Method for measuring width of casting blank through industrial CCD (charge coupled device) based on AI (artificial intelligence)

Abstract

The invention provides a method for measuring the width of a casting blank by an industrial CCD (Charge Coupled Device) based on AI (Artificial Intelligence), belongs to the technical field of casting blank width measurement, can overcome the problems of low measurement precision, easiness in environmental influence and the like in the prior art by combining an industrial CCD measurement technology with an AI model, and provides a casting blank width measurement method which is more accurate and stable. According to the method, the accuracy and efficiency of measurement can be improved, the dependence of manual operation can be reduced, and the error rate is reduced, so that the requirement of the ferrous metallurgy industry on high-precision measurement is met. In addition, the method is also suitable for other industrial fields requiring high-precision measurement, and has a wide application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of slab width measurement, and particularly to a method for measuring the width of a slab based on AI using an industrial CCD. Background Art

[0002] In traditional slab width measurement, contact measurement or simple non-contact measurement methods are often used. These methods have problems such as low measurement accuracy and being easily affected by the environment. The contact measurement method may cause measurement errors due to the wear of the measurement head or the contact pressure with the object to be measured. For non-contact measurement methods, such as laser ranging, although they avoid the defects of contact measurement, they are easily affected by factors such as environmental light and temperature, resulting in unstable measurement accuracy. Especially in the steel metallurgy environment with high temperature and high dust, the performance of these non-contact measurement methods is often greatly reduced. In addition, most of the existing measurement methods lack intelligent processing, and the analysis and processing of measurement data rely on manual operation, which is inefficient and error-prone.

[0003] Therefore, there is an urgent need in this field for a technical solution with high measurement accuracy and not affected by the environment.

[0004] The information disclosed in this background art section is only intended to enhance the overall understanding of the present invention and should not be regarded as an admission or any form of implication that this information constitutes prior art already known to those of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide a technical solution with high measurement accuracy and not affected by the environment.

[0006] To achieve the above purpose, the present invention provides the following solution:

[0007] A method for measuring the width of a slab based on AI using an industrial CCD, comprising:

[0008] The auxiliary line laser projects a visible laser line onto the surface of the slab to be measured. After the laser is reflected by the surface of the steel billet, it reaches the focal plane of the CCD detector through the lens. When the image of the laser line presented at the boundary of the steel billet shows segmentation / deflection, after extracting the boundary points using an image processing algorithm, the distance from the lens to the surface of the steel billet is measured by a height measurement sensor, and then the width of the steel billet can be calculated according to the size measurement principle of CCD optoelectronic imaging.

[0009] Optionally, the camera used for image acquisition is a high-speed camera, and the frame rate needs to be higher than the movement speed of the slab; the resolution is selected according to the displacement accuracy requirement; the trigger mode is external trigger synchronization; the camera is installed above the outlet roller table of the continuous caster, and the field of view covers the surface of the slab. The camera resolution is N, the pixel / field of view width is FOV, and the actual size represented by each pixel is:

[0010]

[0011] Optionally, feature point marking and extraction adopt natural feature method and optical flow method. The natural feature method uses SIFT and ORB algorithms to extract the scale oxide texture feature points, and the optical flow method tracks the displacement of feature points in adjacent frames.

[0012] Optionally, CCD width measurement is based on geometric optical imaging and digital image analysis. The CCD camera images the edge contour of the lens blank onto the photosensitive surface of the CCD sensor. The object-image relationship formula is:

[0013]

[0014] where W is the object width, W' is the projected width on the imaging plane, L is the distance from the object to the lens, and f is the focal length of the lens.

[0015] Optionally, the extraction of boundary points using the image processing algorithm includes:

[0016] Using the gray gradient method to locate the rough pixel position of the edge;

[0017] Improving the edge positioning accuracy to within 0.1 pixel through the difference algorithm.

[0018] Optionally, displacement calculation includes:

[0019] Calculating the actual displacement according to the pixel displacement and the calibration coefficient:

[0020] Δs = k·Δp;

[0021] where Δp is the pixel displacement of the feature point and k is the calibration coefficient;

[0022] Adopting Kalman filtering or moving average method to eliminate noise.

[0023] Optionally, it also includes calibration and dynamic error compensation.

[0024] Calibration includes static calibration, using a calibration block with a known size to determine the conversion coefficient between pixels and actual size; the calibration formula:

[0025] Wactual = k·(X2 - X1) + b;

[0026] where X1 and X2 are the pixel coordinates of the left and right edges, k is the calibration coefficient, and b is the compensation error; [[ID= fifty]] [[ID= fifty-one]]

[0027] [[ID= fifty-two]]Dynamic error compensation includes motion blur correction and vibration compensation; [[ID= fifty-three]] [[ID= fifty-four]]

[0028] [[ID= fifty-five]]Motion blur correction: short exposure time combined with high-frequency light source to freeze motion; [[ID= fifty-six]] [[ID= fifty-seven]]

[0029] Deconvolution algorithm for restoring blurred images;

[0030] Vibration compensation: Install shock-absorbing brackets or eliminate vibration displacement by referring to the background fixed points.

[0031] If the casting speed is v and the camera exposure time is t, the error caused by motion blur is Delta = v * t;

[0032] Eliminate the error by triggering synchronous exposure through the encoder or image deblurring algorithm.

[0033] Optionally, it further includes: temperature compensation and optical distortion correction:

[0034] Temperature compensation:

[0035] The thermal expansion of the casting blank will cause changes in the actual width. The compensation formula:

[0036] W 修正 = W 测量 / (1 + α·ΔT);

[0037] Where α is the coefficient of thermal expansion of the material;

[0038] Optical distortion correction is to use a telecentric lens to reduce perspective.

[0039] Optionally, it further includes:

[0040] Real-time monitor the surface temperature of the casting blank through an infrared thermometer and correct it according to the coefficient of thermal expansion of the casting blank material, so as to obtain a more accurate displacement measurement result.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The process method of the present invention is mainly used for the precise measurement of the width of casting blanks in the iron and steel metallurgy industry. By combining industrial CCD measurement technology with an AI model, the present invention can overcome problems such as low measurement accuracy and susceptibility to environmental influence in the prior art, and provide a more accurate and stable method for measuring the width of casting blanks. This method can not only improve the accuracy and efficiency of measurement, but also reduce the dependence on manual operation and reduce the error rate, thus meeting the requirements of the iron and steel metallurgy industry for high-precision measurement. In addition, the method of the present invention is also applicable to other industrial fields that require high-precision measurement and has a wide application prospect. Description of the Drawings

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0044] Figure 1 Schematic diagram of slab width measurement provided by the embodiment of the present invention.

[0045] Figure 2 Schematic diagram of size measurement principle of CCD optoelectronic imaging provided by the embodiment of the present invention. Detailed implementation manners

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] The object of the present invention is to provide a technical solution with high measurement accuracy and not affected by the environment.

[0048] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0049] Embodiment 1:

[0050] This embodiment provides a method for measuring the width of a continuous casting billet based on AI using a single-line array on the CCD focal plane. Taking the single-line array photosensitive element of the CCD as an example, the schematic diagram of the size measurement principle of CCD optoelectronic imaging is as Figure 2 shown. Assuming the distance between the single-line array photosensitive elements is P, the lens magnification of the CCD optoelectronic imaging system is:

[0051] Length L of the object to be measured x The length is:

[0052] L x = N x ·P / M

[0053] The resolution of the single-line array on the CCD focal plane is:

[0054] S = L t / N t

[0055] It can be seen from the above analysis that when the distance P between the single-line array photosensitive elements on the CCD focal plane and the lens magnification M of the CCD optoelectronic imaging system are determined, only by measuring the number of photosensitive elements (pixel number) of the object to be measured on the CCD, the size information of the object to be measured can be solved.

[0056] Based on the dimensional measurement principle of CCD photoelectric imaging, the auxiliary line laser emits a visible laser line toward the surface of the slab to be measured. The laser is reflected by the surface of the billet and then passes through the lens to the focal plane of the CCD detector. The image presented by the laser line at the boundary of the billet will be segmented / deflected. After the boundary points are extracted using the image processing algorithm, the distance from the lens to the billet surface is measured by the height measurement sensor. Then, based on the dimensional measurement principle of CCD photoelectric imaging, the billet width can be calculated. The schematic diagram of billet width measurement is shown in the figure below. Figure 1 shown.

[0057] An imaging CCD with a pixel size of 2.5μm × 2.5μm and a resolution of 5120 × 5120 is proposed. ROI sampling can be set based on the on-site segment spacing, and the maximum resolution across the slab width is 5120. Taking a CCD covering 3m across the slab width as an example, it can be calculated that the resolution of width measurement is better than 0.6mm.

[0058]

[0059] Basic principles:

[0060] The core principle of CCD width measurement (i.e., the technology of measuring the width of an object using an industrial CCD camera) is to convert the physical size of an object into pixel information through optical imaging and image processing technology, and then obtain the actual width through calibration and calculation. The following is a step-by-step explanation of its core principle:

[0061] Optical imaging principle:

[0062] CCD width measurement is based on geometric optical imaging and digital image analysis:

[0063] The CCD camera uses the lens to image the edge contour of the ingot onto the photosensitive surface of the CCD sensor.

[0064] If the width of the object is (W), the width of its projection on the imaging plane is (W'), according to the object-image relationship formula:

[0065]

[0066] Where (L) is the object distance (the distance from the object to the lens) and (f) is the focal length of the lens.

[0067] Image processing and edge detection

[0068] Extract edge pixel locations from an image using an algorithm:

[0069] 1. Edge detection:

[0070] Use grayscale gradient methods (such as Sobel and Canny operators) to locate the rough pixel position of the edge.

[0071] Sub-pixel subdivision: The edge positioning accuracy is improved to within 0.1 pixel through interpolation algorithms (such as Zernike moments and Gaussian fitting).

[0072] 2. Key parameters:

[0073] If the camera resolution is (N) pixels per field of view width (FOV), then the actual size represented by each pixel is:

[0074]

[0075] For example: 2000mm field of view width + 5000 pixel resolution → each pixel corresponds to 0.4mm, and the accuracy can reach ±0.04mm after sub-pixel subdivision.

[0076] Width calculation and calibration:

[0077] Static calibration:

[0078] Use a calibration block with a known size (such as a standard ruler) to determine the conversion coefficient between pixels and actual size.

[0079] Calibration formula:

[0080] Wactual = k·(X2 - X1) + b

[0081] Where, (x1, x2) are the pixel coordinates of the left and right edges, (k) is the calibration coefficient, and (b) is the compensation error.

[0082] Dynamic compensation:

[0083] If the casting speed is (v) and the camera exposure time is (t), then the error caused by motion blur is (Delta = v·t).

[0084] Eliminate the error through encoder-triggered synchronous exposure or image deblurring algorithm.

[0085] Environment and error compensation

[0086] Temperature compensation:

[0087] The thermal expansion of the casting will cause a change in the actual width. Compensation formula:

[0088] W 修正 = W 测量 / (1 + α·αT).

[0089] Where, (α) is the thermal expansion coefficient of the material.

[0090] ΔT is the change in the casting temperature;

[0091] Optical distortion correction:

[0092] Use a telecentric lens to reduce perspective distortion, or pre-calculate the distortion coefficient through a calibration plate and perform reverse correction.

[0093] Hardware system design:

[0094] CCD camera selection:

[0095] High-speed camera: The frame rate needs to be higher than the casting speed (for example, when the drawing speed is 2 m / min, the frame rate ≥ 50 fps).

[0096] Resolution: Select according to the displacement accuracy requirements (such as above 5 million pixels for sub-millimeter accuracy).

[0097] Trigger mode: External trigger synchronization (synchronized with the encoder or PLC to avoid missing frames).

[0098] Light source and marking:

[0099] Natural marking method: Use natural features such as scale oxide and scratches on the surface of the casting blank as tracking points.

[0100] Installation layout:

[0101] Camera position: Installed above the exit roller table of the continuous caster, and the field of view covers the surface of the casting blank.

[0102] Image processing and algorithms:

[0103] Feature point extraction:

[0104] Natural feature method: Use SIFT and ORB algorithms to extract the texture feature points of the scale oxide.

[0105] Optical flow method (Lucas-Kanade) to track the displacement of feature points in adjacent frames.

[0106] Displacement calculation:

[0107] Single-frame displacement: Calculate the actual displacement according to the pixel displacement and the calibration coefficient.

[0108] Δs = k·Δp

[0109] Δp: Pixel displacement of the feature point.

[0110] k: Calibration coefficient (mm / pixel).

[0111] Multi-frame filtering: Use Kalman filtering or moving average method to eliminate noise.

[0112] Calibration and error compensation:

[0113] Geometric calibration:

[0114] Pixel-actual size conversion: Place a standard scale (such as 1000 mm) in the field of view and calculate the actual length k corresponding to each pixel.

[0115] Dynamic error compensation:

[0116] Motion blur correction:

[0117] Freeze motion with a short exposure time (<1 ms) in combination with a high-frequency light source.

[0118] Use deconvolution algorithm to repair the blurred image.

[0119] Vibration compensation:

[0120] Install a shock-absorbing bracket or eliminate vibration displacement by referring to the background fixed point.

[0121] Temperature influence:

[0122] Thermal expansion correction: Correct the actual displacement according to the slab temperature (assistance of an infrared thermometer is required):

[0123] Δs actual = Δs measured (1 + α·ΔT);

[0124] α: Coefficient of thermal expansion of the slab material (α for steel ≈ 12×10-6 / °C);

[0125] ΔT: Change in slab temperature (°C).

[0126] In actual measurement, since the slab is just drawn out of the high-temperature mold, its surface temperature is relatively high and there is a thermal gradient, which will cause thermal expansion of the slab in the length direction. In order to accurately measure the actual width of the slab, this thermal expansion effect must be considered. Therefore, the present invention monitors the surface temperature of the slab in real time through an infrared thermometer and corrects it according to the coefficient of thermal expansion of the slab material, so as to obtain a more accurate displacement measurement result. This method can significantly improve the accuracy and stability of measurement, especially in the steel metallurgy environment with high temperature and high dust, and has significant technical advantages.

[0127] Based on the above principle, the industrial CCD method can achieve high-precision, non-contact real-time measurement of the casting speed of the continuous caster, providing key data support for the optimization of the continuous casting process. The application of this technology not only improves the accuracy and efficiency of measurement, but also reduces the interference to the production process, making the operation of the continuous caster more stable and efficient. In addition, the real-time obtained data is of great significance for timely adjusting process parameters and optimizing product quality, thus providing strong technical support for the quality control and cost saving of the entire production process.

[0128] In the present specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0129] In this article, specific examples are used to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method of the present invention and its core idea. At the same time, for those of ordinary skill in the art, based on the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation on the present invention.

Claims

1. A method for measuring the width of a continuous casting slab based on AI using an industrial CCD, characterized in that, Including: The auxiliary line laser projects a visible laser line onto the surface of the slab to be measured. After the laser is reflected by the surface of the billet, it reaches the focal plane of the CCD detector through the lens. The image of the laser line presented at the billet boundary shows segmentation / deflection. After using the image processing algorithm to extract the boundary points, the distance from the lens to the billet surface is measured by the height measurement sensor, and then the width of the billet can be calculated according to the size measurement principle of CCD optoelectronic imaging.

2. The method for measuring the width of a continuous casting slab based on AI according to claim 1, wherein, The camera used for image acquisition is a high-speed camera, with a frame rate higher than the moving speed of the continuous casting billet; the resolution is selected according to the displacement accuracy requirements; the trigger mode is external trigger synchronization; the camera is installed above the outlet roller table of the continuous casting machine, and the field of view covers the surface of the continuous casting billet; the camera resolution is N, the pixel / field of view width is FOV, and the actual size represented by each pixel is:

3. The method for measuring the width of a continuous casting slab based on AI according to claim 1, wherein Feature point marking and extraction adopt natural feature method and optical flow method. The natural feature method uses SIFT and ORB algorithms to extract the texture feature points of the scale, and the optical flow method tracks the displacement of the feature points in adjacent frames.

4. The method for measuring the width of a continuous casting slab based on AI according to claim 1, wherein CCD width measurement is based on geometric optical imaging and digital image analysis. The CCD camera images the edge contour of the billet through the lens onto the photosensitive surface of the CCD sensor, and the object-image relationship formula is: Where, W is the width of the object, W’ is the projected width on the imaging plane, L is the distance from the object to the lens, and f is the focal length of the lens.

5. The method for measuring the width of a continuous casting slab based on AI according to claim 1, wherein The extraction of the boundary points by using the image processing algorithm includes: Using the gray gradient method to locate the rough pixel position of the edge; Improving the edge positioning accuracy to within 0.1 pixel through the difference algorithm.

6. The method for measuring the width of a continuous casting billet based on AI industrial CCD according to claim 1, wherein It also includes calibration and dynamic error compensation. Calibration includes static calibration, using a calibration block with a known size to determine the conversion coefficient between pixels and actual size; calibration formula: Wactual = k·(X2 - X1) + b; Where, X1 and X2 are the pixel coordinates of the left and right edges, k is the calibration coefficient, and b is the compensation error; Dynamic error compensation includes motion blur correction and vibration compensation; Motion blur correction: short exposure time combined with a high-frequency light source to freeze the motion; The deconvolution algorithm is used to repair the blurred image; Vibration compensation: install a shock-absorbing bracket, or eliminate the vibration displacement by referring to the fixed point of the background; If the moving speed of the billet is v and the exposure time of the camera is t, the error caused by motion blur is Delta = v·t; Eliminate the error through encoder-triggered synchronous exposure or image deblurring algorithm.

7. The method for measuring the width of a continuous casting slab based on AI industrial CCD according to claim 1, characterized in that It also includes: Temperature compensation and optical distortion correction: Temperature compensation: The thermal expansion of the billet will cause a change in the actual width, and the compensation formula is: W 修正 = W 测量 / (1 + α·ΔT); Where, α is the thermal expansion coefficient of the material; Optical distortion correction is to use a telecentric lens to reduce perspective distortion, or pre-calculate the distortion coefficient through a calibration plate and perform reverse correction.

8. The method for measuring the width of a continuous casting billet based on AI industrial CCD according to claim 1, wherein, It also includes: The surface temperature of the billet is monitored in real time by an infrared thermometer and corrected according to the thermal expansion coefficient of the billet material, so as to obtain a more accurate displacement measurement result.

Images