ONLINE QUANTITATIVE ASSESSMENT METHODS AND SYSTEMS FOR ACID WASHING QUALITY BASED ON MACHINE VISION

Abstract

This invention belongs to the field of cold rolling line processing technology, specifically referring to a method and system for online quantitative assessment of acid wash quality based on machine vision. The method includes: using tension rollers to form a pre-set wrap angle in the imaging area to suppress vibration of the steel strip; using lights to illuminate the surface of the steel strip; and a camera collecting images of the steel strip's surface. The surface image data captured by the camera is collected via a data acquisition card; defects such as insufficient and excessive acid wash are identified based on the surface image data, and the acid wash quality level is determined and displayed on the screen. Based on machine vision, the online assessment of the acid wash quality of the steel strip is performed with good real-time compliance, high objectivity, and good consistency. Furthermore, the formation of a specific wrap angle in the imaging area through the tension roller suppresses the vibration of the steel strip, resulting in clearer image quality.This study can statistically analyze the acid-wash quality of steel strips of different types, specifications, and acid-washing technology parameters, propose improvements to the acid-washing technology, and address existing problems in manual acid-washing quality assessment, with broad applicability.

Description

An online quantitative evaluation method and system for pickling quality based on machine vision Technical Field

[0001] The present disclosure relates to the technical field of cold rolling processing lines, and in particular to an online quantitative evaluation method and system for pickling quality based on machine vision. Background Art

[0002] In the cold-rolled strip pickling unit, different process parameters such as acid temperature and acid concentration are set for steel coils of a certain type and a certain coiling temperature. The final pickling endpoint time is different, the pickling quality effect is also different, and the pickling cost is also different. The current pickling parameter settings are mainly based on the operating procedures. In order to avoid under-pickling, the operating procedures set the acid temperature and acid concentration values ​​too high. On the one hand, it leads to a large consumption of energy media and increases the operating cost. On the other hand, it will lead to more over-pickling and increase the pickling iron loss.

[0003] One of the reasons is that the current pickling quality evaluation is based on manual qualitative evaluation. The pickling quality is sampled, qualitative, and subjectively evaluated by the human eye. The pickling quality is visually inspected at intervals, which has a large lag and only roughly distinguishes between under-pickling and over-pickling. The severity of under-pickling and over-pickling is not subdivided, and no quantitative indicators are formed, making it difficult to effectively guide the quantitative adjustment of pickling process parameters.

[0004] Summary of the Invention

[0005] The present disclosure aims to solve at least one of the technical problems existing in the prior art, and proposes an online quantitative evaluation method and system for pickling quality based on machine vision.

[0006] In a first aspect, the present disclosure provides an online quantitative evaluation method for pickling quality based on machine vision, comprising:

[0007] Use tension rollers to form a preset wrap angle in the shooting area to suppress strip vibration;

[0008] A light source is used to project light onto the surface of the strip, and a camera collects the surface image of the strip;

[0009] The surface image data of the strip captured by the camera is acquired through a data acquisition card;

[0010] Under-pickling and over-pickling defects are identified based on the surface image data, a pickling quality grade is given, and a screen prompt is provided.

[0011] Preferably, the light source is a strip LED light source, and the camera uses a hyperspectral band.

[0012] Preferably, the identifying of under-pickling and over-pickling defects based on the surface image data and providing a pickling quality grade specifically includes:

[0013] Under-pickling and over-pickling defects are identified based on the surface image data, and characteristic values ​​of defect quantity, defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor and defect grayscale are counted, and the pickling quality grade is given based on the characteristic values.

[0014] Preferably, the pickling quality grade is given according to the characteristic value, specifically including:

[0015] A quantitative evaluation model for pickling quality is constructed based on the AHP method to quantitatively evaluate the pickling quality of strip steel; the quantitative evaluation model for pickling quality includes a target layer O, a criterion layer D, a criterion layer C, and a solution layer P;

[0016] The target layer O is used to display the comprehensive quality evaluation output result;

[0017] The criterion layer D is used to store different defect levels, and each defect level corresponds to a level weight;

[0018] The criteria layer C is used to store different defect types, each defect type corresponds to a defect right;

[0019] The solution layer P is used to store different steel plates.

[0020] Preferably, the target layer O includes: Σ 得分 =Σ 缺陷权 ×Level Rights

[0021] Among them, Σ 得分 is the comprehensive quality evaluation output result, Σ 缺陷权 is the right of the identified steel plate defect, and the grade right is the grade of the identified steel plate defect.

[0022] Preferably, the level weight value is determined as follows:

[0023] Construct the level weight judgment matrix A;

[0024] Solve the maximum eigenvalue of the level weight judgment matrix A and perform consistency test.

[0025] Preferably, the level weight judgment matrix A is as follows:

[0026] Among them, a ij >0,a ji =1 / a ij , a ii =1, i, j = 1, 2, ... n, n is a positive integer, a ijThe value of is the defect level weight factor, and its size is the value of the pairwise comparison of defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor and defect grayscale eigenvalue.

[0027] Preferably, solving the maximum eigenvalue of the level weight judgment matrix A specifically includes:

[0028] (1) Normalize each column of the level weight judgment matrix A;

[0029] (2) The normalized judgment matrix of each column is summed row by row;

[0030] (3) Normalize the vector obtained after summation, and the result W is the desired eigenvector, ω i is the corresponding defect level weight factor eigenvalue, as shown below:

[0031] Where k is the corresponding column number.

[0032] In a second aspect, the present disclosure provides an online quantitative evaluation system for pickling quality based on machine vision. The system can be used to implement an online quantitative evaluation method for pickling quality based on machine vision. The system includes: a tension roller, a light source, a camera, and pickling quality quantitative evaluation software. The specific principles are as follows:

[0033] Use tension rollers to form a preset wrap angle in the shooting area to suppress strip vibration;

[0034] A light source is used to project light onto the surface of the strip, and a camera collects the surface image of the strip;

[0035] The surface image data of the strip captured by the camera is acquired through a data acquisition card;

[0036] Under-pickling and over-pickling defects are identified based on the surface image data, a pickling quality grade is given, and a screen prompt is provided.

[0037] In a third aspect, the present disclosure provides an electronic device, comprising:

[0038] one or more processors;

[0039] a memory for storing one or more programs;

[0040] When the one or more programs are executed by the one or more processors, the one or more processors implement an online quantitative evaluation method for pickling quality based on machine vision. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] FIG1 is a flow chart of a method for online quantitative evaluation of pickling quality based on machine vision provided by an embodiment of the present disclosure;

[0042] FIG2 is a block diagram of the principle of online quantitative evaluation of pickling quality based on machine vision provided by an embodiment of the present disclosure;

[0043] FIG3 is a diagram of a quantitative evaluation model for pickling quality of steel strips based on AHP according to an embodiment of the present disclosure;

[0044] FIG4 is a structural block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the technical solutions of the present disclosure, the present disclosure is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0046] Unless otherwise defined, the technical or scientific terms used in this disclosure should have the usual meanings understood by people with ordinary skills in the field to which this disclosure belongs. The words "first", "second" and similar terms used in this disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one", "an" or "the" are not limited to quantity, but rather to the presence of at least one. Words such as "include" or "comprise" mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used for relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0047] In each accompanying drawing, identical elements are represented by similar reference numerals. For the sake of clarity, the various parts in the accompanying drawings are not all drawn to scale. In addition, some well-known parts may not be shown in the drawings.

[0048] Many specific details of the present disclosure are described below, such as component structures, materials, dimensions, processing techniques, and technologies, to provide a clearer understanding of the present disclosure. However, as will be appreciated by those skilled in the art, the present disclosure may be implemented without following these specific details.

[0049] As shown in Figures 1 and 2, a machine vision-based online quantitative evaluation method for pickling quality is used in a machine vision-based online quantitative evaluation system for pickling quality. The machine vision-based online quantitative evaluation system for pickling quality includes: a tension roller C1, a steel strip C2, a light source C3, a camera lens C4, a network cable C5, a data acquisition card C6, an image processing server C7, and pickling quality quantitative evaluation software C8.

[0050] The method specifically includes the following steps:

[0051] S1, using the tension roller to form a preset wrap angle in the shooting area to suppress the vibration of the strip;

[0052] S2, using a light source to project light onto the surface of the strip, and a camera to capture the surface image of the strip;

[0053] S3, acquiring the surface image data of the strip captured by the camera through a data acquisition card;

[0054] S4, identifying under-pickling and over-pickling defects based on the surface image data, giving a pickling quality grade, and providing a screen prompt.

[0055] It should be noted that, in the above steps S1-S4, the order of each step is not limited and can be arbitrarily disrupted.

[0056] Tension roller C1 forms a certain wrap angle within the imaging area, suppressing vibrations in the steel strip C2 and ensuring clear imaging quality. Light source C3 projects light onto the surface of the steel strip C2, and camera lens C4 captures the surface image of the steel strip C2. A network cable C5 connects camera lens C4 to data acquisition card C6, which is installed in the PCI slot of image processing server display C7. Pickling quality quantitative evaluation software C8 runs on image processing server display C7.

[0057] The camera lens C4 captures the surface image of the steel strip C2. The image processing server C7 is connected to the camera lens C4 via the data acquisition card C6 and the network cable C5 to obtain the surface image of the steel strip captured by the camera lens C4. The pickling quality quantitative evaluation software C8 uses an image processing algorithm to identify under-pickling and over-pickling defects from the high-resolution image of the steel strip surface, and counts characteristic values ​​such as the number of defects, defect area, defect length, defect perimeter, slenderness ratio of the defect area, defect shape factor and defect grayscale, to give a pickling quality grade. The pickling quality quantitative evaluation software C8 also provides a screen prompt to provide data support for the adjustment of process parameters in the pickling process section.

[0058] Taking into account the service life, brightness, uniformity and price, the strip LED light source is selected as the system light source C3.

[0059] The camera lens C4 uses a hyperspectral band, which has a high resolution and can detect subtle features of matter with a high degree of precision.

[0060] The network cable C5 uses Cameralink network cable for high-speed image transmission, and the image processing server C7 is equipped with a high-performance GPU. The image processing algorithm identifies under-pickling and over-pickling defects from high-resolution images of the strip surface.

[0061] Run the pickling quality quantitative evaluation software C8. The data acquisition and visualization module of the pickling quality quantitative evaluation software C8 collects the surface image of the strip, stores it in the hard disk of the image processing server C7, and displays it visually on the monitor.

[0062] The main components of the pickling quality quantitative evaluation software C8 include: data acquisition visualization module, pickling defect recognition module, and pickling quality quantitative evaluation module.

[0063] The data acquisition and visualization module reads strip surface images captured by a camera via a CameraLink network cable, stores them on the image processing server's hard drive, and displays them on the server's monitor. The pickling defect recognition module reads high-resolution images of the strip surface and uses image processing algorithms to detect under-pickling and over-pickling defects. It also stores characteristic value data such as defect number, defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor, and defect grayscale.

[0064] The pickling quality quantitative evaluation software C8 builds a quantitative pickling quality evaluation model based on the AHP method to quantitatively assess the pickling quality of steel strips, as shown in Figure 3. The model comprises a target layer O, a criterion layer D, a criterion layer C, and a solution layer P. The target layer O displays the comprehensive quality evaluation output; the criterion layer D stores different levels; the criterion layer C stores different defect types; and the solution layer P stores different steel plates.

[0065] Target layer O uses a final quality score as an indicator to measure the pickling quality of cold-rolled strip. Target layer O includes:

[0066] Σ score = Σ defect weight × level weight

[0067] The Σ score is the comprehensive quality evaluation output, the Σ defect weight is the identified steel plate defect weight, and the grade weight is the identified steel plate grade. This can be used to quantitatively assess the pickling quality level of cold-rolled strip.

[0068] The criterion layer D is the different levels of under-pickling and over-pickling defects. Each level has a weight. The defect level weight mainly considers the defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor and defect grayscale characteristic value.

[0069] The criterion layer C represents the under-pickling and over-pickling defect types. These defects are assigned defect weights, which primarily consider the degree of impact of the defects on pickling quality. These defect weights are calculated by constructing a defect judgment matrix. The solution layer P represents each roll of steel with under-pickling and over-pickling defects, containing the raw data for all pickling quality defects for that roll. The difficulty lies in determining the level weights, which are determined by constructing a judgment matrix, finding the maximum eigenvalue of the judgment matrix, and performing a consistency check.

[0070] During the production process, image acquisition is carried out continuously to collect strip pickling images under different steel grades, different specifications, different strip speeds and different pickling process parameters.

[0071] Browse a large number of collected images, select suitable and sufficient images as samples and mark them. Based on the marked sample images, select Yolo, PicoDet, or FastRCNN model for training until the error is controlled within the allowable range.

[0072] The difficulty lies in determining the value of the level weight. This is determined by constructing a level weight judgment matrix A, finding the maximum eigenvalue of the level weight judgment matrix A, and performing a consistency test. The level weight judgment matrix A is as follows.

[0073] Among them, a ij >0,a ji =1 / a ij , a ii =1, i, j = 1, 2, ... n, n is a positive integer, a ij The value is the defect level weight factor, and its size is the ratio of the defect characteristic value i to the defect characteristic value j. Specifically, its size is the value of the pairwise comparison of the defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor and defect grayscale characteristic value. In a specific implementation scenario, i, j represent the variable numbers related to the defect level weight, such as defect area 1, defect length 2, defect perimeter 3, defect area slenderness ratio 4, defect shape factor 5 and defect grayscale characteristic value 6. These six variables have 6 numbers, for example, a 12 It represents the ratio of the defect area 1 to the defect length 2.

[0074] The process of calculating eigenvalues ​​using the sum-product method is as follows:

[0075] (1) Normalize each column of the level weight judgment matrix A;

[0076] (2) The normalized judgment level weight judgment matrix A of each column is summed row by row;

[0077] (3) Normalize the vector obtained after summation, and the result W is the desired eigenvector, ω i is the corresponding defect level weight factor eigenvalue, as shown below, and k is the corresponding column number.

[0078] Where k is the corresponding column number of the level weight judgment matrix A. The above matrix has rows and columns, and k is the column number of the matrix. For example, the above judgment matrix A has 6 rows and 6 columns, and the range of k is 1 to 6.

[0079] Based on the online quantitative evaluation system for pickling quality, the incoming strip material characteristics and pickling process parameters are comprehensively considered to statistically analyze the pickling quality of strip steel under different steel grades, different specifications and different pickling process parameters, and provide suggestions for improving the pickling process. Beneficial effects:

[0080] The system is based on machine vision and can realize online evaluation of strip pickling quality with good real-time performance;

[0081] The system uses a hyperspectral camera for imaging. The hyperspectral camera has a rich band composition and high resolution, and can detect subtle features of materials with a high degree of precision.

[0082] The system uses a camera to capture images of the strip surface. It identifies and detects under-pickling and over-pickling defects from high-resolution images of the strip surface and stores characteristic value data such as defect number, defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor, and defect grayscale. This system is highly objective and consistent.

[0083] The system uses the data on all the under-pickling and over-pickling defects and defect levels of each coil of strip steel to build a strip pickling quality evaluation model based on the AHP method to quantitatively evaluate the strip pickling quality.

[0084] The system forms a certain wrap angle in the shooting area through the tension roller to suppress the vibration of the strip and ensure clear image quality.

[0085] The present disclosure also provides an online quantitative evaluation system for pickling quality based on machine vision. The system can be used to implement the above-mentioned online quantitative evaluation method for pickling quality based on machine vision. The system includes: a tension roller, a light source, a camera, and pickling quality quantitative evaluation software. The specific principles are as follows:

[0086] S1, using the tension roller to form a preset wrap angle in the shooting area to suppress the vibration of the strip;

[0087] S2, using a light source to project light onto the surface of the strip, and a camera to capture the surface image of the strip;

[0088] S3, acquiring the surface image data of the strip captured by the camera through a data acquisition card;

[0089] S4, identifying under-pickling and over-pickling defects based on the surface image data, giving a pickling quality grade, and providing a screen prompt.

[0090] FIG4 is a schematic diagram of an embodiment of an electronic device according to an embodiment of the present invention. As shown in FIG4 , an embodiment of the present invention provides an electronic device 1300, comprising a memory 1310, a processor 1320, and a computer program 1311 stored in the memory 1310 and executable on the processor 1320. When the processor 1320 executes the computer program 1311, the following steps are implemented: S1, using a tension roller to form a preset wrap angle in the shooting area to suppress vibration of the strip;

[0091] S2, using a light source to project light onto the surface of the strip, and a camera to capture the surface image of the strip;

[0092] S3, acquiring the surface image data of the strip captured by the camera through a data acquisition card;

[0093] S4, identifying under-pickling and over-pickling defects based on the surface image data, giving a pickling quality grade, and providing a screen prompt.

[0094] It should be noted that the computer-readable medium described in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or component. In the present disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wireline, optical fiber cable, RF, or any suitable combination thereof.

[0095] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the aforementioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0096] It will be understood that the above embodiments are merely exemplary embodiments for illustrating the principles of the present invention, and the present invention is not limited thereto. Those skilled in the art will appreciate that various modifications and improvements can be made without departing from the spirit and substance of the present invention, and such modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for online quantitative evaluation of pickling quality based on machine vision, characterized in that: include: Use tension rollers to form a preset wrap angle in the shooting area to suppress strip vibration; A light source is used to project light onto the surface of the steel strip, and a camera collects the surface image of the steel strip; The surface image data of the strip captured by the camera is acquired through a data acquisition card; According to the surface image data, under-pickling and over-pickling defects are identified, a pickling quality grade is given, and a screen prompt is given.

2. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 1 is characterized in that: The light source is a strip LED light source, and the camera uses a high-spectrum band.

3. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 1 is characterized in that: The method of identifying under-pickling and over-pickling defects according to the surface image data and providing a pickling quality grade specifically includes: According to the surface image data, under-pickling and over-pickling defects are identified, and characteristic values ​​of defect quantity, defect area, defect length, defect perimeter, defect area slenderness ratio, defect shape factor and defect grayscale are counted, and the pickling quality grade is given according to the characteristic values.

4. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 3 is characterized in that: The pickling quality grade is given according to the characteristic value, specifically including: A quantitative evaluation model for pickling quality is constructed based on the AHP method to quantitatively evaluate the pickling quality of strip steel; the quantitative evaluation model for pickling quality includes a target layer O, a criterion layer D, a criterion layer C and a solution layer P; The target layer O is used to display the comprehensive quality evaluation output result; The criterion layer D is used to store different defect levels, each defect level corresponds to a level weight; The criterion layer C is used to store different defect types, each defect type corresponds to a defect right; The solution layer P is used to store different steel plates.

5. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 4 is characterized in that: The target layer O includes: Σ 得分 =Σ 缺陷权 ×Level Rights Among them, Σ 得分 is the comprehensive quality evaluation output result, Σ 缺陷权 is the right of the identified steel plate defect, and the grade right is the grade of the identified steel plate defect.

6. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 4 is characterized in that: The value determination process of the level weight is as follows: Construct the level weight judgment matrix A; Solve the maximum eigenvalue of the level weight judgment matrix A and perform a consistency check.

7. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 6 is characterized in that: The level weight judgment matrix A is as follows: Among them, a ij >0,a ji =1 / a ij , a ii =1, i, j = 1, 2, ... n, n is a positive integer, a ij The value of is the defect level weight factor, which is the ratio of the defect characteristic value i to the defect characteristic value j.

8. The on-line quantitative evaluation method for pickling quality based on machine vision according to claim 7 is characterized in that: The step of solving the maximum eigenvalue of the level weight judgment matrix A specifically includes: (1) Normalize each column of the level weight judgment matrix A; (2) The normalized judgment level weight judgment matrix A is summed row by row in each column; (3) The vector obtained after summation is normalized again, and the result W is the desired eigenvector, ω i is the corresponding defect level weight factor characteristic value, as shown below: Among them, k is the corresponding column number of the level weight judgment matrix A.

9. An online quantitative evaluation system for pickling quality based on machine vision, characterized in that: The system can be used to implement the online quantitative evaluation method of pickling quality based on machine vision as described in any one of claims 1 to 8 above. The system includes: a tension roller, a light source, a camera and a quantitative evaluation software for pickling quality. The specific principle is as follows: Use tension rollers to form a preset wrap angle in the shooting area to suppress strip vibration; A light source is used to project light onto the surface of the steel strip, and a camera collects the surface image of the steel strip; The surface image data of the strip captured by the camera is acquired through a data acquisition card; According to the surface image data, under-pickling and over-pickling defects are identified, a pickling quality grade is given, and a screen prompt is given.

10. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the online quantitative evaluation method for pickling quality based on machine vision as described in any one of claims 1 to 8.

Images