Detection method for tuyere temperature measurement based on dual-light image fusion matrix

Abstract

The invention discloses a tuyere temperature measurement detection method based on a dual-light image fusion matrix, and the method comprises the steps: collecting an image video stream through a pan-tilt camera, reading visible light and thermal infrared images, carrying out the preprocessing of the images through Gaussian filtering, extracting a target object ROI region in the images, carrying out the bilinear interpolation zooming under the same resolution according to the features of the ROI region, and carrying out the detection of the tuyere temperature measurement. Feature point detection is carried out by using an AKAZE algorithm, a digital transformation matrix is solved, and image information fusion is carried out; the method comprises the following steps: acquiring a visible light image, identifying a temperature-measuring object target by using a YOLOV5 algorithm, projecting object contour coordinates identified by the visible light image into an infrared image by using an image fusion matrix, extracting the contour of the measured object from the infrared image, and measuring the temperature of related high-temperature areas of a belly pipe, a tuyere and a gun bag by combining temperature data of the infrared image. The computer vision technology is used, the problem that a high-temperature point cannot be found through infrared general measurement is solved, and the field worker inspection efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of blast furnace tuyere temperature measurement, and in particular to a detection method for tuyere temperature measurement based on a dual-light image fusion matrix. Background Art

[0002] The blast furnace tuyere is on the blast furnace of the ironmaking plant. The high-temperature hot air blown into the blast furnace tuyere and the coke at the bottom of the furnace are oxidized and burned to generate CO. CO is reduced to iron in the form of oxides at high temperatures. It is the most important part of steelmaking and ironmaking. The material of the blast furnace tuyere is mainly high-purity copper. Considering the strength, rigidity, and crack resistance of the tuyere, the material state includes forging, copper plate rolling, and casting. The copper plate rolling tuyere has low weight and low cost, but if the wall thickness is too thin and the rigidity is insufficient, it is easy to deform. Because the blast furnace tuyere is cooled by water to keep the tuyere body running at a low temperature to maintain strength and rigidity, the purity of copper is very important. The traditional method of checking the tuyere is to measure the temperature of the tuyere on site with a manual handheld temperature measuring gun, and the on-site environment is very harmful to the human body. Summary of the invention

[0003] The purpose of the present invention is to provide a detection method for temperature measurement of air outlets based on a dual-light image fusion matrix. By cooperating with the algorithm of robot autonomous navigation, the temperature measurement work is carried out on-site at the air outlets. The high-temperature areas of the direct blowing pipe, air outlet, gun package and related positions can be automatically inspected regularly or as needed, realizing temperature measurement without dead angles and eliminating the blind spots of manual point temperature measurement, so as to solve the problems raised by the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a detection method for air outlet temperature measurement based on a dual light image fusion matrix, comprising the following steps:

[0005] S1: Call the image video stream collected by the inspection robot's pan-tilt camera to read visible light and thermal infrared images;

[0006] S2: Use Gaussian filtering to preprocess the image to improve image quality for subsequent ROI area extraction;

[0007] S3: Extract the target ROI area in the image, and perform bilinear interpolation and scaling to the same resolution according to the ROI area features for subsequent feature point detection and image information fusion;

[0008] S4: Use AKAZE algorithm to detect feature points, obtain digital transformation matrix, and perform image information fusion;

[0009] S5: Use the YOLOV5 algorithm to identify temperature measurement objects, including the blowpipe, tuyere, and gun bag;

[0010] S6: Use the image fusion matrix to project the coordinates of the object contour identified by the visible light image into the infrared image, extract the contour of the object to be measured from the infrared image, and then combine it with the temperature data of the infrared image to achieve temperature measurement of the high-temperature areas related to the direct blowpipe, tuyere, and gun package.

[0011] Furthermore, the inspection robot in S1 has a built-in autonomous navigation programming program, which is used to pre-set the robot's inspection tasks.

[0012] Furthermore, the specific method in S2 is: use Gaussian filtering to traverse the image, use a pixel point in the image as the center of the convolution kernel, use the convolution kernel to make a weighted average of the neighboring pixels around the pixel point, and use the calculation result as the new pixel value of the current pixel point to achieve Gaussian filtering and noise removal of the image.

[0013] Furthermore, the method for extracting image ROI region features in S3 is as follows:

[0014] S301: Perform OTSU binarization threshold segmentation and Canny contour extraction on the image, and then use the shape scale feature of the aspect ratio to screen and obtain the required target object ROI area;

[0015] S302: extracting the contour shape features of the ROI area, performing minimum bounding rectangle fitting, determining the scaling ratio according to the size of the bounding rectangle, and performing image scaling using a bilinear interpolation method.

[0016] Furthermore, the calculation method of bilinear interpolation in S3 is as follows:

[0017] S301': define four points of the image Q11 (x1, y1), Q12 (x1, y2), Q21 (x2, y1), Q22 (x2, y2), and the pixel values ​​are f(Q11), f(Q12), f(Q21), f(Q22);

[0018] S302': For the pixel value f(P) of any point P(x, y) between the four points, linear interpolation is first performed in the x direction, and then linear interpolation is performed in the y direction to obtain the following relationship:

[0019]

[0020] The pixel value of point P can be solved using this relationship.

[0021] Furthermore, in S4, the FED algorithm is used to solve the nonlinear diffusion equation when using the AKAZE algorithm. The FED algorithm solves the nonlinear diffusion equation: L i+1 =(I+τA(L i ))L i , i=0,1,Λ,n-1;

[0022] Where I is the identity matrix, A(L i ) is the conduction matrix of the image in dimension i, constructed by the scale-affected gradient histogram after Gaussian filtering, τ is the time step, and t is the difference in evolution time i-1 -t i .

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. The detection method of the temperature measurement of the tuyere based on the dual-light image fusion matrix of the present invention, in conjunction with the autonomous navigation of the robot, can promptly detect abnormalities in the hot blast furnace straight blowing pipe, reducer, upper corrugation, and lower corrugation, thereby realizing the temperature measurement of the blast furnace tuyere without blind spots.

[0025] 2. The detection method of air outlet temperature measurement based on the dual-light image fusion matrix of the present invention can greatly reduce the labor intensity of workers, lock high-temperature positions and high-temperature points, improve work efficiency, reduce the safety risks of construction operations, and improve the economic benefits of the enterprise.

[0026] 3. The detection method of air outlet temperature measurement based on the dual-light image fusion matrix of the present invention uses two different types of light sources or wavelengths to provide different types of information. This multimodal information acquisition method can more comprehensively understand the state of the object being measured. By combining visual recognition and infrared temperature measurement, high-resolution and high-precision temperature measurements can be performed in complex environments. Specific high-temperature points can be found, as well as tiny problem points that are difficult for the human eye to detect, such as pipe cracks and holes.

[0027] 4. The detection method for air outlet temperature measurement based on the dual-light image fusion matrix of the present invention uses a deep learning model to first identify the object to be measured, which increases the accuracy of the method, can provide more image information, and contributes to more accurate temperature measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a logic block diagram of the detection method of the present invention;

[0029] Figure 2 A pixel definition diagram for the bilinear interpolation algorithm of the present invention;

[0030] Figure 3 It is a schematic diagram of the framework of the YOLOV5 algorithm of the present invention;

[0031] Figure 4 This is a state diagram of a case implementation in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] like Figure 1 As shown, an embodiment of the present invention provides a method for detecting temperature of an air outlet based on a dual-light image fusion matrix, comprising the following steps:

[0034] S1: Call the image video stream collected by the inspection robot's pan-tilt camera to read visible light and thermal infrared images; the inspection robot has a built-in autonomous navigation programming program, which can pre-set the robot's inspection tasks, realizing the inspection and temperature measurement of the blast furnace tuyere without blind spots;

[0035] S2: Use Gaussian filtering to preprocess the image to improve the image quality for subsequent ROI area extraction; specifically, for the original RGB and thermal infrared image data, use Gaussian filter to eliminate noise; Gaussian filtering is a weighted average filter, and its convolution kernel has a coefficient for averaging; the reciprocal of the sum of all values ​​in the matrix is ​​the coefficient of the convolution kernel; in actual filtering, the image is traversed, and a certain point in the image is used as the center of the convolution kernel. The convolution kernel is used to perform a weighted average of the neighboring pixels around the pixel point, and the calculated result is used as the new pixel value of the current pixel point; finally, Gaussian filtering of the image is achieved to remove noise, providing high-quality image data for subsequent ROI extraction and feature point detection.

[0036] Among them, the Gaussian distribution can select weights according to the Gaussian function. The one-dimensional form and two-dimensional form of the function when the mean μ=0 are shown in the formula: where σ is the standard deviation of the normal distribution, and its value determines the decay speed of the function.

[0037] One-dimensional form when μ=0:

[0038] Two-dimensional form when μ=0:

[0039] S3: Extract the target ROI area in the image, and perform bilinear interpolation and scaling to the same resolution according to the ROI area features for subsequent feature point detection and image information fusion;

[0040] Among them, the method for extracting image ROI area features is as follows:

[0041] S301: Perform OTSU binarization threshold segmentation and Canny contour extraction on the image, and then use the shape scale feature of the aspect ratio to screen and obtain the required target object ROI area;

[0042] S302: extracting the contour shape features of the ROI area, performing minimum bounding rectangle fitting, determining the scaling ratio according to the size of the bounding rectangle, and performing image scaling using a bilinear interpolation method.

[0043] Among them, the calculation method of bilinear interpolation is as follows:

[0044] S301': define four points of the image Q11 (x1, y1), Q12 (x1, y2), Q21 (x2, y1), Q22 (x2, y2), and the pixel values ​​are f(Q11), f(Q12), f(Q21), f(Q22);

[0045] S302': For the pixel value f(P) of any point P(x, y) between the four points, linear interpolation is first performed in the x direction, and then linear interpolation is performed in the y direction to obtain the following relationship:

[0046]

[0047] The pixel value of point P is solved by this relationship to ensure the quality of the scaled image, such as Figure 2 shown.

[0048] S4: Use the AKAZE algorithm to detect feature points, obtain the digital transformation matrix, and perform image information fusion; specifically, the AKAZE algorithm uses the FED (fast display algorithm) algorithm to solve the nonlinear diffusion equation. The FED algorithm can converge at any step size to obtain a stable scale space, which improves repeatability and specificity. In the description of the feature direction, it improves the calculation speed while improving the calculation accuracy.

[0049] FED algorithm solves the nonlinear diffusion equation: L i+1 =(I+τA(L i ))L i ,i=0,1,Λ,n-1

[0050] Where I is the identity matrix, A(L i ) is the conduction matrix of the image in dimension i, constructed by the scale-affected gradient histogram after Gaussian filtering, τ is the time step, and t is the difference in evolution time i-1 -t i .

[0051] S5: Use the YOLOV5 algorithm to identify temperature measurement objects, including the blowpipe, tuyere, and gun bag; Figure 3As shown in the figure, YOLOv5 is a single-stage target detection algorithm. This algorithm adds some new improvement ideas based on YOLOv4, which greatly improves its speed and accuracy. The main improvement ideas are as follows:

[0052] Input: During the model training phase, some improvement ideas were proposed, including Mosaic data enhancement, adaptive anchor box calculation, and adaptive image scaling.

[0053] Benchmark network: Integrates some new ideas from other detection algorithms, mainly including: Focus structure and CSP structure;

[0054] Neck network: The target detection network often inserts some layers between BackBone and the final Head output layer. Yolov5 adds the FPN+PAN structure.

[0055] Head output layer: The anchor box mechanism of the output layer is the same as that of YOLOv4. The main improvements are the loss function GIOU_Loss during training and DIOU_nms for prediction box screening.

[0056] S6: Use the image fusion matrix to project the coordinates of the object contour identified by the visible light image into the infrared image, extract the contour of the object to be measured from the infrared image, and then combine it with the temperature data of the infrared image to realize the temperature measurement function of the high-temperature areas related to the direct blow pipe, tuyere, and gun package.

[0057] In order to further better explain the embodiment of the present invention, the method of the present invention is also implemented in the blast furnace area of ​​the main plant, such as Figure 4 As shown, the high-temperature areas of the direct blow pipe, tuyere, gun package and related locations can be automatically inspected regularly or as needed, achieving temperature measurement without dead angles and eliminating the blind spots of manual point temperature measurement. When an abnormal situation is found, the system will automatically issue an alarm immediately to prevent safety accidents.

[0058] In summary: the detection method for temperature measurement of air outlets based on the dual-light image fusion matrix of the present invention, aimed at the temperature measurement requirements of the on-site environment, cooperates with the robot's autonomous navigation algorithm, sets the robot's inspection task, calls the image video stream collected by the robot's pan-tilt camera, reads visible light and thermal infrared images, performs image preprocessing, extracts contour features in the image, uses feature point detection to obtain the transformation matrix, and uses this to perform image registration and information fusion to obtain new image data; based on the visible light image, the direct blowing pipe, air outlet, gun package and related positions are identified and contours are extracted, and combined with the temperature information of the thermal infrared image, an air outlet patrol temperature detection algorithm based on visible light and thermal infrared images is realized, and automatic inspection of high-temperature areas of the direct blowing pipe, air outlet, gun package and related positions is realized, temperature measurement without dead angles is realized, and the blind spots of manual point temperature measurement are eliminated.

[0059] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A detection method for air outlet temperature measurement based on dual light image fusion matrix, It is characterized in that The following steps are involved: S1: Call the image video stream collected by the inspection robot's pan-tilt camera to read visible light and thermal infrared images; S2: Use Gaussian filtering to preprocess the image to improve image quality for subsequent ROI area extraction; S3: Extract the target ROI area in the image, and perform bilinear interpolation and scaling to the same resolution according to the ROI area features for subsequent feature point detection and image information fusion; S4: Use AKAZE algorithm to detect feature points, obtain digital transformation matrix, and perform image information fusion; S5: Use the YOLOV5 algorithm to identify temperature measurement objects, including the blowpipe, tuyere, and gun bag; S6: Use the image fusion matrix to project the coordinates of the object contour identified by the visible light image into the infrared image, extract the contour of the object to be measured from the infrared image, and then combine it with the temperature data of the infrared image to achieve temperature measurement of the high-temperature areas related to the direct blowpipe, tuyere, and gun package.

2. According to claim 1, a detection method for air outlet temperature measurement based on dual light image fusion matrix, It is characterized in that The S1 inspection robot has a built-in autonomous navigation programming program that is used to pre-set the robot's inspection tasks.

3. According to the detection method of air outlet temperature measurement based on dual light image fusion matrix as described in claim 1, It is characterized in that The specific method in S2 is: use Gaussian filtering to traverse the image, take a certain pixel in the image as the center of the convolution kernel, use the convolution kernel to make a weighted average of the neighboring pixels around the pixel, and use the calculated result as the new pixel value of the current pixel to achieve Gaussian filtering and noise removal of the image.

4. According to claim 1, a detection method for air outlet temperature measurement based on dual light image fusion matrix, It is characterized in that The method of extracting image ROI regional features in S3 is as follows: S301: Perform OTSU binarization threshold segmentation and Canny contour extraction on the image, and then use the shape scale feature of the aspect ratio to screen and obtain the required target object ROI area; S302: extracting the contour shape features of the ROI area, performing minimum bounding rectangle fitting, determining the scaling ratio according to the size of the bounding rectangle, and performing image scaling using a bilinear interpolation method.

5. According to claim 4, a method for measuring air outlet temperature based on a dual light image fusion matrix, It is characterized in that The calculation method of bilinear interpolation in S3 is as follows: S301': define four points of the image Q11 (x1, y1), Q12 (x1, y2), Q21 (x2, y1), Q22 (x2, y2), and the pixel values ​​are f(Q11), f(Q12), f(Q21), f(Q22); S302': For the pixel value f(P) of any point P(x, y) between the four points, linear interpolation is first performed in the x direction, and then linear interpolation is performed in the y direction to obtain the following relationship: The pixel value of point P can be solved using this relationship.

6. According to claim 1, a method for measuring air outlet temperature based on a dual light image fusion matrix, It is characterized in that In S4, the FED algorithm is used to solve the nonlinear diffusion equation when using the AKAZE algorithm. The FED algorithm solves the nonlinear diffusion equation: L i+1 =(I+τA(L i ))L i , i=0,1,Λ,n-1; Where I is the identity matrix, A(L i ) is the conduction matrix of the image in dimension i, constructed by the scale-affected gradient histogram after Gaussian filtering, τ is the time step, and t is the difference in evolution time i-1 -t i .

Images