CN115861304A - Method and system for detecting steel strip-shaped structure based on image processing - Google Patents
Method and system for detecting steel strip-shaped structure based on image processing Download PDFInfo
- Publication number
- CN115861304A CN115861304A CN202310132215.8A CN202310132215A CN115861304A CN 115861304 A CN115861304 A CN 115861304A CN 202310132215 A CN202310132215 A CN 202310132215A CN 115861304 A CN115861304 A CN 115861304A
- Authority
- CN
- China
- Prior art keywords
- image
- banded
- steel
- structure information
- rating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 107
- 239000010959 steel Substances 0.000 title claims abstract description 107
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000002436 steel type Substances 0.000 claims abstract description 76
- 238000005070 sampling Methods 0.000 claims abstract description 49
- 238000001514 detection method Methods 0.000 claims abstract description 22
- 238000013139 quantization Methods 0.000 claims abstract description 17
- 238000004364 calculation method Methods 0.000 claims abstract description 12
- 238000013145 classification model Methods 0.000 claims abstract description 6
- 230000008520 organization Effects 0.000 claims abstract description 6
- 238000001914 filtration Methods 0.000 claims description 14
- 238000003708 edge detection Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 11
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 230000009467 reduction Effects 0.000 claims description 7
- 230000000149 penetrating effect Effects 0.000 claims description 6
- 230000002452 interceptive effect Effects 0.000 claims description 3
- 229910001562 pearlite Inorganic materials 0.000 description 17
- 229910000859 α-Fe Inorganic materials 0.000 description 17
- 238000000605 extraction Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 239000000463 material Substances 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 239000012535 impurity Substances 0.000 description 3
- 238000001000 micrograph Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000001629 suppression Effects 0.000 description 3
- 241001085205 Prenanthella exigua Species 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 230000035515 penetration Effects 0.000 description 2
- 238000011002 quantification Methods 0.000 description 2
- 238000005452 bending Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002542 deteriorative effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The application provides a method and a system for detecting a steel strip-shaped structure based on image processing, wherein the method comprises the following steps: acquiring a sampling image of a sample block to be graded to obtain a first steel grade image; extracting strip-shaped structure information in the first steel type image according to the color features in the first steel type image; performing a quantization calculation on the strip-shaped organization structure information to obtain quantization structure information; inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result; the banded structure quantitative rating model is a classification model constructed according to banded structure information of preset different types of steel. The digital image-based detection method for the steel strip-shaped structure can continuously detect and improve the grading efficiency.
Description
Technical Field
The application belongs to the technical field of metal microstructure analysis, and particularly relates to a steel strip structure detection method and system based on image processing.
Background
Digital image processing is also called computer image processing, which refers to a process of converting an image signal into a digital signal and processing it with a computer. Digital image processing has found widespread use in many fields.
The grade of the band structure in steel may reflect a strong hardness defect characteristic inside the steel. Since the band structure is commonly present in various types of steel materials and significantly affects the mechanical properties of the steel, deteriorating the cold and hot workability of materials and the fatigue life of parts, products requiring high quality require accurate inspection and rating of the band structure of the steel.
The rating of the banded structure refers to the rating of how significant ferrite and pearlite appear as banded features in the image, respectively. The banding patterns exhibited by different steel grades also differ. For example, the steel with the mark HL610A has white pearlite and black ferrite in the photomicrograph. For example, the steel with the mark of 42CrMo-N2 has bright white pearlite and greenish black ferrite in the micrograph. The steel with the grade 1E0170 has lighter pearlite and darker gray ferrite in the micrograph. The 20 g steel showed lighter gray pearlite and black ferrite in the micrograph. The steel mills are rated manually, inspectors shoot sample images by using a microscope firstly, and then compare and analyze experimental images and standard images manually according to rating standard files of banded structures, so that the rating is further realized, and the efficiency is low.
Disclosure of Invention
The application provides a method and a system for detecting a steel strip-shaped structure based on image processing, which aim to solve the problem that the efficiency of manually grading the steel strip-shaped structure is low.
In one aspect, the present application provides a method for detecting a strip-shaped structure in steel based on image processing, including:
acquiring a sampling image of a sample block to be graded to obtain a first steel grade image;
extracting banded texture structure information in the first steel type image according to the color features in the first steel type image;
performing a quantization calculation on the strip-shaped organization structure information to obtain quantization structure information;
inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result; the banded structure quantitative rating model is a classification model constructed according to banded structure information of preset different types of steel.
Optionally, the step of obtaining the sample image of the sample block to be evaluated includes:
scanning the sample block to be evaluated to obtain a sampling image of the sample block to be evaluated and the steel type of the sample block to be evaluated; calling a rating standard file corresponding to the steel type;
if the sampling image corresponding to the steel type accords with the rating standard file, generating a first steel type image;
and if the sampling image corresponding to the steel type does not accord with the rating standard file, rescanning the sample block to be rated.
Optionally, after the step of calling the rating standard file corresponding to the steel type, the method further includes: detecting sampling parameters of the sampling image, wherein the sampling parameters comprise contrast ratio, brightness, reduction color and resolution; analyzing parameter intervals in the rating standard file, wherein the parameter intervals comprise a contrast rate interval, a brightness interval, a reduction color interval and a resolution interval;
if all the sampling parameters are in the corresponding parameter intervals, determining that the sampling images corresponding to the steel types conform to the rating standard file;
if any one of the sampling parameters is not in the corresponding parameter interval, determining that the sampling image corresponding to the steel type does not conform to the rating standard file;
optionally, the step of extracting the banded texture structure information in the first steel grade image according to the color feature in the first steel grade image includes: performing color space conversion on the first steel grade image according to the first steel grade image to obtain a second steel grade image; filtering the interference structure in the second steel type image to obtain a third steel type image; and extracting the banded structure information in the third steel grade image.
Optionally, performing color space conversion on the first steel grade image includes: and converting the RGB color space of the first steel type image into HSV color space.
Optionally, the step of filtering the interference structure in the second steel type image to obtain a third steel type image includes: setting a characteristic threshold value of the color of the banded tissue; verifying the second steel type image according to the RGB color feature, the HSV color feature and the feature threshold; marking structures in the second steel grade image where the RGB color features and/or the HSV color features do not meet the feature threshold; the structure which does not meet the characteristic threshold value is an interference structure; and filtering the interference structure in the second steel grade image to obtain a third steel grade image.
Optionally, the extracting of the banded structure information in the third steel type image includes: determining the edge contour of the banded structure in the image of the third steel grade by using a Canny edge detection method; marking the banded texture region in the third steel grade image according to the edge profile.
Optionally, the step of performing a quantitative calculation on the strip-shaped tissue structure information includes: creating a sliding window, wherein the length of the sliding window is the length of the image of the third steel grade, and the width of the sliding window is set according to the type of the steel grade; traversing the banded tissues in the third steel grade image to acquire the contour structure information of the banded tissues; the contour structure information comprises an area value, a length value and a width value of the banded tissue; recording the outline structure information into a storage unit of the sliding window; and traversing the storage unit to obtain the quantization structure information.
Optionally, the step of inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result includes: calculating all band-shaped tissue characteristics in the third steel grade image according to the quantitative structure information; classifying the banded tissues into penetrating banded tissues, semi-penetrating banded tissues and non-banded tissues according to the characteristics of the banded tissues, and respectively counting the number of the banded tissues under each classification; and inputting the number of the banded tissues to generate a rating result through the banded tissue rating model.
In another aspect, the present application provides an image processing-based steel strip structure detection system, comprising; the device comprises an image preprocessing unit, a banded tissue identification unit, a banded tissue characteristic rating unit and a rating result output unit;
the image preprocessing unit is used for: acquiring a sampling image of a sample block to be graded, and preprocessing the sampling image;
the strip-shaped tissue identification unit is used for: extracting band-shaped organization structure information in the first steel type image;
the banded tissue characteristic rating unit is for: according to the banded tissue structure information, performing quantization calculation on the banded tissue structure information to obtain quantized structure information;
the rating result output unit is used for: and inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result.
As can be seen from the above technical solutions, the present application provides a method and a system for detecting a steel strip structure based on image processing, the method including: acquiring a sampling image of a sample block to be graded to obtain a first steel grade image; extracting strip-shaped structure information in the first steel type image according to the color features in the first steel type image; performing a quantization calculation on the strip-shaped organization structure information to obtain quantization structure information; inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result; the banded structure quantitative rating model is a classification model constructed according to banded structure information of preset different types of steel. The digital image detection method for the steel strip-shaped structure can be used for continuously detecting and improving the grading efficiency.
Drawings
In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments are briefly described below, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a flow chart of a method for detecting a strip-shaped structure in steel according to an embodiment of the present application;
FIG. 2 is a schematic view of a sliding window provided in an embodiment of the present application;
FIG. 3 is a schematic structural diagram of a steel strip structure detection system provided in an embodiment of the present application;
FIG. 4 is a schematic diagram of an image preprocessing flow provided in an embodiment of the present application;
fig. 5 is a schematic view of a Canny edge detection process provided in the embodiment of the present application.
Detailed Description
In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The existence of the strip-shaped structure can cause the uneven structure of the steel, influence the performance of the steel, form anisotropy, reduce the plasticity, impact toughness and section shrinkage rate of the steel, cause the bad results of improper cold bending, high stamping rejection rate and easy deformation of the steel during heat treatment, and influence the service life of the steel. In order to express the performance of the steel material, the steel material may be graded in a band structure according to the band structure. When the banded tissue is graded, the most serious visual field of the banded tissue on the test surface is firstly found, and then the banded tissue grade is obtained by comparing the standard grading atlas manually. However, the judgment process is inefficient when compared with the standard rating map.
The method aims to solve the problems of strong subjectivity, high labor intensity and low efficiency of manual grading of the strip-shaped structure in the steel. The application provides a method for detecting a strip-shaped structure in steel based on image processing, which comprises the following steps of:
s10: and acquiring a sampling image of the sample block to be graded.
The step of obtaining the sampling image of the sample block to be graded comprises the steps of scanning the sample block to be graded by controlling a metallographic detection control platform, obtaining the sampling image and evaluating the quality of the sampling image.
Specifically, after opening metallographic examination control platform, metallographic examination control platform can automatic control microscope take a picture to the sample piece on the scanning charging tray to the sampling image and the steel grade information that will shoot reach this detecting system, this detecting system can carry out the analysis to the sampling image of uploading, judges whether sampling image quality accords with the rating standard.
And (4) because the types of the steel grades with the grading sample blocks are different, judging whether the sampled image meets the grading standard or not according to different conditions, and calling a grading standard file corresponding to the steel grade type according to the steel grade information.
If the sampling image corresponding to the steel type meets the rating standard file, generating a first steel type image;
and if the sampling image corresponding to the steel type does not accord with the rating standard file, scanning the sample block to be rated again.
The method comprises the steps of calling a rating standard file corresponding to the steel type, and detecting sampling parameters of a sampled image, wherein the sampling parameters comprise contrast ratio, brightness, reduction color and resolution; selecting a corresponding rating standard file according to the type of steel, and then analyzing parameter intervals in the rating standard file, wherein the parameter intervals comprise a contrast rate interval, a brightness interval, a reduction color interval and a resolution interval;
if all items in the sampling parameters of the sampling image are in the corresponding parameter interval, determining that the sampling image corresponding to the steel type meets the rating standard file;
if any one of the sampling parameters is not in the corresponding parameter interval, determining that the sampling image corresponding to the steel type does not conform to the rating standard file;
in this embodiment, the metallographic detection control platform can sequentially scan sample blocks to be graded on six trays, and send the sampled images and steel type information to the detection system, and the detection system performs image quality analysis after receiving the sampled images, for example, when the received images are completely black, the contrast ratio, brightness, reduction color and resolution ratio of the received images do not meet the grading standard file of the corresponding steel type, and the system can automatically send information which does not meet the grading requirement to the metallographic detection control platform, and continue to read the next sampled image.
S20: preprocessing an image;
after acquiring the adopted image, the sampled image needs to be preprocessed, as shown in fig. 4, the preprocessing of the image includes the following steps:
s201: performing color space conversion on the first steel grade image according to the first steel grade image to obtain a second steel grade image;
s202: filtering the interference structure in the second steel type image to obtain a third steel type image;
s203: and extracting the banded structure information in the third steel grade image.
Specifically, before extracting the banded structure through the color features of pearlite different from ferrite, the color space of the first steel type image needs to be converted to achieve a better extraction effect.
The image has three channels of red (R), green (G) and blue (B), each channel consisting of different values (0-255), which constitutes a multicolored image, called the color space of the image. In image processing, the HSV space has better separability and operability.
HSV is a method of representing points in the RGB color space in an inverted cone. HSV is Hue (Hue), saturation (Saturation), and lightness (Value). Hue is a basic attribute of color, such as red, yellow, and the like. The saturation (S) is the purity of the color, and the higher the color is, the more pure the color is, and the lower the color is, the gray gradually becomes, and the value is 0 to 100%. Lightness (V), 0-max (HSV value range is related to the length of storage). The color space conversion model of HSV is as follows:
let (R, G, B) be the red, green and blue channel parameter values for a color, respectively, which are real numbers between 0 and 255;
let max (R, G, B) equal the maximum of R, G, B;
let min (R, G, B) be equal to the smallest of R, G, B;
wherein, H represents the color information, namely the position of the spectral color; s represents the shade degree, and when S =0, represents only the gray scale; v represents the brightness of the color.
In some embodiments, the RGB color space of the first steel grade image is converted to an HSV color space to obtain the second steel grade image.
And obtaining a second steel type image after color space conversion, wherein the second steel type image has some interference structures except pearlite and ferrite, so that the banded structure in the second steel type image cannot be directly extracted, the interference structures need to be removed, and the influence of the interference structures on the detection structure is reduced.
The color characteristics of the interfering structures are different from those of the band-shaped tissue, and the interfering structures can be distinguished by the color characteristics.
In some embodiments, since the color characteristics of the banded structures of different steel types are different, a characteristic threshold value of the corresponding banded structure color needs to be set for each steel type, for example, the steel with the grade HL610A has white pearlite and black ferrite in a sampled image. For example, the steel with the mark of 42CrMo-N2 has bright white pearlite and greenish black ferrite in a sampling image. The steel with the grade of 1E0170 has pearlite slightly light grey and ferrite slightly dark grey in a sampling image. The 20 g steel showed lighter gray pearlite and darker black ferrite in the sampled image. Pearlite and ferrite of each steel type show different colors in the sampled image, so a characteristic threshold value of different zonal structure colors is set.
Verifying the second steel type image according to the RGB color feature, the HSV color feature and the feature threshold;
marking a structure of which the RGB color feature and/or the HSV color feature in the second steel type image do not meet a feature threshold; the structure which does not meet the characteristic threshold value is an interference structure;
and filtering the interference structure in the second steel grade image to obtain a third steel grade image.
In some embodiments, firstly, the steel type of the inspected second steel type image is determined according to the steel type information, the feature threshold of pearlite/ferrite corresponding to the steel type is called, the RGB color features and/or HSV color features of all banded structures in the second steel type image are verified and compared with the feature threshold, banded structures with the RGB color features and/or HSV color features not in the feature threshold range are screened out, and then the interference structures are filtered and removed.
It is understood that when the RGB color characteristic and/or the HSV color characteristic of the zonal tissue are within the range of the characteristic threshold, it means that the zonal tissue is pearlite and ferrite.
Through the steps, interference impurities which are not in the range of the characteristic threshold value are filtered and rejected pertinently, the accuracy of rating can be improved, different characteristic threshold values are set according to different steel types, interference structures can be better screened out, and the effect of filtering and rejecting the interference structures is better.
And filtering the interference structure of the image of the second steel grade to obtain an image of a third steel grade, wherein the banded structures in the image of the third steel grade are ferrite and pearlite.
S30: extracting banded tissue characteristics;
in the strip-shaped tissue feature extraction, firstly, the edge profile of the strip-shaped tissue in the third steel type image needs to be determined, and in the embodiment, a Canny edge detection method is adopted for obtaining the edge profile of the strip-shaped tissue.
Specifically, the extraction of the banded tissue features is data for establishing a basis for a banded feature quantitative rating model, and the information of the width, length, area and number of bands of the banded tissue needs to be acquired, so that the contour edge of the banded tissue needs to be extracted when the banded tissue feature data is acquired.
In some embodiments, the edge contour of the banded tissue is extracted using a Canny edge detection method, which, as shown in fig. 5, includes the steps of:
s301, smoothing the image by applying Gaussian filtering; by convolving the original number with a gaussian mask, the resulting image is slightly blurred compared to the original image. By this operation, the noise of a single pixel can be made to have little effect on the image subjected to gaussian smoothing.
S302, finding the intensity gradient of the image; the basic idea of Canny edge detection is to find the location in an image where the intensity of the gray scale changes most strongly. The strongest variation is referred to as the gradient direction. The gradient of each pixel point in the smoothed image can be obtained by a Sobel operator (a convolution operation) (a packaged function is provided in opencv, and the nth derivative of each pixel point in the image can be solved). First, the gradients g in the horizontal (x) and vertical (y) directions are determined, respectively x And g y 。
Wherein, G is the gradient size of the pixel point, theta is the gradient direction of the pixel point, and arctan is an arctangent function. The gradient angle θ ranges from radian- π to π, and is approximated to four directions, representing horizontal, vertical, and two diagonal directions (0, 45, 90, 135), respectively.
S303, eliminating boundary false detection by using a non-maximum suppression technology; the purpose of applying the non-maximum suppression technique is to clarify the blurred boundary, i.e. to retain the maximum value of the gradient intensity at each pixel point and to delete other values.
For each pixel point, the following operations are carried out: the gradient direction of the pixel is approximate to one of the following values (0, 45, 90, 135, 180, 225, 270, 315) (namely, the directions of the upper, lower, left and right directions and 45 degrees), the gradient strength of the pixel and the gradient direction positive and negative direction of the pixel are compared, if the gradient strength of the pixel is the maximum, the pixel is kept, and if the gradient strength of the pixel is the maximum, the pixel is inhibited (deleted, namely, set to be 0).
S304, determining a possibly existing boundary by applying a double-threshold method; after suppression by non-maximum values, there are still many possible edge points, and a double threshold, i.e. a low threshold and a high threshold, is further set. And if the gray level change is larger than the high threshold value, setting the gray level change as a strong edge pixel, and rejecting the gray level change below the low threshold value. The setting between the low and high thresholds is a weak edge. And further judging, if the strong edge pixels exist in the field, reserving the strong edge pixels, and if the strong edge pixels do not exist in the field, rejecting the strong edge pixels.
The Canny edge detection algorithm aims at finding an optimal edge detection algorithm, the meaning of the optimal edge detection is optimal detection, the algorithm can identify actual edges in an image as much as possible, and the probability of missing detection of the actual edges and the probability of false detection of non-edges are both as small as possible; the optimal positioning criterion is that the position of the detected edge point is closest to the position of the actual edge point, or the degree that the detected edge deviates from the real edge of the object due to the influence of noise is minimum; the detection points are in one-to-one correspondence with the edge points, and the edge points detected by the operator should be in one-to-one correspondence with the actual edge points. By Canny edge detection, the contour of the banded structure can be extracted as much as possible, and the accuracy of detecting the banded structure in the steel is improved.
After determining the edge profile of the strip-shaped tissue by the Canny edge detection method, the profile structure information of all strip-shaped tissues needs to be counted.
Wherein the information of the outline structure of all the banded tissues comprises the following steps: and creating a sliding window, wherein the length of the sliding window is the length (horizontal axis direction) of the image of the third steel grade, and the width of the sliding window is set according to the type of the steel grade.
Traversing the banded tissues in the third steel type image to acquire the outline structure information of the banded tissues; the contour structure information comprises an area value, a length value and a width value of the banded tissues; recording the outline structure information into a storage unit of the sliding window; and traversing the storage unit to obtain the quantization structure information.
Specifically, as shown in fig. 2, a sliding window is first created, the length of the sliding window is equal to the length (horizontal axis direction) of the third steel type image, the width of the sliding window is related to the steel type, and since the rating requirements of different steel types are different, the corresponding width needs to be set according to the steel type.
After the sliding window is arranged, traversing the whole third steel type image from the upper end of the third steel type image downwards, counting all the outline structure information of the banded structures in the range of the sliding window, recording the outline structure information into a storage unit of the sliding window, traversing the outline structure information in the storage unit, comparing the outline structure information in the storage unit with the image length (horizontal axis direction) and the window width, and generating the quantized structure information.
S40: quantitative rating of band-shaped features;
calculating all band-shaped tissue characteristics in the third steel grade image according to the quantized structure information;
classifying the banded tissues into penetrating bands, semi-penetrating bands and non-banded tissues according to the characteristics of the banded tissues, and respectively counting the number of the banded tissues under each classification;
wherein a band having a band length (horizontal axis direction) that is greater than 0.8 in the ratio to the image length (horizontal axis direction) and a sliding window width (vertical axis direction) that is greater than twice as large as the band width is classified as a through band; classifying the strip with the ratio of the strip length (horizontal axis direction) to the image length (horizontal axis direction) larger than 0.5 and smaller than 0.8 as a semi-penetration strip; the remainder was non-banded tissue.
The quantized structure information is classified, and the number of the band-shaped tissues of each class is calculated, for example, there are 10 pieces of the quantized structure information in the storage unit, where there are 3 pieces of the band satisfying that the occupation ratio of the image length (horizontal axis direction) is larger than 0.8 and the width of the sliding window where the band width (vertical axis direction) is larger than twice, and 2 pieces of the band satisfying that the occupation ratio of the band length (horizontal axis direction) with respect to the image length (horizontal axis direction) is larger than 0.5 and smaller than 0.8, the number of the penetrated bands is 3, the number of the semi-penetrated bands is 2, and the number of the non-band tissues is 5.
S50: and outputting a rating result.
And inputting the number of the banded tissues to generate a rating result through a banded tissue rating model.
In some embodiments, the model assigns rating rules of:
level 0: without banded tissue.
Level 0.5: a rating of 0.5 for tissues containing hard or ferritic bands; there is no strip of through-view and the width of the strip is less than 1 times the width of the single identification box.
Level 1.0: and the area ratio of the banded tissues is less than 0.5.
1.5-2.5 grade: there were through field of view strips rated 1.5; accumulating according to the number of the penetration strips, and adding 0.5 grade for 1 strip; if the area of the strip occupies more than 0.5 of the area of the sliding calculation window and the width of the strip is more than two times of the width of the sliding calculation window, the sum of the broadband and the sliding calculation window is 0.5 level.
2.5-3.5: more than 3 run-through strips, if there is one wide strip plus 0.5 level.
And 4, stage 4: and 4-level capping, if more than 3 penetrating strips exist, and the penetrating strips belong to the broadband.
The number of the banded structures existing in the third steel grade image and the degree of the banded structures penetrating are calculated. The number of the banded tissues is comprehensively considered. The ranking rule principle given by the model is to comprehensively evaluate the influence degree of the banded tissues on the subsequent use and quality, and the greater the number of the bands and the greater the banded penetration degree, the greater the influence on the subsequent use and quality, and thus the ranking level is higher.
And inputting the number of the banded structures into a banded structure rating model, outputting a final rating result by the banded structure rating model according to the number of the through bands, the semi-through bands and the non-banded structures in the number of the banded structures, and uploading the rating result to a metallographic detection control platform together with a third steel type image after outputting the rating result so as to ensure the traceability of subsequent data.
In some embodiments, as shown in fig. 3, the present application further provides an image processing-based in-steel strip tissue detection system comprising: the device comprises an image preprocessing unit, a banded tissue identification unit, a banded tissue characteristic rating unit and a rating result output unit;
the image preprocessing unit also comprises an image receiving and reading module and a color space conversion module;
the banded structure identification unit also comprises an interference impurity filtering module, a ferrite/pearlite identification module and a contour information extraction module;
the banded tissue characteristic rating unit comprises a banded tissue characteristic quantification module and a banded tissue quantification rating model.
The image receiving and reading module is used for acquiring and obtaining a sampling image of the sample block to be evaluated;
the color space conversion module is used for converting the RGB color space of the sampling image into HSV color space;
the interference impurity filtering module is used for carrying out targeted filtering and removing on interference structures which are not in the range of the characteristic threshold value;
the ferrite/pearlite identification module is used for extracting a banded structure within a characteristic threshold range;
the contour information extraction module is used for extracting contour edges of the banded tissues;
the banded characteristic quantization module is used for converting the outline structure information into quantization structure information;
the banded structure quantitative rating model is used for presetting a classification model constructed by banded structure information of different types of steel;
the rating result output unit is used for: and inputting the quantitative structure information into the banded structure quantitative rating model to obtain a rating result, and uploading the rating result and the third steel grade image to a metallographic detection control platform.
The functions and effects of the system in the application of the foregoing method can be referred to the description of the foregoing method embodiments, and are not repeated herein.
In view of the above technical solutions, the present application provides a method and a system for detecting a steel strip structure based on image processing, where the method includes: acquiring a sampling image of a sample block to be graded to obtain a first steel grade image; extracting the strip-shaped structure information in the first steel type image according to the color characteristics in the first steel type image; performing a quantization calculation on the strip-shaped organization structure information to obtain quantization structure information; inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result; the banded structure quantitative rating model is a classification model constructed according to banded structure information of preset different types of steel. The digital image detection method for the steel strip-shaped structure can be used for continuously detecting and improving the grading efficiency.
The embodiments provided in the present application are only a few examples of the general concept of the present application, and do not limit the scope of the present application. Any other embodiments extended according to the scheme of the present application without inventive efforts will be within the scope of protection of the present application for a person skilled in the art.
Claims (10)
1. A steel strip structure detection method based on image processing is characterized by comprising the following steps:
acquiring a sampling image of a sample block to be graded to obtain a first steel grade image;
extracting banded texture structure information in the first steel type image according to the color features in the first steel type image;
performing a quantization calculation on the strip-shaped organization structure information to obtain quantization structure information;
inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result; the banded structure quantitative rating model is a classification model constructed according to banded structure information of preset different types of steel.
2. The method of claim 1, wherein the step of obtaining a sample image of a sample block to be rated comprises:
scanning the sample block to be evaluated to obtain a sampling image of the sample block to be evaluated and the steel type of the sample block to be evaluated;
calling a rating standard file corresponding to the steel type;
if the sampling image corresponding to the steel type accords with the rating standard file, generating a first steel type image;
and if the sampling image corresponding to the steel type does not accord with the rating standard file, rescanning the sample block to be rated.
3. The method of claim 2, wherein after the step of calling the rating standard file corresponding to the steel type, the method further comprises:
detecting sampling parameters of the sampling image, wherein the sampling parameters comprise contrast ratio, brightness, reduction color and resolution;
analyzing parameter intervals in the rating standard file, wherein the parameter intervals comprise a contrast rate interval, a brightness interval, a reduction color interval and a resolution interval;
if all the sampling parameters are in the corresponding parameter intervals, determining that the sampling images corresponding to the steel types conform to the rating standard file;
and if any one of the sampling parameters is not in the corresponding parameter interval, determining that the sampling image corresponding to the steel grade type does not conform to the rating standard file.
4. The method of claim 1, wherein the step of extracting band-like texture information in the first steel type image based on color features in the first steel type image comprises:
performing color space conversion on the first steel grade image according to the first steel grade image to obtain a second steel grade image;
filtering the interference structure in the second steel type image to obtain a third steel type image;
and extracting the banded structure information in the third steel grade image.
5. The method of claim 4, wherein color space converting the first steel grade image comprises:
and converting the RGB color space of the first steel type image into HSV color space.
6. The method of claim 5, wherein filtering interfering structures in the image of the second steel grade to obtain an image of a third steel grade comprises:
setting a characteristic threshold value of the color of the banded tissue;
verifying the second steel type image according to the RGB color feature, the HSV color feature and the feature threshold;
marking structures in the second steel grade image where the RGB color features and/or the HSV color features do not meet the feature threshold; the structure which does not meet the characteristic threshold value is an interference structure;
and filtering the interference structure in the second steel grade image to obtain a third steel grade image.
7. The method of claim 4, wherein extracting the band-shaped texture information in the third steel grade image comprises:
determining the edge profile of the banded structure in the third steel type image by using a Canny edge detection method;
marking the banded texture region in the third steel grade image according to the edge profile.
8. The method of claim 4, wherein the step of performing a quantitative calculation on the band-shaped tissue structure information comprises:
creating a sliding window, wherein the length of the sliding window is the length of the image of the third steel grade, and the width of the sliding window is set according to the type of the steel grade;
traversing the banded tissues in the third steel type image to acquire the outline structure information of the banded tissues; the contour structure information comprises an area value, a length value and a width value of the banded tissue;
recording the outline structure information into a storage unit of the sliding window;
and traversing the storage unit to obtain the quantization structure information.
9. The method of claim 4, wherein inputting the quantitative structure information into a banded tissue quantitative rating model, and obtaining a rating result comprises:
calculating all band-shaped tissue characteristics in the third steel grade image according to the quantitative structure information;
classifying the banded tissues into penetrating banded tissues, semi-penetrating banded tissues and non-banded tissues according to the characteristics of the banded tissues, and respectively counting the number of the banded tissues under each classification;
and inputting the quantity of the banded tissues to generate a rating result through the banded tissue quantitative rating model.
10. A steel strip structure detection system based on image processing is characterized by comprising; the device comprises an image preprocessing unit, a banded tissue identification unit, a banded tissue characteristic rating unit and a rating result output unit;
the image pre-processing unit is configured to: acquiring a sampling image of a sample block to be graded, and preprocessing the sampling image;
the strip-shaped tissue identification unit is configured to: extracting banded structure information in the first steel type image;
the banded tissue feature rating unit is for: according to the banded tissue structure information, performing quantization calculation on the banded tissue structure information to obtain quantized structure information;
the rating result output unit is used for: and inputting the quantitative structure information into a banded tissue quantitative rating model to obtain a rating result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310132215.8A CN115861304A (en) | 2023-02-20 | 2023-02-20 | Method and system for detecting steel strip-shaped structure based on image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310132215.8A CN115861304A (en) | 2023-02-20 | 2023-02-20 | Method and system for detecting steel strip-shaped structure based on image processing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115861304A true CN115861304A (en) | 2023-03-28 |
Family
ID=85658367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310132215.8A Pending CN115861304A (en) | 2023-02-20 | 2023-02-20 | Method and system for detecting steel strip-shaped structure based on image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115861304A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116385472A (en) * | 2023-06-07 | 2023-07-04 | 深圳市锦红兴科技有限公司 | Hardware stamping part deburring effect evaluation method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877074A (en) * | 2009-11-23 | 2010-11-03 | 常州达奇信息科技有限公司 | Tubercle bacillus target recognizing and counting algorithm based on diverse characteristics |
| CN103020605A (en) * | 2012-12-28 | 2013-04-03 | 北方工业大学 | Bridge identification method based on decision-making layer fusion |
| CN105787931A (en) * | 2016-02-17 | 2016-07-20 | 中国工商银行股份有限公司 | Seal image detection method and system |
| CN106248721A (en) * | 2016-07-08 | 2016-12-21 | 首钢总公司 | A kind of structural steel bar banded structure rank Fast Prediction method |
| CN109117764A (en) * | 2018-07-29 | 2019-01-01 | 国网上海市电力公司 | Using the method for color threshold method identification target object region electrical symbol in power monitoring |
| CN112634200A (en) * | 2020-12-01 | 2021-04-09 | 宝钢特钢韶关有限公司 | Method for evaluating strip structure in steel |
| CN112801391A (en) * | 2021-02-04 | 2021-05-14 | 科大智能物联技术有限公司 | Artificial intelligent scrap steel impurity deduction rating method and system |
| CN113112513A (en) * | 2021-04-27 | 2021-07-13 | 汇鸿智能科技(辽宁)有限公司 | Method and device for AI (artificial intelligence) recognition of banded tissue, computer equipment and storage medium |
| CN113447512A (en) * | 2021-03-16 | 2021-09-28 | 首钢集团有限公司 | Quantitative evaluation method for hypoeutectoid steel strip-shaped structure |
-
2023
- 2023-02-20 CN CN202310132215.8A patent/CN115861304A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877074A (en) * | 2009-11-23 | 2010-11-03 | 常州达奇信息科技有限公司 | Tubercle bacillus target recognizing and counting algorithm based on diverse characteristics |
| CN103020605A (en) * | 2012-12-28 | 2013-04-03 | 北方工业大学 | Bridge identification method based on decision-making layer fusion |
| CN105787931A (en) * | 2016-02-17 | 2016-07-20 | 中国工商银行股份有限公司 | Seal image detection method and system |
| CN106248721A (en) * | 2016-07-08 | 2016-12-21 | 首钢总公司 | A kind of structural steel bar banded structure rank Fast Prediction method |
| CN109117764A (en) * | 2018-07-29 | 2019-01-01 | 国网上海市电力公司 | Using the method for color threshold method identification target object region electrical symbol in power monitoring |
| CN112634200A (en) * | 2020-12-01 | 2021-04-09 | 宝钢特钢韶关有限公司 | Method for evaluating strip structure in steel |
| CN112801391A (en) * | 2021-02-04 | 2021-05-14 | 科大智能物联技术有限公司 | Artificial intelligent scrap steel impurity deduction rating method and system |
| CN113447512A (en) * | 2021-03-16 | 2021-09-28 | 首钢集团有限公司 | Quantitative evaluation method for hypoeutectoid steel strip-shaped structure |
| CN113112513A (en) * | 2021-04-27 | 2021-07-13 | 汇鸿智能科技(辽宁)有限公司 | Method and device for AI (artificial intelligence) recognition of banded tissue, computer equipment and storage medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116385472A (en) * | 2023-06-07 | 2023-07-04 | 深圳市锦红兴科技有限公司 | Hardware stamping part deburring effect evaluation method |
| CN116385472B (en) * | 2023-06-07 | 2023-08-08 | 深圳市锦红兴科技有限公司 | Hardware stamping part deburring effect evaluation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112419250B (en) | Pavement crack digital image extraction, crack repair and crack parameter calculation method | |
| CN109934802B (en) | Cloth defect detection method based on Fourier transform and image morphology | |
| CN111145161B (en) | Pavement crack digital image processing and identifying method | |
| CN113592861B (en) | Bridge crack detection method based on dynamic threshold | |
| CN109507192B (en) | Magnetic core surface defect detection method based on machine vision | |
| CN111260616A (en) | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization | |
| US10453195B2 (en) | Method of detecting tissue area of interest in digital pathology imaging by executing computer-executable instructions stored on a non-transitory computer-readable medium | |
| CN109472788B (en) | Method for detecting flaw on surface of airplane rivet | |
| US20070189606A1 (en) | Automatic detection and correction of non-red eye flash defects | |
| KR101776355B1 (en) | Apparatus and methods for setting up optical inspection parameters | |
| CN115861304A (en) | Method and system for detecting steel strip-shaped structure based on image processing | |
| CN114037691A (en) | Carbon fiber plate crack detection method based on image processing | |
| KR101045393B1 (en) | Determination method of internal crack and center segregation in steel process | |
| CN111724375A (en) | Screen detection method and system | |
| CN106960424B (en) | Tubercle bacillus image segmentation and identification method and device based on optimized watershed algorithm | |
| CN110533626B (en) | All-weather water quality identification method | |
| CN113744326B (en) | Fire detection method based on seed region growth rule in YCRCB color space | |
| CN117974627A (en) | Surface crack detection method based on image processing | |
| CN114998346B (en) | Waterproof cloth quality data processing and identifying method | |
| CN112465817B (en) | Pavement crack detection method based on directional filter | |
| JP2001028059A (en) | Method and device for color unevenness inspection | |
| CN114913099A (en) | Method and system for processing video file | |
| JP3505496B2 (en) | Image processing method and image processing apparatus | |
| CN116883401B (en) | Industrial product production quality detection system | |
| JP2000028540A (en) | Free lime extraction method for concrete surface |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |