A method for on-line monitoring of the temperature of quenched steel pipes in full field of view

Abstract

The present application belongs to the field of metallurgical industry automatic detection technology, and particularly relates to a kind of full field of view online monitoring method for the temperature of quenched steel pipe.The method comprises synchronously collecting the visible light image and thermal imaging temperature data of the transverse chain bed area where the quenched steel pipe is located;based on deep learning algorithm, the steel pipe area in the visible light image is intelligently identified and segmented;the identified steel pipe area is spatially coordinate registered and data fused with the thermal imaging temperature data, and the full-length temperature distribution data of the steel pipe surface is extracted;the steel pipe process parameters are obtained and associated with the temperature distribution data;through the visual interface, the quenched steel pipe full-length temperature is non-contact, full field of view and online continuously monitored.The present application realizes non-contact, full field of view and online continuous monitoring of the quenched steel pipe full-length temperature, without manual intervention, improves the temperature measurement efficiency and safety, and provides accurate data support for steel pipe heat treatment process optimization.

Description

Technical Field

[0001] This invention belongs to the field of automated detection technology in the metallurgical industry, and specifically relates to a method for online monitoring of the temperature of quenched steel pipes across the entire field of view. Background Technology

[0002] In steel pipe rolling and heat treatment production lines, the quenching process is a crucial step that determines the final mechanical properties and service life of the steel pipe. After water or oil quenching, the uniformity of the surface temperature distribution directly affects the phase transformation process, residual stress distribution, and the effectiveness of subsequent tempering processes. Excessive temperature differences along the entire length of the steel pipe can easily lead to quality defects such as insufficient local hardness, warping, and even cracking. Therefore, continuous and accurate temperature monitoring along the entire length of the quenched steel pipe is an important means of process optimization and quality control.

[0003] Currently, the industry commonly uses handheld infrared thermometers to randomly inspect quenched steel pipes. Operators must enter high-temperature, high-humidity areas such as transverse conveyor belts to measure the surface temperature of the steel pipes point by point. This method has three significant drawbacks: First, the temperature measurement efficiency is extremely low, making it difficult to achieve continuous monitoring of every steel pipe along its entire length, often requiring temporary shutdowns and severely impacting production rhythm. Second, the on-site environment is harsh, with high-temperature heat radiation, high-pressure water mist, and moving mechanical parts, posing a threat to the personal safety of operators. Third, manual sampling only obtains discrete "point" temperature data, failing to reflect the complete temperature distribution along the axial direction of the steel pipe, making it difficult for process engineers to accurately judge quenching uniformity and lacking sufficient data to support closed-loop process optimization. In recent years, some companies have attempted to install fixed infrared thermal imagers to measure the temperature of steel pipe areas. However, these systems typically lack intelligent recognition capabilities, failing to automatically segment the steel pipe itself from complex backgrounds (chain bed, water mist, adjacent steel pipes), and are susceptible to interference from environmental heat sources. Furthermore, the thermal imaging images have low resolution and are separate from the on-site visible light monitoring screen, making it difficult for operators to intuitively correlate temperature data with specific steel pipe locations. In addition, existing systems are not integrated with the production control system (PLC / MES), failing to automatically bind process parameters such as steel pipe numbers and specifications; the data is isolated and cannot be used for quality traceability.

[0004] Therefore, there is an urgent need for a full-field monitoring method that can integrate visible light and thermal imaging information, possess intelligent steel pipe identification capabilities, and extract the full-length temperature distribution online and automatically bind it to process parameters, in order to overcome the shortcomings of existing technologies. Summary of the Invention

[0005] To address the aforementioned problems, the purpose of this invention is to provide a full-field-of-view online monitoring method for the temperature of quenched steel pipes. This invention achieves full-field-of-view, online, non-contact temperature monitoring of quenched steel pipes. Through deep learning, it automatically identifies and segments the steel pipe region, and combines visible light and thermal imaging data fusion to accurately extract the temperature distribution along the entire length of the steel pipe. This distribution is automatically linked and bound to process parameters, supporting real-time visualization and over-temperature warnings. This significantly improves temperature measurement efficiency and safety, providing complete and reliable data support for quenching process optimization and quality traceability. Furthermore, the system has excellent adaptability to harsh environments.

[0006] The technical solution of this invention is: a method for online monitoring of the temperature of quenched steel pipes across the entire field of view, comprising the following steps: S1: Synchronously acquire visible light images and thermal imaging temperature data of the monitoring area of ​​the transverse chain bed where the quenched steel pipe is located; S2: Based on deep learning algorithms, the steel pipe region in the visible light image is identified and segmented in real time to obtain the pixel-level contour mask and position information of the steel pipe; S3: Pre-establish a mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates. Register the steel pipe outline and position information obtained in step S2 with the thermal imaging temperature data collected in step S1 in spatial coordinates. Extract the surface temperature distribution data corresponding to the steel pipe area through a data fusion algorithm to generate a two-dimensional temperature distribution matrix on the surface of the steel pipe. Extract the temperature distribution dataset along the entire length of the steel pipe along the axial direction of the steel pipe. S4: Obtain the steel pipe process parameters and associate and bind the temperature distribution dataset of the entire length of the steel pipe with the steel pipe process parameters; S5: Through a visual interface, the system displays in real time visible light images, thermal images, axial temperature trend curves of the steel pipe, and associated steel pipe process parameters of the monitored area, enabling non-contact, full-field, online continuous monitoring of the temperature of the entire length of the quenched steel pipe.

[0007] Furthermore, the deep learning algorithm described in step S2 is a steel pipe instance segmentation model based on Mask R-CNN. This model includes a backbone network, a region proposal network, an ROI Align layer, and classification branches, bounding box regression branches, and mask branches. The model's loss function is the sum of the classification loss, bounding box regression loss, and mask loss, where: The classification loss uses cross-entropy loss L cls Specifically: ; In the formula, L cls is the classification loss value, used to measure the difference between the model's predicted class and the true class, where i is the detected steel pipe instance in the image. The i-th instance belongs to the real category The probability of , taking values ​​in the range (0,1), This is the actual category label for this instance; The bounding box regression loss uses the Smooth L1 loss, specifically: ; In the formula, L box t represents the bounding box regression loss value, which measures the difference between the predicted bounding box and the true bounding box. i Parameterize the coordinate vector of the bounding box predicted by the model. The parameterized coordinate vector of the true bounding box. or ; The mask loss uses binary cross-entropy loss L mask Specifically: ; In the formula, L mask , where u and v are the pixel coordinates (row and column) in the image, representing the difference between the predicted mask and the real mask. M(u,v) is the value of the real mask at pixel (u,v), which takes the value 0 (non-steel pipe) or 1 (steel pipe). M(u,v) is the probability of the model predicting the mask at pixel (u,v), which takes the value (0,1).

[0008] Furthermore, the specific steps in step S3 for establishing the mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates are as follows: Several pairs of corresponding points are obtained using a calibration plate, and the homography matrix H is solved using the direct linear transformation method, satisfying... , where (ut, vt) are the pixel coordinates of the thermal image, and (uc, vc) are the pixel coordinates of the visible light image.

[0009] Furthermore, step S3, which involves extracting the surface temperature distribution data corresponding to the steel pipe region using a data fusion algorithm, specifically includes: S31: For each pixel of the steel pipe mask Mi(uc, vc)=1 in the visible light image, calculate the corresponding thermal imaging coordinates based on the inverse of the homography matrix: ; S32: Extract the temperature value at this coordinate Tc(uc, vc) = T(round(ut), round(vt)), where T is the thermal imaging temperature data matrix; S33: Generate a two-dimensional temperature distribution matrix Wi(uc, vc) = Tc(uc, vc) Mi(uc, vc)=1) or NaN on the surface of the steel pipe.

[0010] Furthermore, the step S3, which involves extracting the temperature distribution dataset along the entire length of the steel pipe, specifically involves: calculating the average temperature column-by-column of the two-dimensional temperature distribution matrix along the axial direction of the steel pipe to form an axial temperature distribution array. In the formula, x is the axial pixel coordinate, Hx is the pixel height of the steel pipe in the direction perpendicular to the axial direction, and the final output is the axial temperature distribution array TempProfile of the steel pipe. i = [Ti(1), Ti(2), ..., Ti(L)], where L is the pixel length of the steel pipe in the axial direction.

[0011] Furthermore, the steel pipe process parameters mentioned in step S4 include at least the steel pipe number, length, and diameter; the association binding specifically means that when a trigger signal is received that a steel pipe has entered the monitoring area, the currently identified steel pipe is matched one by one with the steel pipe process parameters read from the control system.

[0012] An online temperature monitoring system for quenched steel pipes based on the above-described method includes: Image acquisition probe: used to acquire visible light images; Thermal imaging temperature probe: used to collect thermal imaging temperature data; Control box: Built-in data processing unit for executing the method steps described above; Client software: Used to provide a visual interface and enable data interaction.

[0013] The image acquisition probe and the thermal imaging temperature measurement probe are respectively installed on the aligned side of the steel pipe and are equipped with a double-layer protective cover and a compressed air purging interface.

[0014] The technical advantages of this invention are as follows: 1. This invention utilizes an image acquisition probe and a thermal imaging temperature measurement probe to simultaneously acquire visible light images and thermal imaging temperature data of the transverse chain bed area. This allows for real-time monitoring of the full-length temperature of multiple steel pipes within the area without machine downtime or manual intervention. Compared to traditional manual handheld temperature gun inspection, this method can achieve continuous temperature acquisition for each steel pipe along its entire length, significantly improving monitoring efficiency and meeting production line cycle requirements. 2. Through an instance segmentation model based on Mask R-CNN, the model can automatically and accurately segment the contour mask of each steel pipe from complex backgrounds (chain bed, water mist, adjacent steel pipes, etc.) at the pixel level and output position information. This model can adapt to steel pipes of different diameters and lengths, effectively solving the problems of traditional fixed infrared temperature measurement devices being easily interfered with by environmental heat sources and unable to intelligently identify the steel pipe itself, thus improving the accuracy of temperature extraction. 3. By establishing a precise mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates (homography transformation), the thermal imaging temperature data is spatially registered and fused with the steel pipe mask to generate a two-dimensional temperature distribution matrix on the steel pipe surface. Then, the full-length temperature distribution array (TempProfile) is extracted along the axial direction. This method upgrades traditional discrete "point" temperature measurement to "full-field" and "continuous" temperature distribution data, comprehensively reflecting the axial temperature uniformity of the steel pipe and providing ample data support for quenching process optimization.

[0015] The following will provide further explanation in conjunction with the accompanying drawings. Attached Figure Description

[0016] Figure 1 This is a flowchart of an online monitoring method for the temperature of quenched steel pipes across the entire field of view, according to the present invention.

[0017] Figure 2 This is a schematic diagram of the installation of the image acquisition probe and the thermal imaging temperature measurement probe of the present invention.

[0018] Figure 3 This is a schematic diagram showing the real-time visualization interface of the present invention. Detailed Implementation

[0019] Example 1 like Figure 1 As shown, a method for online monitoring of the temperature of quenched steel pipes across the entire field of view includes the following steps: S1: Synchronously acquire visible light images and thermal imaging temperature data of the monitoring area of ​​the transverse chain bed where the quenched steel pipe is located; S2: Based on deep learning algorithms, the steel pipe region in the visible light image is identified and segmented in real time to obtain the pixel-level contour mask and position information of the steel pipe; S3: Pre-establish a mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates. Register the steel pipe outline and position information obtained in step S2 with the thermal imaging temperature data collected in step S1 in spatial coordinates. Extract the surface temperature distribution data corresponding to the steel pipe area through a data fusion algorithm to generate a two-dimensional temperature distribution matrix on the surface of the steel pipe. Extract the temperature distribution dataset along the entire length of the steel pipe along the axial direction of the steel pipe. S4: Obtain the steel pipe process parameters and associate and bind the temperature distribution dataset of the entire length of the steel pipe with the steel pipe process parameters; S5: Through a visual interface, the system displays in real time visible light images, thermal images, axial temperature trend curves of the steel pipe, and associated steel pipe process parameters of the monitored area, enabling non-contact, full-field, online continuous monitoring of the temperature of the entire length of the quenched steel pipe.

[0020] Furthermore, the deep learning algorithm described in step S2 is a steel pipe instance segmentation model based on Mask R-CNN. This model includes a backbone network, a region proposal network, an ROI Align layer, and classification branches, bounding box regression branches, and mask branches. The model's loss function is the sum of the classification loss, bounding box regression loss, and mask loss, where: The classification loss uses cross-entropy loss L cls Specifically: ; In the formula, L cls is the classification loss value, used to measure the difference between the model's predicted class and the true class, where i is the detected steel pipe instance in the image. The i-th instance belongs to the real category The probability of , taking values ​​in the range (0,1), This is the actual category label for this instance; The bounding box regression loss uses the Smooth L1 loss, specifically: ; In the formula, L box t represents the bounding box regression loss value, which measures the difference between the predicted bounding box and the true bounding box. i Parameterize the coordinate vector of the bounding box predicted by the model. The parameterized coordinate vector of the true bounding box. or ; The mask loss uses binary cross-entropy loss L mask Specifically: ; In the formula, L mask, where u and v are the pixel coordinates (row and column) in the image, representing the difference between the predicted mask and the real mask. M(u,v) is the value of the real mask at pixel (u,v), which takes the value 0 (non-steel pipe) or 1 (steel pipe). M(u,v) is the probability of the model predicting the mask at pixel (u,v), which takes the value (0,1).

[0021] Furthermore, the specific steps in step S3 for establishing the mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates are as follows: Several pairs of corresponding points are obtained using a calibration plate, and the homography matrix H is solved using the direct linear transformation method, satisfying... , where (ut, vt) are the pixel coordinates of the thermal image, and (uc, vc) are the pixel coordinates of the visible light image.

[0022] Furthermore, step S3, which involves extracting the surface temperature distribution data corresponding to the steel pipe region using a data fusion algorithm, specifically includes: S31: For each pixel of the steel pipe mask Mi(uc, vc)=1 in the visible light image, calculate the corresponding thermal imaging coordinates based on the inverse of the homography matrix: ; S32: Extract the temperature value at this coordinate Tc(uc, vc) = T(round(ut), round(vt)), where T is the thermal imaging temperature data matrix; S33: Generate a two-dimensional temperature distribution matrix Wi(uc, vc) = Tc(uc, vc) Mi(uc, vc)=1) or NaN on the surface of the steel pipe.

[0023] Furthermore, the step S3, which involves extracting the temperature distribution dataset along the entire length of the steel pipe, specifically involves: calculating the average temperature column-by-column of the two-dimensional temperature distribution matrix along the axial direction of the steel pipe to form an axial temperature distribution array. In the formula, x is the axial pixel coordinate, Hx is the pixel height of the steel pipe in the direction perpendicular to the axial direction, and the final output is the axial temperature distribution array TempProfile of the steel pipe. i = [Ti(1), Ti(2), ..., Ti(L)], where L is the pixel length of the steel pipe in the axial direction.

[0024] Furthermore, the steel pipe process parameters mentioned in step S4 include at least the steel pipe number, length, and diameter; the association binding specifically means that when a trigger signal is received that a steel pipe has entered the monitoring area, the currently identified steel pipe is matched one by one with the steel pipe process parameters read from the control system.

[0025] Example 2 An online temperature monitoring system for quenched steel pipes based on the above-described method includes: Image acquisition probe: used to acquire visible light images; Thermal imaging temperature probe: used to collect thermal imaging temperature data; Control box: Built-in data processing unit for executing the method steps described above; Client software: Used to provide a visual interface and enable data interaction.

[0026] The image acquisition probe and the thermal imaging temperature measurement probe are respectively installed on the aligned side of the steel pipe and are equipped with a double-layer protective cover and a compressed air purging interface.

[0027] Example 3 The online monitoring method for the full field of view of the temperature of the quenched steel pipe, as described in Example 1, is used to monitor the transverse chain bed area of ​​the quenched steel pipe. The specific process is as follows: S1: Data Acquisition In the transverse chain bed area after the steel pipe has been quenched, an image acquisition probe and a thermal imaging temperature measurement probe are installed on the aligned side of the steel pipe. The image acquisition probe is specifically an industrial camera, and the thermal imaging temperature measurement probe is specifically a medium-wave thermal imager. This ensures that their field of view covers the entire steel pipe inspection area. The two probes work synchronously to acquire the visible light video stream of this area in real time with a resolution of 1440×1080. The thermal imaging temperature data stream has a resolution of 640×480 and a temperature measurement range of 0-800℃. The probes are equipped with double-layer protective covers and are connected to compressed air for cooling and purging to prevent interference from high temperature and water mist. S2: Intelligent recognition and segmentation of steel pipes: The acquired visible light images are transmitted in real time to the data processing unit in the control box. This unit runs a pre-trained deep learning model, a steel pipe instance segmentation model based on Mask R-CNN. This model can automatically identify the contours of all steel pipes in the image and generate a pixel-level mask for each steel pipe, accurately outputting the position information of the steel pipe in the visible light image. S3: Data Fusion and Temperature Extraction The data processing unit performs spatial coordinate registration between the steel pipe position information obtained in step S2 and the thermal imaging data collected in step S1. Specifically, a coordinate mapping matrix between the thermal imaging camera and the visible light camera is established in advance through a calibration plate. Based on this matrix, the fusion algorithm accurately maps the temperature value of each pixel in the thermal imaging image to each pixel in the steel pipe mask area in the visible light image, thereby generating a two-dimensional temperature distribution matrix on the surface of the steel pipe. By processing the matrix along the axial direction of the steel pipe, the temperature values ​​of each point along the entire length of the steel pipe can be extracted to form a steel pipe axial temperature distribution dataset. S4: Data Linking and Uploading The system establishes communication with the production line control system PLC / secondary system via industrial Ethernet. When the water cooling of the steel pipe ends and it is loaded onto the transverse conveyor bed, the control system sends a trigger signal. After receiving the signal, the system automatically matches and binds the currently identified steel pipe with the process parameters such as the steel pipe number, length, and diameter read from the control system. Then, it packages the "steel pipe number + time information + full-length temperature distribution array" and uploads it back to the control system via protocols such as Web Socket for process closed-loop control. S5: Visual Display The system client software provides a main interface, such as... Figure 3 As shown, the interface includes: Video display area: Real-time visible light images are displayed, with each steel pipe showing its number and current average or highest temperature floating above it. Thermal image display area: Real-time display of thermal imaging pseudo-color images. Trend curve display area: After a user clicks on a steel pipe, the axial temperature trend curve of that steel pipe from head to tail will be displayed. If the temperature at a certain point exceeds a set threshold, the curve color will change, and the corresponding steel pipe area in the video area will turn red as an alarm. Process parameter area: Real-time display of information such as the current steel pipe number, length, and diameter read from the control system; Through the above steps, this method achieves full-field, online, and intelligent monitoring of the temperature of quenched steel pipes; To verify the applicability of the method of the present invention, an experimental verification was conducted in a production line for the method of online monitoring of the temperature of quenched steel pipes across the entire field of view. A schematic diagram of the experimental system is shown below. Figure 2 The main interface of the experimental system client software can be found here. Figure 3 The experiment yielded good results and can be further promoted in production.

[0028] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for online monitoring of the temperature of quenched steel pipes across the entire field of view, characterized in that: Includes the following steps: S1: Synchronously acquire visible light images and thermal imaging temperature data of the monitoring area of ​​the transverse chain bed where the quenched steel pipe is located; S2: Based on deep learning algorithms, the steel pipe region in the visible light image is identified and segmented in real time to obtain the pixel-level contour mask and position information of the steel pipe; S3: Pre-establish a mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates. Register the steel pipe outline and position information obtained in step S2 with the thermal imaging temperature data collected in step S1 in spatial coordinates. Extract the surface temperature distribution data corresponding to the steel pipe area through a data fusion algorithm to generate a two-dimensional temperature distribution matrix on the surface of the steel pipe. Extract the temperature distribution dataset along the entire length of the steel pipe along the axial direction of the steel pipe. S4: Obtain the steel pipe process parameters and associate and bind the temperature distribution dataset of the entire length of the steel pipe with the steel pipe process parameters; S5: Through a visual interface, the system displays in real time visible light images, thermal images, axial temperature trend curves of the steel pipe, and associated steel pipe process parameters of the monitored area, enabling non-contact, full-field, online continuous monitoring of the temperature of the entire length of the quenched steel pipe.

2. The method for online monitoring of the temperature of quenched steel pipes across the entire field of view according to claim 1, characterized in that, The deep learning algorithm described in step S2 is a steel pipe instance segmentation model based on Mask R-CNN. This model includes a backbone network, a region proposal network, an ROI Align layer, and classification branches, bounding box regression branches, and mask branches. The model's loss function is the sum of the classification loss, bounding box regression loss, and mask loss, where: The classification loss uses cross-entropy loss L cls Specifically: ； In the formula, L cls is the classification loss value, used to measure the difference between the model's predicted class and the true class, where i is the detected steel pipe instance in the image. The i-th instance belongs to the real category The probability of , taking values ​​in the range (0,1), This is the actual category label for this instance; The bounding box regression loss uses the Smooth L1 loss, specifically: ； In the formula, L box t represents the bounding box regression loss value, which measures the difference between the predicted bounding box and the true bounding box. i Parameterize the coordinate vector of the bounding box predicted by the model. The parameterized coordinate vector of the true bounding box. or ; The mask loss uses binary cross-entropy loss L mask Specifically: ； In the formula, L mask , where u and v are the pixel coordinates (row and column) in the image, representing the difference between the predicted mask and the real mask. M(u,v) is the value of the real mask at pixel (u,v), which takes the value 0 (non-steel pipe) or 1 (steel pipe). M(u,v) is the probability of the model predicting the mask at pixel (u,v), which takes the value (0,1).

3. The method for online monitoring of the temperature of quenched steel pipes across the entire field of view according to claim 2, characterized in that, Step S3, establishing the mapping matrix between thermal imaging pixel coordinates and visible light image pixel coordinates, specifically involves: obtaining several sets of corresponding point pairs through a calibration plate, and solving for the homography matrix H using the direct linear transformation method, satisfying... , where (ut, vt) are the pixel coordinates of the thermal image, and (uc, vc) are the pixel coordinates of the visible light image.

4. The method for online monitoring of the temperature of quenched steel pipes across the entire field of view according to claim 3, characterized in that, Step S3, which involves extracting the surface temperature distribution data corresponding to the steel pipe region using a data fusion algorithm, specifically includes: S31: For each pixel of the steel pipe mask Mi(uc, vc)=1 in the visible light image, calculate the corresponding thermal imaging coordinates based on the inverse of the homography matrix: ; S32: Extract the temperature value at this coordinate Tc(uc, vc) = T(round(ut), round(vt)), where T is the thermal imaging temperature data matrix; S33: Generate a two-dimensional temperature distribution matrix Wi(uc, vc) = Tc(uc, vc) on the surface of the steel pipe. Mi(uc, vc) = 1 or NaN.

5. The method for online monitoring of the temperature of quenched steel pipes across the entire field of view according to claim 4, characterized in that, Step S3, which involves extracting the temperature distribution dataset along the entire length of the steel pipe along its axial direction, specifically involves calculating the average temperature column-by-column of the two-dimensional temperature distribution matrix to form an axial temperature distribution array. ； In the formula, x is the axial pixel coordinate, Hx is the pixel height of the steel pipe in the direction perpendicular to the axial direction, and the final output is the axial temperature distribution array TempProfile of the steel pipe. i = [Ti(1), Ti(2), ..., Ti(L)], where L is the pixel length of the steel pipe in the axial direction.

6. The method for online monitoring of the temperature of a quenched steel pipe across the entire field of view according to claim 5, characterized in that, The steel pipe process parameters mentioned in step S4 include at least the steel pipe number, length, and diameter; the association binding specifically means that when a trigger signal is received that a steel pipe has entered the monitoring area, the currently identified steel pipe is matched one by one with the steel pipe process parameters read from the control system.

7. An online temperature monitoring system for quenched steel pipes based on the method of any one of claims 1 to 6, characterized in that, include: Image acquisition probe: used to acquire visible light images; Thermal imaging temperature probe: used to collect thermal imaging temperature data; Control box: Built-in data processing unit for executing the steps of the method described in claim 1; Client software: Used to provide a visual interface and enable data interaction.

8. The online temperature monitoring system for quenched steel pipes according to claim 7, characterized in that, The image acquisition probe and the thermal imaging temperature measurement probe are respectively installed on the aligned side of the steel pipe and are equipped with a double-layer protective cover and a compressed air purging interface.

Images