High-temperature plate blank jet printing character recognition system and method

Abstract

The invention provides a high-temperature plate blank jet printing character recognition system and method. The high-temperature plate blank jet printing character recognition system comprises an operation terminal, a roller way control system, a process control system, a communication module, an industrial camera, an image recognition module and an image processing module. All the modules are connected through a network to form a service function system, and the industrial camera is installed above a plate blank conveying roller way and used for capturing images of plate blank jet printing characters; the image processing module uses a linear expansion algorithm to carry out strong light suppression processing on the captured image; and the image recognition module performs character recognition on the processed image by adopting an OCR (Optical Character Recognition) character recognition algorithm. The method effectively eliminates the red light influence caused by a high-temperature plate blank through stretching the interval between the low-gray-scale region and the high-gray-scale region in the image and compressing the medium-gray-scale region, thereby remarkably improving the accuracy of character recognition, enabling the character recognition technology to work stably in an industrial scene with high-temperature red light interference, and improving the recognition efficiency. And the requirements of manufacturing industries such as iron and steel enterprises on high-precision character recognition are met.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing and recognition, and in particular to a system and method for high-temperature slab printing character recognition. Background Art

[0002] After years of development, traditional character recognition technology has been perfected, especially in fixed image recognition scenarios under stable lighting conditions. However, in the high-temperature slab production environment, the red light interference generated by the high temperature of the newly produced slab makes the slab number image captured by the camera blurred, which seriously affects the recognition accuracy of the slab number. Therefore, solving this problem is crucial to expand the application of image recognition technology in red light interference industrial scenarios such as high-temperature slab production.

[0003] With the transformation and upgrading of China's manufacturing industry, the demand for high-temperature slab character recognition technology is growing. China's steel companies are seeking to improve production efficiency and product quality through automation and intelligent technology, thereby reducing costs and improving market competitiveness. Domestic research institutions and universities are actively conducting research on related technologies, and developing character recognition systems that can work stably in high-temperature environments through machine learning, deep learning, image processing and other technologies.

[0004] Globally, high temperature slab character recognition technology is an important research direction in the field of intelligent manufacturing and industrial automation. Developed countries such as the United States, Germany and Japan have relatively mature technologies and products in this field. Some large international steel companies and automation equipment suppliers have applied high temperature slab character recognition technology to production practice, and have achieved optimization of the production process and improvement of quality through real-time monitoring and data analysis.

[0005] The patent document with publication number CN117688162A discloses a full-text retrieval method and system based on OCR recognition, which converts unstructured data into searchable text data, realizes automatic entry and automatic proofreading of files to be stored, completes comprehensive retrieval of text content, helps users find the required information quickly and accurately, and improves the efficiency and accuracy of information retrieval. This invention recognizes traditional text and is not suitable for industrial scenarios.

[0006] In the patent document with the publication number CN116824225A, an improved handwritten slab number recognition method based on the attention mechanism is disclosed. The handwritten slab number image data is preprocessed in combination with complex production scenarios to ensure the quality of the model input image; the residual neural network module is used to extract image features; the bidirectional long short-term memory network extracts image-based sequence features; finally, the information flow in the sequence is captured based on the attention mechanism, and the features of each character are integrated to form a text feature vector to predict the output sequence. After field testing, the accuracy rate of the handwritten slab number recognition task on the surface of the steel plate reached 93.49%. The results show that the algorithm is feasible and effective and meets the actual production needs. The recognition situation of this invention can realize handwriting recognition, but it cannot recognize high-temperature slabs, and the recognition rate is less than 95%.

[0007] A deep learning-based slab number recognition method is disclosed in the patent document with publication number CN116052151A, which includes the following steps: obtaining a video recording of the on-site slab moving process, intercepting the video recording frame by frame to obtain a large number of images with slab numbers; marking the slab number on each slab image, and making a data set required for slab number recognition model training; training and debugging the recognition model; deploying the recognition model on-site, each time recognizing a section of the video of the entire rail car passing by, and finally processing the matching results, and obtaining the number of slabs and the corresponding slab numbers. This invention is based on deep learning for recognition, and does not realize high-temperature slab recognition.

[0008] Therefore, it is necessary to propose a new technical solution to improve the above technical problems. Summary of the invention

[0009] In view of the defects in the prior art, the object of the present invention is to provide a high-temperature slab printing character recognition system and method.

[0010] A high-temperature slab printing character recognition system provided by the present invention comprises: an operation terminal, a roller control system, a process control system, a communication module, an industrial camera, an image recognition module and an image processing module; the operation terminal, the roller control system, the process control system, the communication module, the industrial camera, the image recognition module and the image processing module are connected through a network to form a business function system, wherein:

[0011] The industrial camera is installed above the slab transport roller and is used to capture the image of the characters printed on the slab;

[0012] The image processing module uses a linear expansion algorithm to perform strong light suppression processing on the captured image;

[0013] The image recognition module uses an OCR character recognition algorithm to perform character recognition on the processed image.

[0014] Preferably, the linear expansion algorithm expands the dynamic range of the image by stretching the interval between low-grayscale and high-grayscale areas in the image and compressing the medium-grayscale area; uses different shutter speeds to capture multiple images, and extracts detail information at the same position from the images to synthesize new images.

[0015] Preferably, the method for synthesizing a new image is to linearly weight the grayscale values ​​of the two images to form a new grayscale value gf(x,y):

[0016] gf(x,y)=a 1 g 1 (x,y)+a 2 g 2 (x,y)

[0017] Among them, a 1 and a 2 is the weight, which is calculated based on the grayscale value of the image and the preset boundary conditions.

[0018] Preferably, the weight a 1 and a 2 The calculation is performed according to the following formula:

[0019] When [g 1 (x,y)+g 2 (x,y)] / 2<=T min When 1 =[g 1 (x,y)+g 2 (x,y)] / (4T min ), a 2 =1-a 1 ;

[0020] When T min <=[g 1 (x,y)+g 2 (x,y)] / 2<=T max When 1 =a 2 =0.5;

[0021] When [g 1 (x,y)+g 2 (x,y)] / 2>=T max When 2 =[256-(g 1 (x,y)+g 2 (x,y)) / 2] / 2(256-T max ), a 1 =1-a 2 ;

[0022] Wherein, Tmin and Tmax are preset grayscale thresholds.

[0023] Preferably, the image recognition module performs character recognition on the processed image through an OCR character recognition algorithm, and feeds back the recognition result to the process control system through the communication module.

[0024] Preferably, the process control system is used to receive the slab arrival information sent by the roller control system, and after determining the slab arrival signal, send down an identification signal, and receive the recognition result of the image recognition module, compare and determine it with the data in the actual production performance, and store the recognition result and determination result in the database.

[0025] Preferably, the operation terminal is used to connect to the process control system database, obtain the slab printing character recognition results, judgment results and statistical analysis data, and display them through the application screen.

[0026] The present invention also provides a method for identifying characters printed on a high-temperature slab. The method uses the above-mentioned system for identifying characters printed on a high-temperature slab. The method comprises the following steps:

[0027] Step S1: installing an industrial camera above the slab transport roller to capture images of characters printed on the slab;

[0028] Step S2: After the roller control system moves the slab to the fixed identification position, it sends the in-position information to the process control system;

[0029] Step S3: After the process control system determines that the slab is in place, it sends a recognition signal and calls the image processing module to process the captured image;

[0030] Step S4: the image processing module uses a linear expansion algorithm to perform strong light suppression processing on the image to eliminate the red light effect caused by the high-temperature slab;

[0031] Step S5: The image recognition module uses an OCR character recognition algorithm to perform character recognition on the processed image and feeds back the recognition result to the process control system.

[0032] Preferably, the processing steps of the linear expansion algorithm in step S4 include: capturing multiple images using different shutter speeds, extracting detail information at the same position, and calculating weights based on the grayscale values ​​of the images and preset boundary conditions, linearly weighting the grayscale values ​​of the two images, and synthesizing a new image.

[0033] Preferably, after receiving the recognition result in step S5, the process control system compares and determines it with the data in the actual production performance, and stores the recognition result and the determination result in a database for subsequent data statistics and result display.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] 1. The present invention effectively eliminates the red light influence caused by high-temperature slabs by stretching the interval between low-grayscale and high-grayscale areas in the image and compressing the medium-grayscale areas, thereby significantly improving the accuracy of character recognition. This enables the character recognition technology to work stably in industrial scenarios with high-temperature red light interference, meeting the needs of manufacturing industries such as steel companies for high-precision character recognition.

[0036] 2. The present invention realizes automatic identification of slab numbers. Through industrial camera capture, image processing, character recognition and other steps, the system can quickly and accurately identify the slab number, and compare the identification result with the data in the actual production performance, thereby reducing manual intervention, improving production efficiency and reducing production costs;

[0037] 3. The present invention is not only suitable for high-temperature slab production environments, but can also be extended to other industrial scenarios where strong light interference exists. By adjusting parameters such as the weights of the linear expansion algorithm, the system can adapt to image recognition requirements under different lighting conditions and has strong adaptability and flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0039] Figure 1 It is a system module connection diagram of the present invention;

[0040] Figure 2 This is a flow chart of the linear expansion algorithm implementation of the present invention. DETAILED DESCRIPTION

[0041] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0042] Embodiment 1:

[0043] Reference Figure 1 and Figure 2According to the present invention, a high-temperature slab printing character recognition system is provided, including: an operation terminal, a roller control system, a process control system, a communication module, an industrial camera, an image recognition module and an image processing module; the operation terminal, the roller control system, the process control system, the communication module, the industrial camera, the image recognition module and the image processing module are connected through a network to form a business function system, the industrial camera is installed above the slab transport roller, and is used to capture the image of the slab printing characters; the image processing module uses a linear expansion algorithm to suppress strong light on the captured image; the image recognition module uses an OCR character recognition algorithm to perform character recognition on the processed image.

[0044] The operation terminal is installed in the on-site operation room to allow on-site operators to check the recognition status. When the planned slab number is inconsistent with the on-site slab number, the operator is reminded to check. The roller control system is used to control the operation status of the roller. When the roller runs to the specified position, it sends an in-place signal to the process control system. The process control system is used to receive and issue production plans. When receiving the in-place signal from the roller control system, the process control system sends production information to the communication module and receives the recognition result from the communication module. The communication module is used to receive production information and send recognition results. After receiving the production information from the process control system, it sends a capture image signal. After completing the recognition, the recognition result is sent to the process control system. The industrial camera is installed above the slab transport roller to capture the image of the printed characters on the slab. After receiving the capture signal from the communication module, the image is captured; the image processing module uses a linear expansion algorithm to suppress the captured image, and after receiving the captured image from the industrial camera, it performs image preprocessing and sends it to the image recognition module; the image recognition module uses an OCR character recognition algorithm to perform character recognition on the processed image, and after receiving the preprocessed image, it recognizes and sends it to the communication module.

[0045] The linear expansion algorithm expands the dynamic range of the image by stretching the interval between low-grayscale and high-grayscale areas in the image and compressing the medium-grayscale area; it uses different shutter speeds to capture multiple images, extracts detail information at the same position from the images, and synthesizes a new image. The method of synthesizing a new image is to linearly weight the grayscale values ​​of the two images to form a new grayscale value gf(x,y):

[0046] gf(x,y)=a 1 g 1 (x,y)+a 2 g 2 (x,y)

[0047] Among them, a 1 and a 2 is the weight, which is calculated based on the gray value of the image and the preset boundary conditions.1 and a 2 The calculation is performed according to the following formula:

[0048] When [g 1 (x,y)+g 2 (x,y)] / 2<=T min When 1 =[g 1 (x,y)+g 2 (x,y)] / (4T min ), a 2 =1-a 1 ;

[0049] When T min <=[g 1 (x,y)+g 2 (x,y)] / 2<=T max When 1 =a 2 =0.5;

[0050] When [g 1 (x,y)+g 2 (x,y)] / 2>=T max When 2 =[256-(g 1 (x,y)+g 2 (x,y)) / 2] / 2(256-T max ), a 1 =1-a 2 ;

[0051] Wherein, Tmin and Tmax are preset grayscale thresholds.

[0052] The linear expansion algorithm expands the dynamic range of the image by stretching the interval between the low grayscale (setting the parameter grayscale value to 32) and high grayscale (setting the parameter grayscale value to 192) areas in the image and compressing the medium grayscale area (setting the parameter grayscale value interval to 32-192) to suppress strong light in the image.

[0053] The image recognition module uses the OCR character recognition algorithm to perform character recognition on the processed image, and feeds back the recognition results to the process control system through the communication module. The process control system is used to receive the slab in-place information sent by the roller control system, and after determining the slab in-place signal, it sends a recognition signal and receives the recognition results of the image recognition module, compares and determines with the data in the actual production performance, and stores the recognition results and determination results in the database. The operation terminal is used to connect to the process control system database, obtain the slab printing character recognition results, determination results and statistical analysis data, and display them through the application screen.

[0054] The OCR character recognition algorithm performs strong light suppression processing through the linear expansion algorithm in the image preprocessing stage, and then solves the high-temperature red light interference and performs OCR character recognition.

[0055] The present invention also provides a method for identifying characters printed on a high-temperature slab. The method uses the above-mentioned system for identifying characters printed on a high-temperature slab. The method comprises the following steps:

[0056] Step S1: installing an industrial camera above the slab transport roller to capture images of characters printed on the slab;

[0057] Step S2: After the roller control system moves the slab to the fixed identification position, it sends the in-position information to the process control system;

[0058] Step S3: After the process control system determines that the slab is in place, it sends a recognition signal and calls the image processing module to process the captured image;

[0059] Step S4: The image processing module uses a linear expansion algorithm to suppress strong light in the image to eliminate the red light effect caused by the high-temperature slab; the processing steps of the linear expansion algorithm include: capturing multiple images with different shutter speeds, extracting detail information at the same position, and calculating weights based on the grayscale value of the image and preset boundary conditions, linearly weighting the grayscale values ​​of the two images, and synthesizing a new image.

[0060] Step S5: The image recognition module uses the OCR character recognition algorithm to perform character recognition on the processed image and feeds back the recognition results to the process control system. After receiving the recognition results, the process control system compares and determines with the data in the actual production performance, and stores the recognition results and determination results in the database for subsequent data statistics and result display.

[0061] Embodiment 2:

[0062] The present invention relates to image processing and recognition in high-temperature scenes of industrial production, and is a system and method for high-temperature slab printing character recognition. In the scene of high-temperature red light interference, the image is processed by using a linear expansion algorithm to effectively eliminate the red light influence brought by the high-temperature slab, and combined with an OCR character recognition algorithm to ensure accurate recognition of the slab number.

[0063] In the scenario of high-temperature red light interference, the image is processed by using a linear expansion algorithm to effectively eliminate the red light influence brought by the high-temperature slab, and combined with the OCR character recognition algorithm to realize automatic recognition of the slab number, which is achieved through the following technical solutions.

[0064] The system consists of six parts: operation terminal, roller control system, process control system, communication module, industrial camera, image recognition module, and image processing module. Each part is connected through a network to form a business function system. For details, please refer to Figure 1 .

[0065] The method comprises the following steps:

[0066] S1. The industrial camera is installed above the slab transport roller. The camera angle is adjusted to align with the fixed recognition position until the screen can capture the slab printed characters, and the focal length is adjusted to enable clear imaging and capture;

[0067] S2, after the roller control system moves the slab to the fixed identification position, it sends the arrival information to the process control system;

[0068] S3, after the process control system determines the arrival signal of the slab, it sends the identification signal, which is received by the communication module and then processed by the image processing module;

[0069] S4, the image processing module calls the industrial camera interface to capture the image, and after acquiring the image, uses the linear expansion algorithm to suppress the strong light to eliminate the red light effect caused by the high-temperature slab. The processed image is sent to the image recognition module for recognition;

[0070] S5, the image recognition module performs character recognition on the processed image through the OCR character recognition algorithm, and the recognition result is fed back to the process control system via the communication module;

[0071] S6. After receiving the recognition result, the process control system compares and determines it with the data in the actual production performance, and stores the recognition result and determination result in the database for data statistics and result display.

[0072] S7. The operator or user connects to the process control system database through the operation terminal to obtain the slab printing character recognition results, judgment results, and statistical analysis data, and displays them through the application screen.

[0073] As a further improvement, in step S4, the linear expansion algorithm expands the dynamic range of the image by stretching the interval between low grayscale and high grayscale areas in the image and compressing the medium grayscale area. This process involves taking multiple images, capturing them with different shutter speeds, and extracting detail information at the same position from these images to synthesize a new image. The specific synthesis method is to linearly weight the grayscale values ​​of the two images to form gf(x,y):

[0074] gf(x,y)=a 1 g 1 (x,y)+a 2 g 2 (x,y)

[0075] where a 1 and a 2 is the weight. According to the sensitivity of human eyes to grayscale (32-192 range), the grayscale range is divided into three parts during image processing, and the weights are taken in segments for synthesis. The choice of weight depends on the brightness of the scene to ensure the ideal image effect under different lighting conditions:

[0076] Table 1 Weight calculation formula

[0077]

[0078] Linear expansion algorithm implementation steps reference Figure 2 .

[0079] As a further improvement, in step S5, the OCR character recognition algorithm can be compared and selected based on existing mature algorithms or combined algorithms.

[0080] Those skilled in the art may understand this embodiment as a more specific description of Embodiment 1.

[0081] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0082] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A high temperature slab printing character recognition system, characterized in that: include: An operation terminal, a roller control system, a process control system, a communication module, an industrial camera, an image recognition module and an image processing module; the operation terminal, the roller control system, the process control system, the communication module, the industrial camera, the image recognition module and the image processing module are connected through a network to form a business function system, wherein: The industrial camera is installed above the slab transport roller and is used to capture the image of the characters printed on the slab; The image processing module uses a linear expansion algorithm to perform strong light suppression processing on the captured image; The image recognition module uses an OCR character recognition algorithm to perform character recognition on the processed image.

2. The high temperature slab printing character recognition system according to claim 1 is characterized in that: The linear expansion algorithm expands the dynamic range of the image by stretching the interval between the low-grayscale and high-grayscale areas in the image and compressing the medium-grayscale area; uses different shutter speeds to capture multiple images, and extracts detail information at the same position from the images to synthesize new images.

3. The high temperature slab printing character recognition system according to claim 2 is characterized in that: The method for synthesizing a new image is to linearly weight the grayscale values ​​of the two images to form a new grayscale value gf(x,y): gf(x,y)=a1g1(x,y)+a2g2(x,y) Among them, a1 and a2 are weights, which are calculated based on the grayscale value of the image and the preset boundary conditions.

4. The high temperature slab printing character recognition system according to claim 3 is characterized in that: The weights a1 and a2 are calculated according to the following formula: When [g1(x,y) + g2(x,y)] / 2 <= T min then, a1 = [g1(x,y) + g2(x,y)] / (4T min ), a2 = 1 - a1; When T min <=[g1(x,y)+g2(x,y)] / 2<=T max When a1=a2=0.5; When [g1(x,y) + g2(x,y)] / 2 >= T max then, a2 = [256 - (g1(x,y) + g2(x,y)) / 2] / 2(256 - T max ), a1 = 1 - a2; Among them, Tmin and Tmax are preset grayscale thresholds.

5. The high temperature slab printing character recognition system according to claim 1, characterized in that: The image recognition module performs character recognition on the processed image through an OCR character recognition algorithm, and feeds back the recognition result to the process control system through the communication module.

6. The high temperature slab printing character recognition system according to claim 1, characterized in that: The process control system is used to receive the slab arrival information sent by the roller control system, and after determining the slab arrival signal, send down an identification signal, and receive the recognition result of the image recognition module, compare and determine it with the data in the actual production performance, and store the recognition result and determination result in the database.

7. The high temperature slab printing character recognition system according to claim 1, characterized in that: The operation terminal is used to connect to the process control system database, obtain the slab printing character recognition results, judgment results and statistical analysis data, and display them through the application screen.

8. A method for character recognition of high temperature slab printing, characterized in that: The method uses the high-temperature slab printing character recognition system according to any one of claims 1 to 7, and the method comprises the following steps: Step S1: installing an industrial camera above the slab transport roller to capture images of characters printed on the slab; Step S2: After the roller control system moves the slab to the fixed identification position, it sends the in-position information to the process control system; Step S3: After the process control system determines that the slab is in place, it sends a recognition signal and calls the image processing module to process the captured image; Step S4: the image processing module uses a linear expansion algorithm to perform strong light suppression processing on the image to eliminate the red light effect caused by the high-temperature slab; Step S5: The image recognition module uses an OCR character recognition algorithm to perform character recognition on the processed image and feeds back the recognition result to the process control system.

9. The high temperature slab printing character recognition method according to claim 8, characterized in that: The processing steps of the linear expansion algorithm in step S4 include: capturing multiple images using different shutter speeds, extracting detail information at the same position, calculating weights based on the grayscale values ​​of the images and preset boundary conditions, linearly weighting the grayscale values ​​of the two images, and synthesizing a new image.

10. The method for high temperature slab printing character recognition according to claim 8, characterized in that: In step S5, after receiving the recognition result, the process control system compares and determines it with the data in the actual production performance, and stores the recognition result and the determination result in the database for subsequent data statistics and result display.

Images