016] Step 2: Use the valve drawing information recognition model of CRNN to pre-recognize the first target valve drawing information. After recognition, perform drawing image information recognition and removal, drawing pixel channel merging, drawing information filtering and thresholding in sequence to obtain the second target valve drawing.

[0017] Step 3: Use the valve drawing information recognition model of the CRNN to identify the information of the area to be identified in the second target valve drawing;

[0018] Step 4: Use the valve drawing information recognition model of the CRNN to extract and collect the information from the recognition results in Step 3, and classify and integrate the collected information;

[0019] Step 5: Evaluate the correctness of the categorized and integrated drawings, and then input the correct and valid information into the valve drawing information database; wherein, the valve drawing information database is associated with the valve drawing information recognition model of the CRNN.

[0020] According to an embodiment of the present invention, step 1 includes:

[0021] Step 11: Use OpenCV's imread() function to read the target valve drawing;

[0022] Step 12: Check and judge the format of the target valve drawing, and convert the non-JPG format target valve drawing to JPG format;

[0023] Step 13: Perform drawing angle checks and drawing pixel height transformations based on pixel coordinates through pixel operations to obtain the first target valve drawing.

[0024] According to an embodiment of the present invention, step 2 includes:

[0025] Step 21: Receive the first target valve drawing, and perform pixel inspection and key valve identification on the first target valve drawing information to obtain a third target valve drawing without abnormal pixels;

[0026] Step 22: Perform image information recognition on the third target valve drawing;

[0027] Step 23: After image recognition is completed, use OpenCV's merge() method to merge the three channels of the third target valve drawing file to obtain the fourth target valve drawing.

[0028] Step 24: Perform bilateral filtering on the merged fourth target valve drawing;

[0029] Step 25: Apply pixel binarization thresholding to the filtered drawing to make the pixel information in the drawing stand out.

[0030] According to an embodiment of the present invention, in step 21, by obtaining the coordinate pixels of the first target valve drawing, a valve drawing information recognition model based on the CRNN is introduced into the abnormal pixel area between 0 and 255 to pre-identify key information.

[0031] If no abnormal pixels are found, proceed to step 22; if abnormal pixels are found, use the valve drawing information recognition model of the CRNN to identify key information. For drawings that fail to identify key information but belong to key valves, report an error to the valve drawing information database. For drawings that fail to identify key information but do not belong to key valves, report the corresponding key information as invalid.

[0032] According to an embodiment of the present invention, in step 22, a CNN-based valve drawing image recognition model is used to remove image information in the third target valve drawing that does not belong to the image information to be identified and collected; wherein, the valve assembly drawing and detailed design detail image information in the third target valve drawing are image information that does not belong to the image information to be identified and collected.

[0033] According to an embodiment of the present invention, in step 3, the valve drawing information recognition model of the CRNN includes a CRNN convolutional layer, a CRNN recurrent layer and a CRNN transcription layer. The CRNN convolutional layer is used to extract text features of the information to be recognized, the CRNN recurrent layer is used to extract sequence features and generate a probability feature sequence, and the CRNN transcription layer is used to predict the probability feature sequence and convert it into a character sequence.

[0034] According to an embodiment of the present invention, step 4 includes:

[0035] Step 41: Receive the identified result information from Step 3, and extract the drawing information using the end-to-end identification method of the valve drawing information identification model of the CRNN;

[0036] Step 42: Classify and integrate the extracted information.

[0037] According to an embodiment of the present invention, step 5 includes:

[0038] Step 51: Receive the drawing information that has been classified and integrated in step 4, compare the drawing information with the design key information database composed of valve design documents, find the error information through comparison and correct the error information.

[0039] Step 52: Organize the correctly verified information into the valve drawing information database; wherein, the valve drawing information database is associated with the valve drawing information recognition model of the CRNN.

[0040] A computer program product according to a second aspect of this application includes a computer program that, when executed by a processor, implements the steps performed by a computer as described in the first aspect.

[0041] According to a third aspect of this application, a computer-readable storage medium has a computer program that, when executed by a processor, implements the steps performed by a computer as described in the first aspect.

[0042] A data acquisition system according to a fourth aspect of this application includes: a memory capable of storing instructions executable by a processor; and a processor capable of executing the instructions to perform the steps performed by a computer as described in the first aspect of the method.

[0043] The positive and progressive effects of this invention are as follows:

[0044] This invention provides a method for identifying and collecting information from nuclear power plant valve drawings. Through drawing processing, information identification, information collection and classification integration, and finally data storage, it can efficiently complete the identification and collection of information from nuclear power plant valve drawings. It can also reduce the cost of nuclear power plant valve drawing review, system design optimization and other work. At the same time, it can carry out model self-reinforcement and self-upgrading for the identification model, so that the model can maintain high-efficiency identification capability, and provide support for nuclear power plant system equipment design optimization, drawing review, data verification and other work. Attached Figure Description

[0045] The above and other features, properties and advantages of the present invention will become more apparent from the following description taken in conjunction with the accompanying drawings and embodiments, wherein:

[0046] Figure 1 This is a flowchart of one embodiment of the nuclear power plant valve drawing information identification and acquisition method of the present invention;

[0047] Figure 2 This is a flowchart of the preprocessing of nuclear power valve drawings for this invention;

[0048] Figure 3 A schematic diagram for pixel inspection and rotation geometry inspection of the drawing;

[0049] Figure 4 Flowchart for judging nuclear power plant valve drawing information;

[0050] Figure 5 Flowchart for identifying nuclear power plant valve drawing information;

[0051] Figure 6 Flowchart for collecting drawing information for nuclear power plant valves;

[0052] Figure 7 Flowchart for evaluating nuclear power plant valve drawing information;

[0053] Figure 8 This is a schematic diagram of the data acquisition system of the present invention. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Embodiments of the invention will now be described in detail with reference to the accompanying drawings. Preferred embodiments of the invention will now be described in detail, examples of which are illustrated in the drawings. Wherever possible, the same reference numerals will be used in all the drawings to denote the same or similar parts. Furthermore, although the terminology used herein is selected from commonly known and used terminology, some terms mentioned in this specification may have been chosen by the applicant at his or her discretion, and their detailed meanings are explained in the relevant sections of the description herein. Moreover, the invention should be understood not only by the actual terms used, but also by the meaning implied by each term.

[0056] This invention proposes a method for identifying and collecting information from nuclear power plant valve drawings, which can be referred to as... Figure 1 , Figure 1 The process of nuclear power plant valve drawing information recognition and acquisition is shown. The nuclear power plant valve drawings are processed sequentially from input to preprocessing, judgment, recognition, acquisition and evaluation. The process supports the training of CRNN valve drawing information recognition model, evaluates and corrects the obtained drawing information, and the classified and integrated output information files can also be used to strengthen model training, ensuring the adaptability and applicability of model self-learning and self-reinforcement.

[0057] Specifically, the data collection method includes the following steps:

[0058] Step 1: Obtain the target valve drawing and preprocess the target valve drawing according to geometric requirements and format to obtain the first target valve drawing.

[0059] In step 1, the target valve drawing is first received, and then the target valve drawing is converted into a file that meets the geometric and format requirements.

[0060] Reference Figure 2 Preprocessing specifically includes:

[0061] Step 11: Use OpenCV's imread() function to read the target valve drawing;

[0062] Step 12: Check and judge the format of the target valve drawing, and convert the non-JPG format target valve drawing to JPG format; for example, convert the valve drawing file in "*.pdf" format to JPG format.

[0063] Step 13: Perform pixel operations to check the drawing angle based on pixel coordinates and transform the drawing pixel height to obtain the first target valve drawing.

[0064] Specifically, in step 13, NumPy is used to perform pixel operations on the drawing and obtain the vertical coordinates of the first column of the black border on the left side of the drawing. For example... Figure 3 As shown, if all coordinates yield a pixel value of 0 (three-channel pixel values ​​B=G=R=0), indicating a pure black pixel, then the drawing is currently in a vertical position and no rotation geometry operation is needed. Otherwise, obtain the pixel coordinates with a value of 0 in the right column. The angle between the two pixel coordinates indicates the drawing is currently skewed. Calculate the angle between the pixel coordinates and the vertical direction, and use OpenCV's getRotationMatrix2D() method to rotate the drawing with the bottom left corner as the rotation coordinate center and the calculated deflection angle as the rotation angle.

[0065] Then, the OpenCV size() method is used to obtain the pixel values ​​of the drawing width and height. Since the CRNN model used in the subsequent drawing information recognition module requires the geometric pixel height of the drawing input to be fixed at 32, the OpenCV dsize() method is also needed to perform geometric transformation on the input files that do not meet the pixel height requirement.

[0066] Step 2: Use the valve drawing information recognition model of CRNN to pre-recognize the first target valve drawing information. After recognition, perform drawing image information recognition and removal, drawing pixel channel merging, drawing information filtering and thresholding in sequence to obtain the second target valve drawing.

[0067] Reference Figure 4 Step 2 includes:

[0068] Step 21: Receive the first target valve drawing and perform pixel checks and key valve identification on the first target valve drawing information to obtain the third target valve drawing without abnormal pixels.

[0069] In step 21, by obtaining the coordinate pixels of the first target valve drawing, a valve drawing information recognition model based on CRNN is introduced to pre-identify key information for abnormal pixel areas between 0 and 255.

[0070] If no abnormal pixels are found, proceed to step 22; if abnormal pixels are found, perform a key information check on the drawing and use the CRNN valve drawing information recognition model to identify the key information. For drawings where key information identification fails but the drawing belongs to a key valve, an error message is sent back to the valve drawing information database. For drawings where key information identification fails but the drawing does not belong to a key valve, the corresponding key information is invalidated. Figure 2 (The process in progress has ended).

[0071] Step 22: Perform image information recognition on the third target valve drawing.

[0072] Because this application targets the recognition and acquisition of character parameter information in drawings, image information such as valve assembly drawings and detailed design details in the drawings is not part of the information to be recognized and acquired, and will interfere with the recognition and acquisition of character parameter information, or even cause information recognition failure. Therefore, this application adopts a CNN (Convolutional Neural Network) valve drawing image recognition model, uses existing valve organization image data for model training, and uses it for image recognition in this step. After recognition, the positions of the valve assembly drawing and detailed design details images are selected and located, and NumPy pixel operations are used to set the image pixel value within the location box to 255. That is, the area in that location becomes pure white to remove the corresponding image information. This processing will not interfere with the recognition and acquisition of character parameter information.

[0073] Step 23: After image recognition is completed, use OpenCV's merge() method to merge the three channels of the third target valve drawing file to obtain the fourth target valve drawing.

[0074] In other words, after the drawing image information is identified and removed, the drawing file does not need to use the three-way RGB (red, green and blue optical primary colors) channels. This step uses OpenCV's merge() method to merge the three-way channels of the drawing file in order to facilitate subsequent drawing pixel thresholding.

[0075] Step 24: Perform bilateral filtering on the merged fourth target valve drawing.

[0076] In other words, the drawing file after pixel channel merging undergoes bilateral filtering to remove noise and reduce detail layer information while effectively preserving image boundary information, thus smoothing the image. Specifically, the OpenCV `bilaterFilter()` method is used, with the pixel centroid located at the drawing centroid pixel coordinates, and the color calculation range set to 255.

[0077] Step 25: Apply pixel binarization thresholding to the filtered drawing to make the pixel information in the drawing stand out.

[0078] In other words, after filtering, the pixel binarization thresholding of the drawing file can make the pixel information in the drawing stand out, forming a strong pixel contrast and distinction effect. Highlighting the pixel information of the drawing helps to improve the recognition effect in subsequent information recognition.

[0079] Optionally, this application uses OpenCV's threshold() method, with a pixel value of 200 as the dividing line and 255 as the maximum value. Pixels with a value less than or equal to 200 are set to 0 and displayed as pure black, while pixels with a value greater than 200 are set to 255 and displayed as pure white.

[0080]

[0081] After pixel binarization thresholding of the drawing, the pixel values ​​of the area to be identified in the drawing show obvious differentiation, which will effectively improve the recognition effect of the drawing information.

[0082] Step 3: Use the valve drawing information recognition model of CRNN to identify the information of the area to be identified in the second target valve drawing.

[0083] Reference Figure 5 The valve drawing information recognition model of CRNN is divided into CRNN convolutional layer, CRNN recurrent layer and CRNN transcription layer, which respectively complete the extraction of text features of the information to be recognized, the extraction of sequence features and generation of probability feature sequence, the prediction of probability feature sequence and conversion into character sequence.

[0084] Specifically, the CRNN convolutional layer is used to extract the features to be identified from the input drawing. After convolution, pooling, and normalization operations, a convolutional feature map is obtained, which serves as the input to the CRNN recurrent layer. The CRNN convolutional layer uses a convolutional neural network as the core tool for feature extraction, constructing 6 convolution operations, 4 max pooling operations, and performing 2 batch normalization operations, for a total of 13 convolutional layer operations. The relevant operation parameters are shown in Table 1, where C is the number of channels, k is the kernel size, s is the kernel stride, p is the padding size, w is the pooling kernel size, and the activation function is the rectified linear function ReLU (Rectified Linear Unit).

[0085] Table 1 CRNN Convolutional Layer Operations

[0086] 1 convolution C:64,k:3×3,s:1,p:1 ReLU 2 Max pooling w:2×2,s:2 - 3 convolution C:128,k:3×3,s:1,p:1 ReLU 4 Max pooling w:2×2,s:2 - 5 convolution C:256,k:3×3,s:1,p:1 ReLU 6 convolution C:256,k:3×3,s:1,p:1 ReLU 7 Max pooling w:1×2,s:2 - 8 convolution C:512,k:3×3,s:1,p:1 ReLU 9 Batch normalization - - 10 convolution C:512,k:3×3,s:1,p:1 ReLU 11 Batch normalization - - 12 Max pooling w:1×2,s:2 - 13 convolution C:512,k:2×2,s:1,p:0 ReLU

[0087] As we can see, the purpose of convolution is feature detection. The input matrix of the convolution operation represents the image region to be identified, and the matrix elements correspond to image pixels, representing their grayscale information. The convolution kernel slides across the image with a stride, and convolution actually performs matrix inner product calculations. After convolution calculation, a feature map is obtained. Figure 1 It is usually a matrix or a tensor.

[0088] Pooling, also known as convergence or downsampling, is an operation for feature selection. Each element of the input matrix in the pooling operation identifies a feature, representing the detected value of that feature. The pooling kernel slides across the image with a stride, taking the maximum value from the matrix parameters covered by the sliding kernel. Pooling outputs a reduced matrix.

[0089] Batch normalization normalizes the net input of each layer in a neural network for each batch of samples. This operation performs a feature scaling transformation, which can accelerate the learning and convergence speed of the neural network and, to some extent, prevent gradient vanishing and gradient explosion. Gradient vanishing and gradient explosion can occur in the gradient descent calculation of neural networks, leading to parameter cessation of updates and parameter overflow, which in turn makes model learning difficult and computational convergence failure. Normalizing the input values ​​to keep them within a certain range, such as between 0 and 1, can reduce the difference in the value range, accelerate learning convergence, and suppress gradient vanishing and gradient explosion.

[0090] Activation functions are a crucial component of neural networks, used to establish nonlinear relationships between inputs and outputs, thereby enhancing the learning and expressive capabilities of the neural network. Real-world problems are generally nonlinear, and activation functions effectively express these nonlinear relationships. Commonly used activation functions include the sigmoid function, hyperbolic tangent function, and rectified linear function (ReLU). This application employs the rectified linear function ReLU. ReLU's derivative is always 1 in the positive interval, avoiding the gradient explosion problem. Furthermore, this function is computationally simpler than other activation functions and maintains good sparse activation performance, improving the model's generalization ability. The functional relationship is shown in the following equation.

[0091] f(x) = relu(x) = max(0,x);

[0092] The information to be identified undergoes multiple convolution, pooling, and normalization operations to extract text features from the drawing information, resulting in a feature map, which serves as the input to the CRNN recurrent layer.

[0093] The CRNN recurrent layer employs bidirectional LSTM modules to form a Bidirectional Long Short-Term Memory Neural Network (BiLSTM). This network extracts textual features from the CRNN convolutional layer input as sequence features and generates a probabilistic feature sequence. BiLSTM is based on LSTM, which, in turn, introduces special gate structures to control the transmission of input information, allowing the network model to retain or discard information. The LSTM input gate is used to update the current unit's information state; the forget gate determines whether the current information is retained or discarded; and the output gate controls the information output by the current unit and passes information to the next LSTM unit.

[0094] When processing information, LSTM adopts a context-to-context order, which cannot fully utilize contextual information in text recognition. BiLSTM breaks the limitations of unidirectional LSTM by using two LSTMs in opposite directions. The two LSTMs work independently without information interaction and finally merge the calculation results for output. This allows for the simultaneous acquisition and processing of contextual information when processing drawing information, resulting in more accurate and reliable prediction results compared to LSTM.

[0095] The CRNN recurrent layer uses BiLSTM to generate probabilistic feature sequences, which are then used as input to the CRNN transcription layer.

[0096] The CRNN transcription layer uses the Connectionist Temporal Classification (CTC) loss function to convert the predicted probability feature sequence input to the CRNN recurrent layer into a character sequence. Essentially, it finds the character sequence with the highest probability combination in the probability feature matrix as the final information recognition result.

[0097] In text recognition, the predicted sequence has no fixed length, and the number of characters corresponding to different features varies. Traditional mapping function methods cannot solve the problem of variable-length text recognition, and variable sequence length text recognition is one of the challenges in designing text recognition models. Furthermore, the recurrent layer processing may introduce blank positions or repeated character information, leading to distorted recognition results and reduced model recognition ability. CRNN uses the CTC loss function, which allows for end-to-end training, omitting intermediate processes and focusing only on the input and output of data, effectively improving recognition efficiency. Simultaneously, the CTC loss function introduces a whitespace "-" mechanism, inserting whitespace "-" between repeated characters to separate redundant characters. Therefore, when the CRNN transcription layer converts the probabilistic feature sequence into a character sequence, it does not merge identical and consecutively occurring characters in the existing character sequence, but instead first deletes repeated characters and whitespace "-" positions, effectively solving the problem of character repetition and redundancy in text information recognition. This process is also called the decoding process.

[0098] The CRNN transcription layer removes duplicates from the probability distribution sequence of the recurrent layer and extracts the character with the highest probability to form the final recognition result.

[0099] Step 4: Use the CRNN valve drawing information recognition model to extract and collect the information from the recognition results in Step 3, and classify and integrate the collected information.

[0100] Reference Figure 6 Step 4 includes:

[0101] Step 41: Receive the identified result information from Step 3, and extract the drawing information using the end-to-end recognition method of the CRNN valve drawing information recognition model.

[0102] Specifically, this application adopts a key feature matching mechanism. Keywords are derived from valve parameter information required to be stored in the valve drawing information database. The units of the parameter information are used as identifiable key features, such as the units of temperature, pressure, and weight. These features are more significant than other text information, which helps to accurately obtain the correct parameter information. The key feature information is used as a key screening indicator to compare with the currently identified information. Information that matches successfully is retained as extracted and collected information, while invalid information is discarded and returned for information extraction, until all key features have been used.

[0103] Step 42: Classify and integrate the extracted information.

[0104] Specifically, the drawing information obtained by the drawing information classification and integration submodule is classified into major categories such as drawing source information and design parameter information. At the same time, a temporary evaluation file is generated to store the information integration results of the current submodule. This result file will be used by the information evaluation module to judge the correctness.

[0105] Step 5: Evaluate the correctness of the categorized and integrated drawings, and then input the correct and valid information into the valve drawing information database; the valve drawing information database is associated with the valve drawing information recognition model of CRNN.

[0106] Reference Figure 7 Step 5 includes:

[0107] Step 51: Receive the drawing information that has been classified and integrated in Step 4, compare the drawing information with the key design information database composed of valve design documents, find the error information through comparison, and correct the error information.

[0108] It is understood that the comparison parameters used in this application are derived from the design key information database composed of valve design documents, which stores correct data of key information and can be used for parameter accuracy comparison. During parameter information judgment, if the judgment information is incorrect, the error information is fed back to the valve drawing information database for that data entry, and the information is rewritten to correct the error; if the judgment information is correct, the process proceeds to step 52.

[0109] Step 52: Organize the correctly verified information into the valve drawing information database; wherein, the valve drawing information database is associated with the valve drawing information recognition model of CRNN.

[0110] In other words, the main information required for data entry includes: drawing file ownership information (nuclear power project number, item number, submitting unit, data sheet number), valve design parameter information (nuclear safety classification, seismic resistance level, design temperature, design pressure, nominal size, valve weight rating, main body material, flow resistance parameters), valve envelope requirement information (valve weight, valve center of gravity), and valve test parameter information (shell hydrostatic test, valve seat sealing test, closure test, packing sealing test). For electric, pneumatic, or electromagnetic valves, relevant information on the electric actuator, pneumatic actuator, etc., is also required.

[0111] The resulting database information is not only correlated with the current drawings but also ensures the accuracy and validity of the identified data. Therefore, the valve drawing information database will be used for further training and reinforcement of the CRNN valve drawing information recognition model. Through self-learning and self-reinforcement, the CRNN valve drawing information recognition model, trained based on existing drawing information for judgment, recognition, collection, and evaluation, will better meet the information recognition and collection needs of nuclear power project drawings, and its model operation will be more targeted and reliable.

[0112] In summary, this application addresses the identification of nuclear power valve drawing information, meeting the requirements for nuclear power valve drawing information processing, including "drawing processing, information identification, information collection, data sorting, data storage, information error correction, and model enhancement," and efficiently completing the identification and collection of nuclear power valve drawing information.

[0113] Moreover, this application can be used to guide the self-learning and self-reinforcement of the nuclear power valve drawing information recognition model, thereby improving the model's capabilities and providing technical support for nuclear power plant valve selection, drawing review, equipment design, system optimization, data verification, and other work, thus reducing the cost of nuclear power valve drawing review, system design optimization, and other work.

[0114] refer to Figure 8 As shown, this application also provides a data acquisition system 1000, including a memory 1001 and a processor 1002. The memory is capable of storing instructions that can be executed by the processor 1002; the processor 1002 is capable of executing instructions to implement the steps performed by the computer in the nuclear power valve drawing information identification and acquisition method described in the above embodiments.

[0115] It is understood that the aforementioned memory and processor are not limited to a specific memory or processor. Furthermore, in embodiments employing a distributed architecture, the specific execution terminal for each step can be adjusted according to actual circumstances, and the specific implementation scheme of each step on a particular terminal should not limit the scope of protection of this application.

[0116] According to another aspect of this application, a computer-readable medium is also provided.

[0117] The computer-readable medium provided in this application has computer instructions thereon. When executed by a processor, these computer instructions can implement the steps performed by the program in the methods described in the above embodiments.

[0118] According to another aspect of this application, a computer program product is also provided.

[0119] The computer-readable medium provided in this application includes a computer program that, when executed by a processor, can implement the steps performed by the program in the method described in the above embodiments.

[0120] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or performed using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in alternatives, it may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.

[0121] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of both. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor such that the processor can read and write information to / from the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and storage medium may reside as discrete components in the user terminal.

[0122] In one or more exemplary embodiments, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functionality may be stored or transmitted as one or more instructions or code on or via a computer-readable medium. A computer-readable medium includes both computer storage media and communication media, encompassing any medium that facilitates the transfer of a computer program from one location to another. A storage medium may be any available medium accessible to a computer. By way of example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and is accessible to a computer. Any connection is also legitimately referred to as a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of a medium. As used in this article, disk and disc include compact discs (CDs), laser discs, optical discs, digital multi-purpose discs (DVDs), floppy disks, and Blu-ray discs. Disks typically reproduce data magnetically, while discs reproduce data optically using lasers. Combinations of these should also be included within the scope of computer-readable media.

[0123] While this application discloses preferred embodiments as described above, it is not intended to limit the scope of this application. Any changes and modifications can be made by those skilled in the art without departing from the spirit and scope of this application. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application without departing from the content of the technical solution of this application shall fall within the protection scope defined by the claims of this application.

Claims

1. A method for identifying and collecting information from nuclear power plant valve drawings, characterized in that, include: Step 1: Obtain the target valve drawing and preprocess the target valve drawing according to geometric requirements and format to obtain the first target valve drawing; Step 2: Use the valve drawing information recognition model of CRNN to pre-recognize the first target valve drawing information. After recognition, perform drawing image information recognition and removal, drawing pixel channel merging, drawing information filtering and thresholding in sequence to obtain the second target valve drawing. Step 2 includes: Step 21: Receive the first target valve drawing, and perform pixel inspection and key valve identification on the first target valve drawing information to obtain a third target valve drawing without abnormal pixels; In step 21, by obtaining the coordinate pixels of the first target valve drawing, a valve drawing information recognition model based on the CRNN is introduced to pre-identify key information for abnormal pixel areas between 0 and 255. If there are no abnormal pixels, proceed to step 22; if there are abnormal pixels, use the valve drawing information recognition model of the CRNN to identify key information. For drawings that fail to identify key information but belong to key valves, report error information to the valve drawing information database. For drawings that fail to identify key information but do not belong to key valves, report the corresponding key information as invalid. Step 22: Perform image information recognition on the third target valve drawing; In step 22, a CNN-based valve drawing image recognition model is used to remove image information from the third target valve drawing that does not belong to the image information to be identified and collected; wherein, the valve assembly drawing and detailed design detail image information in the third target valve drawing are image information that does not belong to the image information to be identified and collected. Step 23: After image recognition is completed, use OpenCV's merge() method to merge the three channels of the third target valve drawing file to obtain the fourth target valve drawing. Step 24: Perform bilateral filtering on the merged fourth target valve drawing; Step 25: Apply a pixel binarization threshold to the filtered drawing to make the pixel information in the drawing stand out; Step 3: Use the valve drawing information recognition model of the CRNN to identify the information of the area to be identified in the second target valve drawing; Step 4: Use the valve drawing information recognition model of the CRNN to extract and collect the information from the recognition results in Step 3, and classify and integrate the collected information; Step 5: Evaluate the correctness of the categorized and integrated drawings, and then input the correct and valid information into the valve drawing information database; wherein, the valve drawing information database is associated with the valve drawing information recognition model of the CRNN.

2. The method for identifying and collecting nuclear power plant valve drawing information according to claim 1, characterized in that, Step 1 includes: Step 11: Use OpenCV's imread() function to read the target valve drawing; Step 12: Check and judge the format of the target valve drawing, and convert the non-JPG format target valve drawing to JPG format; Step 13: Perform drawing angle checks and drawing pixel height transformations based on pixel coordinates through pixel operations to obtain the first target valve drawing.

3. The method for identifying and collecting nuclear power plant valve drawing information according to claim 1, characterized in that, In step 3, the valve drawing information recognition model of the CRNN includes a CRNN convolutional layer, a CRNN recurrent layer, and a CRNN transcription layer. The CRNN convolutional layer is used to extract text features of the information to be recognized. The CRNN recurrent layer is used to extract sequence features and generate a probability feature sequence. The CRNN transcription layer is used to predict the probability feature sequence and convert it into a character sequence.

4. The method for identifying and collecting nuclear power plant valve drawing information according to claim 1, characterized in that, Step 4 includes: Step 41: Receive the identified result information from Step 3, and extract the drawing information using the end-to-end identification method of the valve drawing information identification model of the CRNN; Step 42: Classify and integrate the extracted information.

5. The method for identifying and collecting nuclear power plant valve drawing information according to claim 1, characterized in that, Step 5 includes: Step 51: Receive the drawing information that has been classified and integrated in step 4, compare the drawing information with the design key information database composed of valve design documents, find the error information through comparison and correct the error information. Step 52: Organize the correctly verified information into the valve drawing information database; wherein, the valve drawing information database is associated with the valve drawing information recognition model of the CRNN.

6. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the steps performed by a computer in the method as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, The method comprises a computer program that, when executed by a processor, implements the steps performed by a computer in the method as described in any one of claims 1-5.

8. A data acquisition system, characterized in that, include: Memory is capable of storing instructions that can be executed by a processor; A processor capable of executing the instructions to implement the steps performed by a computer in the method as described in any one of claims 1-5.

Images