CN115797353A - Intelligent detection system and method for quality of cold-rolled strip steel - Google Patents
Intelligent detection system and method for quality of cold-rolled strip steel Download PDFInfo
- Publication number
- CN115797353A CN115797353A CN202310082637.9A CN202310082637A CN115797353A CN 115797353 A CN115797353 A CN 115797353A CN 202310082637 A CN202310082637 A CN 202310082637A CN 115797353 A CN115797353 A CN 115797353A
- Authority
- CN
- China
- Prior art keywords
- oil stain
- area
- detected
- determining
- index value
- 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.)
- Granted
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 142
- 239000010959 steel Substances 0.000 title claims abstract description 142
- 238000001514 detection method Methods 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title abstract description 19
- 239000013598 vector Substances 0.000 claims description 52
- 239000011159 matrix material Substances 0.000 claims description 29
- 230000035772 mutation Effects 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 6
- 238000003708 edge detection Methods 0.000 claims description 6
- 230000007547 defect Effects 0.000 abstract description 13
- 238000012545 processing Methods 0.000 abstract description 8
- 238000009826 distribution Methods 0.000 description 14
- 238000005096 rolling process Methods 0.000 description 11
- 239000002184 metal Substances 0.000 description 9
- 230000008569 process Effects 0.000 description 9
- 238000004519 manufacturing process Methods 0.000 description 8
- 230000003647 oxidation Effects 0.000 description 8
- 238000007254 oxidation reaction Methods 0.000 description 8
- 239000002932 luster Substances 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 239000010960 cold rolled steel Substances 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000011109 contamination Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 206010039509 Scab Diseases 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 239000002923 metal particle Substances 0.000 description 2
- 230000035699 permeability Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 238000000137 annealing Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 238000005246 galvanizing Methods 0.000 description 1
- 239000004519 grease Substances 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to an intelligent detection system and method for the quality of cold-rolled strip steel, wherein the gray image of the surface image of the strip steel to be detected is obtained, and the gray image is subjected to image processing, so that each oil stain area to be detected is obtained; and determining the strip characteristic position significant value and the texture characteristic index value of each oil stain area to be detected, further determining the oil stain adhesion significant degree of each oil stain area to be detected, and determining the quality of the strip steel to be detected according to the oil stain adhesion significant degree. According to the invention, the gray level image of the surface image of the detected strip steel is obtained, and the image processing is carried out on the gray level image, so that the quality of the strip steel can be accurately determined finally, and the problem that the detection of the oil stain defect on the surface of the existing strip steel is inaccurate, and further the detection of the surface quality of the strip steel is unreliable is effectively solved.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an intelligent detection system and method for the quality of cold-rolled strip steel.
Background
The cold-rolled strip steel has the advantages of smooth and clean surface, high size precision, good mechanical property and the like, is widely applied to various types of manufacturing industries and is mainly applied to the industries of automobiles, shipbuilding, electric power, coal, light industry, household appliances, hardware and the like. At present, in some industrially developed countries, the production of cold-rolled steel strip accounts for 30% of the total steel production. In order to ensure the quality of the strip steel and improve the utilization rate of the strip steel, the surface quality detection of the cold-rolled strip steel becomes an important part in the quality monitoring of steel plate production.
The existing on-line detection technology for the surface quality of the cold-rolled strip steel mainly comprises a manual visual detection method, a stroboscopic detection method, an eddy current detection method, a magnetic flux leakage detection method and the like. The manual visual detection method and the stroboscopic detection method are both manually observed, the consistency, the scientificity and the economy of detection are lacked, meanwhile, the spot inspection rate is low, the real-time performance is poor, and misjudgment and missed inspection are easily caused. The eddy current detection method and the magnetic leakage detection method are only suitable for certain application occasions with low requirements, and the limitations of the detection principle result in extremely limited detectable defect types and defect quantitative description parameters, so that the surface quality condition of a product cannot be comprehensively evaluated.
With the development of computer vision technology, image processing-based detection methods are increasingly applied to product surface quality detection. However, the existing image detection methods are all used for detecting substantial damages or defects on the surface of the strip steel, such as scabs, scratches, edge cracks and the like, and have insufficient resolution capability for the problems of 'benign' defects on the surface of the cold-rolled strip steel, such as excessive oil stain adhesion on the surface and the like. The special optical system for detecting the oil stain adhesion defect without substantial damage on the surface of the cold-rolled strip steel has a complex structure and poor maintainability and upgradability, an oil film on the surface of the cold-rolled strip steel can seriously influence a laser light path, great noise is generated on signals, the recognition capability of the system is seriously reduced, the detection accuracy is deteriorated, and finally the surface quality detection of the strip steel is unreliable.
Disclosure of Invention
The invention aims to provide an intelligent detection system and method for the quality of cold-rolled strip steel, which are used for solving the problem that the detection of the surface quality of the strip steel is unreliable due to inaccurate detection of the oil stain defects on the surface of the existing strip steel.
In order to solve the technical problem, the invention provides an intelligent detection method for the quality of cold-rolled strip steel, which comprises the following steps:
acquiring a gray image of a surface image of the strip steel to be detected, and carrying out edge detection on the gray image so as to obtain each oil stain area to be detected;
determining each positioning point corresponding to each edge pixel point of each oil stain area to be detected, determining a characteristic position distance and a target positioning point of each edge pixel point according to each edge pixel point and the position of each corresponding positioning point, and further determining a position index value of each oil stain area to be detected;
determining a shape index value and a main direction index value of each oil stain area to be tested, and determining a strip characteristic position significant value of each oil stain area to be tested according to the shape index value, the main direction index value and the position index value of each oil stain area to be tested;
determining a background area in the gray-scale image according to the gray-scale image and each oil stain area to be detected, determining texture feature vectors corresponding to each oil stain area to be detected and the background area according to gray-scale values of each pixel point in each oil stain area to be detected and the background area, and further determining texture feature index values of each oil stain area to be detected;
and determining the oil stain adhesion significance of each oil stain area to be detected according to the strip characteristic position significance and the texture characteristic index value of each oil stain area to be detected, and further determining the quality of the strip steel to be detected according to the oil stain adhesion significance.
Further, confirm each setpoint that every marginal pixel point of every greasy dirt region to be measured corresponds, include:
and determining a positioning line of each edge pixel point of each oil stain area to be detected in the gray-scale image, wherein the positioning line is a line segment formed by all pixel points of each edge pixel point in a column in the gray-scale image, and determining two end points and a center point of the positioning line as positioning points of corresponding edge pixel points.
Further, determining the feature position distance and the target positioning point of each edge pixel point includes:
calculating the distance value between each edge pixel point of each oil stain area to be measured and each corresponding locating point, determining the minimum distance value as the characteristic position distance of the corresponding edge pixel point, and determining the locating point corresponding to the minimum distance value as the target locating point of the corresponding edge pixel point.
Further, and then confirm the position index value in every greasy dirt region of awaiting measuring, include:
calculating an average value of all characteristic position distances according to the characteristic position distances of each edge pixel point of each oil stain area to be detected, and determining the average value as a first position index value;
determining the positioning serial number of the target positioning point of each edge pixel point of each oil stain to-be-detected area according to the position of the target positioning point of each edge pixel point of each oil stain to-be-detected area on the corresponding positioning line;
constructing a positioning serial number sequence according to the positioning serial number of the target positioning point of each edge pixel point of each oil stain area to be detected and the position of each edge pixel point in the corresponding oil stain area to be detected, carrying out mutation point detection on the positioning serial number sequence, and determining the number of the detected mutation points as a second position index value;
and determining the first position index value and the second position index value as the position index value of each oil stain area to be measured.
Further, confirm the shape index value and the principal direction index value of every greasy dirt region of awaiting measuring, include:
determining the area length and the area width of each oil stain area to be tested, and determining the ratio of the area length to the area width as a shape index value of the corresponding oil stain area to be tested;
and determining the principal component direction of each oil stain area to be tested, calculating the difference between a set direction coefficient and the principal component direction, and determining the principal component direction and the maximum value of the difference as the principal direction index value of the corresponding oil stain area to be tested.
Further, determining the significant value of the strip-shaped characteristic position of each oil stain area to be measured includes:
and calculating the sum of the first position index value and the second position index value of each oil stain area to be detected, calculating the product of the shape index value and the main direction index value of each oil stain area to be detected, and determining the ratio of the product to the sum as the strip characteristic position significant value of the corresponding oil stain area to be detected.
Further, confirm the texture feature vector that every greasy dirt region to be measured and background region correspond, include:
taking any one of the background area and each oil stain area to be detected as a target area, determining LBP texture values of all pixel points in the target area according to gray values of all pixel points in the target area, and further determining all target LBP texture values and LBP texture stability index values;
determining a gray level co-occurrence matrix according to the gray level value of each pixel point in the target area, and further determining the energy value and the inverse difference moment of the gray level co-occurrence matrix;
and arranging the target LBP texture values, the LBP texture stability index values, the energy values and the inverse difference moments according to a set sequence to form a row vector or a column vector, and determining the row vector or the column vector as a texture feature vector corresponding to the target area, so as to obtain the texture feature vector corresponding to each oil stain area to be detected and the background area.
Further, and then confirm the texture characteristic index value in every greasy dirt region that awaits measuring, include:
determining texture feature matrixes corresponding to each oil stain area to be detected and the background area according to the texture feature vectors corresponding to each oil stain area to be detected and the background area;
and calculating the absolute value of the difference between corresponding elements in the texture feature matrix corresponding to each oil stain region to be detected and the background region, and determining the average value of all the absolute values of the difference as the index value of the texture feature of the corresponding oil stain region to be detected.
Further, determining the quality of the strip steel to be detected comprises the following steps:
respectively judging whether the oil stain adhesion significance of each oil stain region to be detected is greater than or equal to a set significance threshold, if so, judging that the corresponding oil stain region to be detected is an oil stain region, otherwise, judging that the corresponding oil stain region to be detected is not an oil stain region;
and if all the areas to be tested with the oil stains are not the oil stain areas, judging that the quality of the strip steel to be tested is qualified, otherwise, judging that the quality of the strip steel to be tested is unqualified, and marking the strip steel to be tested as the position of the area to be tested with the oil stains in the oil stain area.
In order to solve the technical problem, the invention further provides an intelligent detection system for the quality of the cold-rolled strip steel, which comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor realizes the steps of the intelligent detection method for the quality of the cold-rolled strip steel when executing the computer program.
The invention has the following beneficial effects: the surface quality of the strip steel to be detected can be accurately determined by acquiring the surface image of the strip steel to be detected and processing the surface image. Specifically, edge detection is performed on the gray level image of the surface image, so that each oil stain area to be detected, which may be an oil stain area, is obtained. According to the characteristic position where the oil stain appears, each positioning point corresponding to each edge pixel point of the oil stain area to be detected is determined, the characteristic position distance and the target positioning point of each edge pixel point of the oil stain area to be detected can be determined by observing the distance from each edge pixel point of the oil stain area to be detected to the corresponding positioning point, and then the position index value of the oil stain area to be detected is determined, and the position index value represents the distribution position characteristic corresponding to the oil stain area to be detected. Based on the fact that the greasy dirt on the surface of the strip steel can present strip-shaped characteristics along the rolling direction of the strip steel, the shape index value and the main direction index value of each greasy dirt region to be detected are determined, and the shape index value and the main direction index value respectively represent the strip-shaped characteristics and the distribution main direction condition of the greasy dirt region to be detected. And comprehensively evaluating the shape and the position of the oil stain area to be detected by combining the position index value, the shape index value and the main direction index value of the oil stain area to be detected, and determining the strip characteristic position significant value of each oil stain area to be detected so as to more accurately determine the oil stain area to be detected corresponding to the oil stain. The shadow generated by the influence of light rays can be mistakenly judged as oil stains, the texture characteristic vectors corresponding to each oil stain area to be detected and the background area are determined according to the oxidation texture on the surface of the oil stains and the texture difference between the smooth metal steel belts, the texture characteristic vectors corresponding to the oil stain area to be detected and the background area are compared, and therefore the texture characteristic index value of each oil stain area to be detected is determined, the texture characteristic index value represents the difference situation between the texture characteristics of the oil stain area to be detected and the smooth metal texture characteristics, and therefore the interference of the light and shadow on the surface of the steel belts is avoided. The strip characteristic position significant value and the texture characteristic index value of each oil stain area to be detected are combined, and the oil stain adhesion significant degree of each oil stain area to be detected is finally determined, so that the accurate judgment of the quality of the strip steel to be detected is realized, and the problem that the detection of the oil stain defects on the surface of the existing strip steel is inaccurate, and the detection of the surface quality of the strip steel is unreliable is effectively solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of an intelligent detection method for quality of cold-rolled strip steel according to an embodiment of the invention.
Detailed Description
To further explain the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the embodiments, structures, features and effects of the technical solutions according to the present invention will be given with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. In addition, all parameters or indices in the formulas referred to herein are numerical values with dimensional effects removed.
In order to accurately identify the surface oil stain defects of the cold-rolled strip steel after production, timely treat the oil stain on the surface of the cold-rolled strip steel and ensure that the surface of the strip steel has no excessive pollutants, thereby avoiding the influence of the oil stain on the subsequent treatment process of the cold-rolled strip steel and ensuring the quality of the cold-rolled strip steel, the embodiment provides an intelligent detection method for the quality of the cold-rolled strip steel, and a corresponding flow chart of the method is shown in figure 1 and comprises the following steps:
step S1: and acquiring a gray image of the surface image of the strip steel to be detected, and carrying out edge detection on the gray image so as to obtain each oil stain area to be detected.
After the cold-rolled strip steel is produced, shooting the surface of the cold-rolled strip steel by using a CCD (charge coupled device) camera so as to obtain a surface image of the cold-rolled strip steel, wherein the surface image is an RGB (red, green and blue) image. In order to facilitate the subsequent identification of the oil stain defects on the surface of the cold-rolled strip steel, the rolling direction of the cold-rolled strip steel is in the horizontal direction in the surface image, that is, in the surface image, the horizontal direction is the rolling direction of the cold-rolled strip steel.
The obtained surface image of the cold-rolled strip steel is converted into a gray image, noise inevitably appears on the obtained gray image in consideration of the influence of factors such as environment, vibration of internal parts of a camera and the like, and in order to avoid the influence of the noise on the subsequent analysis result, the gray image is denoised by using median filtering to improve the precision and quality of the image, so that the final gray image is obtained, and the subsequent gray image refers to the final gray image. The median filtering and denoising is a known technology, and the specific process is not described in detail.
And (5) carrying out edge detection on the gray level image by using a canny edge detection algorithm, thereby obtaining an edge image which is recorded as a strip steel edge image. Marking each area in the gray level image according to the position of each pixel point in the area divided by each edge on the strip steel edge image, marking the areas as the areas to be tested for oil stains, wherein the areas to be tested for the oil stains are the areas which are possible to be the oil stains in the gray level image.
Step S2: determining each positioning point corresponding to each edge pixel point of each oil stain area to be detected, determining a characteristic position distance and a target positioning point of each edge pixel point according to each edge pixel point and the position of each corresponding positioning point, and further determining a position index value of each oil stain area to be detected.
The oil stains on the surface of the cold-rolled strip steel are mainly oil drips of a vertical clamp, oil drips of a reduction gearbox, residual oil of an inner cover flange, grease of a steel wire rope and the like, and after the oil stains drip on the surface of the cold-rolled strip steel, the oil stains are distributed in a strip shape along the rolling direction of the strip steel due to the influence of the rolling production flow. The oil stains on the surface of the cold-rolled steel strip are generally divided into two types, namely permeability and non-permeability. Permeable oil stains generally occur at the edge of the strip steel and diffuse from the edge to the inside of the strip steel; the impermeable oil stain is generally positioned in the middle of the inside of the strip steel, and no spots spread from the edge part to the inside of the strip steel exist. Wherein the permeable oil stain amount is more than 90% of the total oil stain amount. Therefore, by analyzing the regional distribution position characteristics and the strip distribution characteristics of the oil stain regions to be detected, the method is beneficial to accurately determining the real oil stain regions subsequently.
Firstly, determining a positioning point according to the characteristic position of oil stains appearing on the surface of the cold-rolled strip steel, namely the position of the middle position or the two side parts in the strip steel, wherein the implementation steps comprise: determining a positioning line of each edge pixel point of each oil stain area to be detected in the gray level image, wherein the positioning line is a line segment formed by all pixel points of each edge pixel point in a row in the gray level image, and determining two end points and a central point of the positioning line as positioning points of corresponding edge pixel points.
Specifically, for any oil stain area to be detected, each pixel point contained on the edge of the oil stain area to be detected is taken, the pixel points are called edge pixel points, a vertical line formed by each pixel point of each edge pixel point in a column of a gray level image is recorded as a positioning line corresponding to the edge pixel point, and each edge pixel point of the oil stain area to be detected is provided with one positioning line. And recording two points at two ends of the positioning line of each edge pixel point as a first positioning point and a second positioning point respectively, and recording the centroid position, namely the midpoint position, of the positioning line as a third positioning point, thereby obtaining each positioning point of each edge pixel point.
For any oil stain area to be detected, after each positioning point of each edge pixel point of the oil stain area to be detected is determined, the characteristic position distance and the target positioning point of each edge pixel point are determined according to each edge pixel point and the position of each corresponding positioning point, and the implementation steps comprise:
calculating the distance value between each edge pixel point of each oil stain area to be measured and each corresponding locating point, determining the minimum distance value as the characteristic position distance of the corresponding edge pixel point, and determining the locating point corresponding to the minimum distance value as the target locating point of the corresponding edge pixel point.
Specifically, considering that the oil stains generally appear at three characteristic positions, namely, the middle position or two side edge positions in the strip steel, and the horizontal direction of the gray level image is the rolling direction of the cold-rolled strip steel, each pixel point included on the edge of the area to be measured of the oil stains, namely, the edge pixel point, should be relatively close to one of the three positioning points corresponding to the edge pixel point. Therefore, the Euclidean distances from each edge pixel point of the oil stain region to be measured to three positioning points on the corresponding positioning line are determined and recorded as
Wherein the lower corner of the Euclidean distance is marked as the corresponding positioning number(s) ((x,y) The coordinates corresponding to the edge pixel points. According to the Euclidean distance from each edge pixel point of the oil stain region to be detected to three positioning points, the characteristic position distance corresponding to each edge pixel point can be obtained:
wherein the content of the first and second substances,c(x,y) For each area to be measured for oil contamination, the coordinates are (x,y) The feature location distance of the edge pixel point of (1),
respectively is the coordinate of (x,y) Respectively reach the Euclidean distance of a first positioning point, a second positioning point and a third positioning point corresponding to the edge pixel point,
the function is a value function, and the function is to take the minimum value in brackets.
The characteristic position distance reflects the distance between each edge pixel point of each oil stain area to be measured and the nearest characteristic position of the three characteristic positions where oil stains appear, and when the distance is smaller, the position corresponding to the edge pixel point is more likely to be the position corresponding to the oil stains. After the positions corresponding to the edge pixel points in the oil stain area to be detected are determined, the edge pixel points are located at the edge of the oil stain area to be detected, so that the positions of the oil stain area to be detected approximately appear can be reflected.
After the characteristic position distance of each edge pixel point of each oil stain area to be measured is determined, considering that the real oil stain area is a horizontally distributed strip area, each strip area is not too wide, and therefore each edge pixel point of each oil stain area to be measured should correspond to the same positioning point. Three positioning points, namely a positioning point I, a positioning point II and a positioning point III, are arranged on a positioning line corresponding to each edge pixel point of each oil stain area to be measured, the three positioning points are numbered in advance, and the number of the positioning points is recorded asa(x,y)，a(x,y) The value range of (1), (2) and (3). Because each edge pixel point of the oil stain region to be detected corresponding to the real oil stain should correspond to the same positioning point, the serial numbers of the positioning points corresponding to the characteristic position distances of each edge pixel point should be the same.
For every greasy dirt region to be measured, when this greasy dirt region to be measured is real greasy dirt region, then the characteristic position distance that each marginal pixel point of this greasy dirt region to be measured corresponds should be littleer, and the location serial number of the locating point that the characteristic position distance of each marginal pixel point corresponds should be the same, consequently can carry out the analysis to the characteristic position distance that each marginal pixel point of this greasy dirt region to be measured corresponds and the location serial number of locating point to confirm the position index value of this greasy dirt region to be measured, the realization step includes:
calculating the average value of all the characteristic position distances according to the characteristic position distance of each edge pixel point of each oil stain area to be measured, and determining the average value as a first position index value;
determining the positioning serial number of the target positioning point of each edge pixel point of each oil stain area to be detected according to the position of the target positioning point of each edge pixel point of each oil stain area to be detected on the corresponding positioning line;
constructing a positioning serial number sequence according to the positioning serial number of the target positioning point of each edge pixel point of each oil stain area to be detected and the position of each edge pixel point in the corresponding oil stain area to be detected, carrying out mutation point detection on the positioning serial number sequence, and determining the number of the detected mutation points as a second position index value;
and determining the first position index value and the second position index value as the position index value of each oil stain area to be measured.
Specifically, to every greasy dirt region to be measured, calculate the mean value of the characteristic position distance of each edge pixel point in this greasy dirt region to be measured to regard this mean value as the first position index value in this greasy dirt region to be measured, first position index value has represented the distance of the whole relative greasy dirt of this greasy dirt region to be measured characteristic position that appears, and when first position index value is less, it is nearer to explain the whole relative greasy dirt characteristic position that appears in this greasy dirt region to be measured, explains that this greasy dirt region to be measured is true greasy dirt very probably.
Meanwhile, for each oil stain area to be detected, according to the predetermined positioning numbers of the three positioning points on the positioning line corresponding to each edge pixel point, namely the positioning numbers corresponding to the positioning point I, the positioning point II and the positioning point III are respectively 1, 2 and 3, the target positioning point corresponding to each edge pixel point is determined to be one of the three positioning points, and therefore the positioning number of the target positioning point is determined. According to the positions of all edge pixel points of the oil stain region to be detected in the gray level image, the positioning numbers of the target positioning points corresponding to the edge pixel points are sequentially arranged from left to right and from top to bottom according to the positions of the edge pixel points, so that a column vector or a row vector can be obtained, and the column vector or the row vector is called as a positioning number sequence. And then carrying out Mann-Kendall mutation point detection on the positioning number sequence, and taking the number of the detected mutation points as a second position index value of the oil stain region to be detected. The specific implementation process of obtaining the number of mutation points by performing Mann-Kendall mutation point detection on the data sequence belongs to the prior art, and is not described herein again. The second position index value represents the distribution situation of the oil stain area to be measured in the vertical direction in the gray level image, for the oil stain area to be measured corresponding to the real oil stain, the area should be distributed along the horizontal direction, the span in the vertical direction is small, the positioning numbers of the positioning points corresponding to the edge pixel points are the same, no mutation point exists, and the second position index value should be zero. When the second position index value is large, the oil stain area to be detected possibly spans a plurality of characteristic positions simultaneously and is widely distributed in the vertical direction, so that the oil stain area to be detected does not accord with the characteristic that the oil stain is distributed in a strip shape along the horizontal direction, and the oil stain area to be detected is possibly not real oil stain.
After obtaining the first position index value and the second position index value of each oil stain area to be detected, the first position index value and the second position index value both represent the distribution position characteristics of the oil stain area to be detected, so that the first position index value and the second position index value are used as the position index values of the oil stain area to be detected, and the position index values and the texture characteristic index values can be combined subsequently, so that whether the oil stain area to be detected is a real oil stain area or not is accurately determined.
And step S3: and determining a shape index value and a main direction index value of each oil stain area to be tested, and determining a strip characteristic position significant value of each oil stain area to be tested according to the shape index value, the main direction index value and the position index value of each oil stain area to be tested.
Because real greasy dirt region is strip distribution along the horizontal direction in the gray level image, consequently can analyze this distribution characteristic of individual greasy dirt region of awaiting measuring, confirm the shape index value and the main direction index value of every greasy dirt region of awaiting measuring to follow-up confirm real greasy dirt region more accurately, realize the step and include:
determining the area length and the area width of each oil stain area to be tested, and determining the ratio of the area length to the area width as a shape index value of the corresponding oil stain area to be tested;
and determining the principal component direction of each oil stain area to be tested, calculating the difference between a set direction coefficient and the principal component direction, and determining the principal component direction and the maximum value of the difference as the principal direction index value of the corresponding oil stain area to be tested.
Specifically, for each oil stain area to be measured, according to the positioning lines corresponding to the edge pixel points of the oil stain area to be measured, the total number of pixel points located inside and on the edge of the oil stain area to be measured on each positioning line is calculated, the total number of the pixel points is used as the area width corresponding to the positioning lines, the average value of the area widths corresponding to all the positioning lines is calculated, and the average value is used as the area width of the oil stain area to be measuredw. The area width reflects the average width of the strip-shaped area of the oil stain area to be measured, and when the value is smaller, the whole width of the area is smaller. Meanwhile, calculating the total number of the positioning lines corresponding to each edge pixel point of the oil stain region to be measuredlAnd determining the total number as the area length of the area to be measured. Because the oil stain is strip-shaped, the ratio of the length to the width is large, so the ratio of the area length to the area width of the oil stain area to be detected is calculated, and the ratio is determined as the shape index value of the oil stain area to be detected. The shape index value represents the shape characteristic of the oil stain area to be measured, and when the value of the shape index value is larger, the shape index value is largerThe thinner and longer the corresponding oil stain area to be measured, the more likely it is oil stain.
PCA principal component analysis is used for each oil stain area to be measured to obtain principal component direction vectors
The principal component direction vector
The corresponding tilt angle is noted
Angle of inclination
That is, in the main component direction of each oil stain area to be measured, because the oil stains are distributed in a strip shape along the rolling direction of the strip steel, that is, in a strip shape along the horizontal direction, when the oil stain area to be measured is a real oil stain area, the obtained inclination angle is closer to 0 degree or 180 degrees. Based on the characteristic, the main direction index value corresponding to each oil stain area to be measured is set as
Wherein, in the step (A),
to set the orientation factor, which takes the value of 180,
the function is a value function, and the maximum value in brackets is taken. The main direction index value represents the distribution main direction condition of the oil stain area to be detected, and when the main direction index value is larger, the larger the main direction index value is, the more the corresponding oil stain area to be detected is distributed along the horizontal direction, the more the oil stain is possible.
Based on the determined shape index value, main direction index value and position index value of each oil stain area to be measured, comprehensively evaluating the shape and position of the oil stain area to be measured, determining the strip characteristic position significant value of each oil stain area to be measured, namely calculating the sum of the first position index value and the second position index value of each oil stain area to be measured, calculating the product value of the shape index value and the main direction index value of each oil stain area to be measured, determining the ratio of the product value and the sum as the strip characteristic position significant value of the corresponding oil stain area to be measured, wherein the corresponding calculation formula is as follows:
wherein, the first and the second end of the pipe are connected with each other,pfor the significant value of the strip-shaped characteristic position of each oil stain area to be measured,
the inclination angle corresponding to the direction vector of the principal component of each oil stain area to be measured,
to set the orientation factor, which takes the value of 180,
the index value of the main direction of each area to be measured with oil stain,lfor each zone length of the area to be examined for oil contamination,wfor the area width of each area to be measured for oil contamination,
the shape index value of each oil stain area to be measured,
a first position index value of each oil stain area to be measured,
and the index value of the second position of each oil stain area to be measured is obtained.
The above strip characteristic position significant value of each oil stain area to be measured reflects the significant degree of the strip characteristics of the oil stain area to be measured along the rolling direction of the strip steel and the distribution position characteristics of the oil stain area to be measured, when the value of the main direction index value and the shape index value of the oil stain area to be measured is larger, it is indicated that the oil stain area to be measured is more likely to be a strip area along the rolling direction of the strip steel, when the value of the first position index value and the value of the second position index value of the oil stain area to be measured are smaller, it is indicated that the oil stain area to be measured is more likely to be located at the characteristic position of the oil stain and concentrated at one characteristic position, and the corresponding strip characteristic position significant value is larger, so that the oil stain area to be measured is more likely to be real oil stain.
And step S4: determining a background area in the gray-scale image according to the gray-scale image and each oil stain area to be detected, determining texture feature vectors corresponding to each oil stain area to be detected and the background area according to gray-scale values of each pixel point in each oil stain area to be detected and the background area, and further determining texture feature index values of each oil stain area to be detected.
After the oil stain is dripped on the surface of the cold-rolled strip steel, the oil stain surface is smoother because the oil stain is originally liquid and the surface has tension. However, along with each process of cold-rolled strip steel production, some metal particles, dust and the like floating in the air can be attached to the surface of the oil stain, the oil stain on the surface of the cold-rolled strip steel can be oxidized under the influence of high temperature in a galvanizing annealing furnace, the surface of the oxidized oil stain is not smooth any more and is converted into a unique oxidation texture, and the oxidation texture is obviously different from the smooth surface with metal luster on the surface of the cold-rolled strip steel. When the image of the cold-rolled steel strip is acquired to identify the oil stain on the surface of the cold-rolled steel strip, the shadow is often influenced by light rays to enable the surface of the steel strip to present a shadow, and when the shadow generated due to the light rays falls on the positioning point of the oil stain in the analysis process and is in a strip shape, misjudgment is easily generated on the shadow, namely the misjudgment of the shadow is the oil stain. However, the shadow generated by the light on the surface of the cold-rolled steel strip does not change the material of the steel strip, the surface of the steel strip is still a smooth surface with metallic luster, the surface texture is unchanged and is consistent with that of a flawless steel strip, and therefore analysis is carried out again according to the difference of the texture characteristics of oil stains and the surface of the steel strip so as to avoid misjudgment.
Based on the analysis, the area formed by the residual pixel points in the gray level image after removing the pixel points in each oil stain area to be detected is recorded as a background area, and the background area corresponds to the cold-rolled strip steel without defects and is used for extracting the smooth metal texture on the surface of the strip steel. Respectively carrying out texture extraction on the background area and each oil stain area to be detected, thereby obtaining the texture characteristic vector corresponding to each oil stain area to be detected and the background area, and the implementation steps comprise:
taking any one of the background area and each oil stain area to be detected as a target area, determining an LBP texture value of each pixel point in the target area according to the gray value of each pixel point in the target area, and further determining each target LBP texture value and an LBP texture stability index value;
determining a gray level co-occurrence matrix according to the gray level value of each pixel point in the target area, and further determining the energy value and the inverse difference moment of the gray level co-occurrence matrix;
and arranging the target LBP texture values, the LBP texture stability index values, the energy values and the inverse difference moments according to a set sequence to form a row vector or a column vector, and determining the row vector or the column vector as a texture feature vector corresponding to the target area, so as to obtain the texture feature vector corresponding to each oil stain area to be detected and the background area.
Specifically, any one of the background region and each oil stain region to be detected is taken as a target region, and the LBP texture value of each pixel point in the target region is extracted according to the gray value of each pixel point in the target region. According to the LBP texture value of each pixel point in the target area, an LBP texture map of the target area can be obtained, the pixel value of each pixel point in the LBP texture map is the LBP texture value of the corresponding pixel point in the target area, and then an LBP distribution histogram is obtained according to the LBP texture map. In the LBP distribution histogram, the abscissa is each LBP texture value of a pixel point in the LBP texture map, and the ordinate is the frequency of occurrence of each LBP texture value in the LBP texture map. And fitting a peak-valley curve in the LBP distribution histogram by using envelope fitting in EMD, and deriving the fitted peak-valley curve to further determine each peak vertex. Determining three peak vertexes with the maximum ordinate in all the peak vertexes, determining LBP texture values corresponding to the three peak vertexes as target LBP texture values, and recording the target LBP texture values as the target LBP texture values respectively
. These three target LBP patternsPhysical value
More concentrated LBP texture values in the analyzed target region are characterized. For the target area corresponding to the oil stain, due to the oxidation, the area is provided with the oxidation texture, the three target LBP texture values extracted from the area are generally dispersed and have larger difference, while the background area is provided with the smooth metal texture, so the three target LBP texture values extracted from the area are generally closer and have smaller difference, and therefore, the three target LBP texture values extracted from the target area corresponding to the oil stain have obvious difference relative to the three target LBP texture values extracted from the background area. For the target area corresponding to the shadow, because the area still has smooth texture with metallic luster, the three target LBP texture values extracted from the area are also relatively close and have small difference, so that the three target LBP texture values extracted from the target area corresponding to the shadow are not very different from the three target LBP texture values extracted from the background area.
Meanwhile, according to the LBP texture value of each pixel point in the LBP texture map corresponding to the target area, calculating the standard deviation of the LBP texture values, and recording the standard deviation assThe standard deviation is calculatedsAnd determining the LBP texture stability index value as the target area. The LBP texture stability index value characterizes the consistency of the LBP texture value in the target area, and the value is smaller when the consistency is larger. For the target area corresponding to the oil stain, because the area is an oxidized texture and the difference of the LBP texture values is large, the LBP texture stability index value corresponding to the area is usually large, and the background area is a smooth metal texture, and the LBP texture values are relatively close to each other, so the LBP texture stability index value corresponding to the area is usually small, and therefore the LBP texture stability index value determined by the target area corresponding to the oil stain is obviously different from the LBP texture stability index value determined by the background area. On the other hand, for the target area corresponding to the shadow, since the area still has a smooth texture with metallic luster, the LBP texture stability index value corresponding to the area is usually smaller, so that the LBP texture stability index value corresponding to the target area corresponding to the shadow is corresponding to the background areaThe LBP texture stability index values are not very different from each other.
In addition, the gray level co-occurrence matrix corresponding to the target area is obtained according to the gray level value of each pixel point in the target area, and since the specific implementation process of obtaining the gray level co-occurrence matrix belongs to the prior art, the detailed description is omitted here. Obtaining corresponding energy value according to the gray level co-occurrence matrixasmSum and inverse difference momentidmWherein the energy valueasmThe method is characterized in that the square sum of each element in the matrix reflects the uniformity degree of gray level distribution and the thickness degree of textures in a target area, when the elements in the gray level co-occurrence matrix are distributed in a concentrated mode, namely the element values are close, the energy value is small, when some elements in the gray level co-occurrence matrix take large values, and when other elements take small values, the energy value is large. Moment of adverse differenceidmThe homogeneity of the texture in the target area is reflected, the local change of the texture in the target area is measured, when the texture in the target area lacks variation and is locally uniform, the inverse difference moment is larger, and when the texture in the target area is not uniformly distributed, the inverse difference moment is smaller. For a target area corresponding to oil stains, because the area is an oxidized texture which is a unique texture formed by changing the property after the oil stains are oxidized, the oil stains at different positions are oxidized differently, and small metal particles can be attached to the surface of the oil stains, the texture on the surface of the oil stains is relatively complex, so that the energy value and the inverse difference moment corresponding to the area are generally relatively small, and the background area is a smooth metal texture. On the other hand, in the target region corresponding to the shadow, since the region is still smooth texture having metallic luster, the energy value and the inverse difference moment corresponding to the region are generally large, and therefore, the energy value and the inverse difference moment corresponding to the target region corresponding to the shadow are not greatly different from those corresponding to the background region.
In the above manner, canObtaining three target LBP texture values corresponding to a background area and each oil stain area to be measured
LBP texture stability index valuesEnergy value ofasmSum and inverse momentidmThe six indexes are obviously different from those of the background area, and the six indexes of the oil stain area to be measured corresponding to the shadow are smaller than those of the background area. Based on the characteristic, three target LBP texture values corresponding to the background area and each oil stain area to be detected respectively
LBP texture stability index valuesEnergy valueasmSum and inverse difference momentidmThe six indexes are used for constructing a texture feature vector
And subsequently, the texture characteristic vector of each oil stain area to be detected can be compared with the texture characteristic vector of the background area, so that the shadow area is distinguished, and the accuracy of oil stain identification is improved. As another embodiment, the above-described texture feature vector may be a row vector, and one column vector may be obtained by taking the transpose of the row vector, and the column vector may be used as a structural texture feature vector.
Comparing the difference between the texture feature vector of each oil stain area to be detected and the texture feature vector of the background area, and further determining the texture feature index value of each oil stain area to be detected, so that the shadow area can be distinguished in a follow-up manner by using the texture feature index value in an auxiliary manner, wherein the implementation steps comprise:
determining texture feature matrixes corresponding to each oil stain area to be detected and the background area according to the texture feature vectors corresponding to each oil stain area to be detected and the background area;
and calculating the absolute value of the difference between corresponding elements in the texture feature matrix corresponding to each oil stain region to be detected and the background region, and determining the average value of all the absolute values of the difference as the index value of the texture feature of the corresponding oil stain region to be detected.
Specifically, in order to compare consistency of the texture feature vector of each oil stain area to be measured with the texture feature vector of the background area, the texture feature matrix corresponding to each oil stain area to be measured and the background area is obtained according to the texture feature vector corresponding to each oil stain area to be measured and the background area, the texture feature matrix is a Gram matrix determined according to the texture feature vectors, and the texture feature matrix of each oil stain area to be measured is recorded as the Gram matrixAThe texture feature matrix of the background area is a Gram matrixB. Since the specific implementation process of acquiring the Gram matrix belongs to the prior art, it is not described herein again.
Because the oil stain area to be measured corresponding to the oil stain is the oil stain oxidation texture, and the shadow generated by the light on the surface of the steel belt is still a smooth surface with metal luster and is consistent with the flawless steel belt, when the oil stain area to be measured corresponds to the oil stain attached to the surface of the steel belt, the texture difference of the flawless steel belt corresponding to the background area is larger, so that the corresponding Gram matrix is largerAAnd Gram matrixBThe difference is large; when the oil stain area to be measured is a shadow generated by light interference, because the corresponding areas of the area and the background area are both flawless steel belts and the surface textures are consistent, the difference between the two matrixes is small, the oxidation texture significance degree of each oil stain area to be measured is evaluated based on the difference, and the corresponding texture characteristic index value is obtained, wherein the corresponding calculation formula is as follows:
wherein the content of the first and second substances,othe index value of the textural features of each oil stain area to be measured,
texture feature matrix for each oil stain area to be measuredAIn
Element values of positions, i.e. texture feature matricesATo middleiGo to the firstjThe value of the element of the column,
texture feature matrix as background regionBIn
Element values of positions, i.e. texture feature matricesBTo middleiGo to the firstjThe value of an element of a column is,
as a texture feature matrixAAndBthe number of element values contained in each of them, in the present embodiment, since the texture feature matrix is constitutedAAndBthe texture feature vector of (1) is composed of six indexes, so
All values of (1, 6)]The number of the integer (c) of (a),
is 36.
The texture feature index value reflects the surface texture difference between each oil stain area to be detected and the background area determined to be a defect-free steel strip, when the difference is larger, the texture feature index value is larger, namely the oil stain area to be detected is more likely to correspond to oil stains attached to the surface of the steel strip, and when the difference is smaller, the texture feature index value is smaller, namely the oil stain area to be detected is more likely to be an interference area formed by light shadow.
Step S5: and determining the oil stain adhesion significance of each oil stain area to be detected according to the strip characteristic position significance and the texture characteristic index value of each oil stain area to be detected, and further determining the quality of the strip steel to be detected according to the oil stain adhesion significance.
And determining the oil stain adhesion significance of each oil stain area to be tested by combining the strip characteristic position significant value and the texture characteristic index value of each oil stain area to be tested, wherein the corresponding calculation formula is as follows:
wherein the content of the first and second substances,cfor the oil stain attachment significance of each area to be tested for oil stains,pfor the significant value of the strip-shaped characteristic position of each oil stain area to be measured,oand (4) obtaining the index value of the texture characteristic of each oil stain area to be detected.
The oil stain adhesion significance is a comprehensive evaluation value of strip-shaped characteristics along the strip steel rolling direction, the degree of similarity between the position of the oil stain adhesion significance and the oil stain characteristic position and the surface texture difference degree of each oil stain area to be detected, when the oil stain adhesion significance is high, the corresponding oil stain area to be detected is more likely to be an oil stain area, and when the oil stain adhesion significance is low, the corresponding oil stain area to be detected is less likely to be an oil stain area.
In order to obtain a uniform judgment standard, the oil stain adhesion significance of each oil stain area to be detected is normalized to be within the range of 0-1, and the oil stain adhesion significance after normalization processing is used as the final oil stain adhesion significance z. Determining the quality of the strip steel to be detected based on the final oil stain adhesion significance z of each oil stain area to be detected, wherein the implementation steps comprise:
respectively judging whether the oil stain adhesion significance of each oil stain region to be detected is greater than or equal to a set significance threshold, if so, judging that the corresponding oil stain region to be detected is an oil stain region, otherwise, judging that the corresponding oil stain region to be detected is not an oil stain region;
and if all the areas to be tested with the oil stains are not the oil stain areas, judging that the quality of the strip steel to be tested is qualified, otherwise, judging that the quality of the strip steel to be tested is unqualified, and marking the strip steel to be tested as the position of the area to be tested with the oil stains in the oil stain area.
Specifically, a set saliency threshold z is preset 0 Setting a significance threshold value z 0 Can be set according to experiments or experiences, and when the quality requirement of the cold-rolled strip steel is higher, the significance threshold value z can be set 0 The small point set, when the quality requirement for cold-rolled strip steel is not so high, can be set the significance threshold z 0 The greater the setting, the present embodiment sets the significance threshold z 0 Is 0.8.
The oil stain attachment significance z of each oil stain area to be measured and a preset significance threshold value z are set 0 Comparing if z is more than or equal to z 0 And if not, determining that the corresponding oil stain area to be detected is not the oil stain on the surface of the cold-rolled strip steel.
When detecting that the oil stain to be detected area in the gray level image corresponds to the oil stain on the surface of the cold-rolled strip steel, outputting the conclusion that the oil stain exists on the surface of the strip steel and the quality of the strip steel is not qualified, and outputting the position of the corresponding oil stain to be detected area so as to conveniently process the oil stain on the surface of the strip steel; and when detecting that the area to be detected without oil stains in the gray level image is judged to correspond to the oil stains on the surface of the cold-rolled strip steel, considering that the quality of the strip steel is good, and outputting a conclusion that the quality of the strip steel is qualified. And under the condition that the oil stains exist on the surface of the strip steel, the oil stains on the surface of the strip steel are cleaned, so that the subsequent judgment of the surface defects of the strip steel and the subsequent processing of the strip steel are prevented from being influenced.
Based on the same inventive concept, the embodiment further provides an intelligent detection system for the quality of the cold-rolled strip, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor implements the steps of the intelligent detection method for the quality of the cold-rolled strip when executing the computer program. The intelligent detection system for the quality of the cold-rolled strip steel has the core that the intelligent detection method for the quality of the cold-rolled strip steel is realized, and the method is described in detail in the content and is not repeated herein.
The method starts from the reason of oil stain generation of the cold-rolled strip steel, analyzes the characteristic that the oil stain on the surface of the cold-rolled strip steel appears as a strip along the rolling direction of the strip steel according to the influence of the production flow of the cold-rolled strip steel on the oil stain dropping on the surface of the cold-rolled strip steel, and judges the distance between the area to be tested for the oil stain on the surface of the strip steel and the characteristic position of the oil stain on the edge or the middle part of the strip steel by combining the reason of the oil drop dropping, thereby determining the obvious value of the strip characteristic position of the area to be tested for the oil stain. And then determining a texture characteristic index value of an oil stain area to be detected according to the oxidation texture on the oil stain surface and the texture difference between smooth metal steel belts, thereby eliminating the influence of light and shadow in the image shooting process, finally combining the strip characteristic position significant value and the texture characteristic index value of the oil stain area to be detected to obtain a comprehensive index of oil stain adhesion significant degree, and accurately determining whether the oil stain area to be detected corresponds to real oil stains according to the comprehensive index, thereby finally determining the quality of the steel belts, effectively solving the problem that the detection of the oil stain defects on the surfaces of the existing steel belts is inaccurate, and further causing the detection of the surface quality of the steel belts to be unreliable, and laying a foundation for accurately judging the subsequent substantial damages to the steel belts, such as scabs, scratches, edge cracks and the like.
It should be noted that: the above-mentioned embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the embodiments of the present application, and they should be construed as being included in the present application.
Claims (10)
1. The intelligent detection method for the quality of the cold-rolled strip steel is characterized by comprising the following steps:
acquiring a gray image of a surface image of the strip steel to be detected, and carrying out edge detection on the gray image so as to obtain each oil stain area to be detected;
determining each positioning point corresponding to each edge pixel point of each oil stain area to be detected, determining the characteristic position distance and the target positioning point of each edge pixel point according to each edge pixel point and the position of each positioning point corresponding to each edge pixel point, and further determining the position index value of each oil stain area to be detected;
determining a shape index value and a main direction index value of each oil stain area to be tested, and determining a strip characteristic position significant value of each oil stain area to be tested according to the shape index value, the main direction index value and the position index value of each oil stain area to be tested;
determining a background area in the gray-scale image according to the gray-scale image and each oil stain area to be detected, determining texture feature vectors corresponding to each oil stain area to be detected and the background area according to gray-scale values of each pixel point in each oil stain area to be detected and the background area, and further determining texture feature index values of each oil stain area to be detected;
and determining the oil stain adhesion significance of each oil stain area to be detected according to the strip characteristic position significance and the texture characteristic index value of each oil stain area to be detected, and further determining the quality of the strip steel to be detected according to the oil stain adhesion significance.
2. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 1, wherein the step of determining each positioning point corresponding to each edge pixel point of each oil stain area to be detected comprises the following steps:
determining a positioning line of each edge pixel point of each oil stain area to be detected in the gray level image, wherein the positioning line is a line segment formed by all pixel points of each edge pixel point in a row in the gray level image, and determining two end points and a central point of the positioning line as positioning points of corresponding edge pixel points.
3. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 1, wherein the step of determining the characteristic position distance and the target positioning point of each edge pixel point comprises the following steps:
calculating the distance value between each edge pixel point of each oil stain area to be measured and each corresponding locating point, determining the minimum distance value as the characteristic position distance of the corresponding edge pixel point, and determining the locating point corresponding to the minimum distance value as the target locating point of the corresponding edge pixel point.
4. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 2, wherein the step of determining the position index value of each oil stain area to be detected comprises the following steps:
calculating an average value of all characteristic position distances according to the characteristic position distances of each edge pixel point of each oil stain area to be detected, and determining the average value as a first position index value;
determining the positioning serial number of the target positioning point of each edge pixel point of each oil stain to-be-detected area according to the position of the target positioning point of each edge pixel point of each oil stain to-be-detected area on the corresponding positioning line;
constructing a positioning serial number sequence according to the positioning serial number of the target positioning point of each edge pixel point of each oil stain area to be detected and the position of each edge pixel point in the corresponding oil stain area to be detected, carrying out mutation point detection on the positioning serial number sequence, and determining the number of the detected mutation points as a second position index value;
and determining the first position index value and the second position index value as the position index value of each oil stain area to be measured.
5. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 1, wherein the step of determining the shape index value and the main direction index value of each oil stain area to be detected comprises the following steps:
determining the area length and the area width of each oil stain area to be tested, and determining the ratio of the area length to the area width as a shape index value of the corresponding oil stain area to be tested;
and determining the principal component direction of each oil stain area to be tested, calculating the difference between a set direction coefficient and the principal component direction, and determining the principal component direction and the maximum value of the difference as the principal direction index value of the corresponding oil stain area to be tested.
6. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 4, wherein the step of determining the strip characteristic position significant value of each oil stain area to be detected comprises the following steps:
and calculating the sum of the first position index value and the second position index value of each oil stain area to be detected, calculating the product of the shape index value and the main direction index value of each oil stain area to be detected, and determining the ratio of the product to the sum as the strip characteristic position significant value of the corresponding oil stain area to be detected.
7. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 1, wherein the step of determining the texture feature vector corresponding to each oil stain area to be detected and the background area comprises the following steps:
taking any one of the background area and each oil stain area to be detected as a target area, determining an LBP texture value of each pixel point in the target area according to the gray value of each pixel point in the target area, and further determining each target LBP texture value and an LBP texture stability index value;
determining a gray level co-occurrence matrix according to the gray level value of each pixel point in the target area, and further determining the energy value and the inverse difference moment of the gray level co-occurrence matrix;
and arranging the target LBP texture values, the LBP texture stability index values, the energy values and the inverse difference moments according to a set sequence to form a row vector or a column vector, and determining the row vector or the column vector as a texture feature vector corresponding to the target area, so as to obtain the texture feature vector corresponding to each oil stain area to be detected and the background area.
8. The intelligent detection method for the quality of the cold-rolled strip steel as claimed in claim 1, wherein the step of determining the index value of the textural features of each oil stain area to be detected comprises the following steps:
determining texture feature matrixes corresponding to each oil stain area to be detected and the background area according to the texture feature vectors corresponding to each oil stain area to be detected and the background area;
and calculating the absolute value of the difference between corresponding elements in the texture feature matrix corresponding to each oil stain region to be detected and the background region, and determining the average value of all the absolute values of the difference as the index value of the texture feature of the corresponding oil stain region to be detected.
9. The intelligent detection method for the quality of the cold-rolled strip steel according to claim 1, wherein the step of determining the quality of the strip steel to be detected comprises the following steps:
respectively judging whether the oil stain adhesion significance of each oil stain region to be detected is greater than or equal to a set significance threshold, if so, judging that the corresponding oil stain region to be detected is an oil stain region, otherwise, judging that the corresponding oil stain region to be detected is not an oil stain region;
if all the areas to be tested with the oil stains are not the oil stain areas, judging that the quality of the strip steel to be tested is qualified, otherwise, judging that the quality of the strip steel to be tested is unqualified, and marking the strip steel to be tested as the position of the area to be tested with the oil stains in the oil stain areas.
10. An intelligent quality detection system for cold-rolled strip, comprising a memory, a processor and a computer program stored in the memory and operable on the processor, wherein the processor implements the steps of the intelligent quality detection method for cold-rolled strip according to any one of claims 1 to 9 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310082637.9A CN115797353B (en) | 2023-02-08 | 2023-02-08 | Intelligent detection system and method for quality of cold-rolled strip steel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310082637.9A CN115797353B (en) | 2023-02-08 | 2023-02-08 | Intelligent detection system and method for quality of cold-rolled strip steel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115797353A true CN115797353A (en) | 2023-03-14 |
| CN115797353B CN115797353B (en) | 2023-05-09 |
Family
ID=85430529
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310082637.9A Active CN115797353B (en) | 2023-02-08 | 2023-02-08 | Intelligent detection system and method for quality of cold-rolled strip steel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115797353B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014119376A1 (en) * | 2013-01-31 | 2014-08-07 | 株式会社日立ハイテクノロジーズ | Defect inspection device and defect inspection method |
| JP2019032268A (en) * | 2017-08-09 | 2019-02-28 | 日立金属株式会社 | Surface defect determination method and surface defect inspection device |
| JP2019078640A (en) * | 2017-10-25 | 2019-05-23 | Ｊｆｅスチール株式会社 | Surface defect detection method and surface defect detector |
| CN114937039A (en) * | 2022-07-21 | 2022-08-23 | 阿法龙(山东)科技有限公司 | Intelligent detection method for steel pipe defects |
| CN115049656A (en) * | 2022-08-15 | 2022-09-13 | 海门市刘氏铸造有限公司 | Method for identifying and classifying defects in silicon steel rolling process |
| CN115100221A (en) * | 2022-08-22 | 2022-09-23 | 启东市云鹏玻璃机械有限公司 | Glass defect segmentation method |
| CN115222733A (en) * | 2022-09-19 | 2022-10-21 | 山东金源不锈钢制品有限公司 | Metal component surface defect detection method based on image recognition |
| CN115311292A (en) * | 2022-10-12 | 2022-11-08 | 南通创铭伊诺机械有限公司 | Strip steel surface defect detection method and system based on image processing |
| CN115375676A (en) * | 2022-10-24 | 2022-11-22 | 微山三利特不锈钢有限公司 | Stainless steel product quality detection method based on image recognition |
-
2023
- 2023-02-08 CN CN202310082637.9A patent/CN115797353B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014119376A1 (en) * | 2013-01-31 | 2014-08-07 | 株式会社日立ハイテクノロジーズ | Defect inspection device and defect inspection method |
| JP2019032268A (en) * | 2017-08-09 | 2019-02-28 | 日立金属株式会社 | Surface defect determination method and surface defect inspection device |
| JP2019078640A (en) * | 2017-10-25 | 2019-05-23 | Ｊｆｅスチール株式会社 | Surface defect detection method and surface defect detector |
| CN114937039A (en) * | 2022-07-21 | 2022-08-23 | 阿法龙(山东)科技有限公司 | Intelligent detection method for steel pipe defects |
| CN115049656A (en) * | 2022-08-15 | 2022-09-13 | 海门市刘氏铸造有限公司 | Method for identifying and classifying defects in silicon steel rolling process |
| CN115100221A (en) * | 2022-08-22 | 2022-09-23 | 启东市云鹏玻璃机械有限公司 | Glass defect segmentation method |
| CN115222733A (en) * | 2022-09-19 | 2022-10-21 | 山东金源不锈钢制品有限公司 | Metal component surface defect detection method based on image recognition |
| CN115311292A (en) * | 2022-10-12 | 2022-11-08 | 南通创铭伊诺机械有限公司 | Strip steel surface defect detection method and system based on image processing |
| CN115375676A (en) * | 2022-10-24 | 2022-11-22 | 微山三利特不锈钢有限公司 | Stainless steel product quality detection method based on image recognition |
Non-Patent Citations (3)
| Title |
|---|
| ZHIPENG LI 等: ""Metal surface defect detection based on MATLAB"" * |
| 岳鹏飞;晏永飞;王楠;: "非接触式带钢表面缺陷检测纹理方向特征分析" * |
| 郭慧;王霄;刘传泽;周玉成;: "基于灰度共生矩阵和分层聚类的刨花板表面图像缺陷提取方法" * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115797353B (en) | 2023-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109682839B (en) | Online detection method for surface defects of metal arc-shaped workpiece | |
| US6779159B2 (en) | Defect inspection method and defect inspection apparatus | |
| CN106404793B (en) | Bearing sealing element defect detection method based on vision | |
| CN109900711A (en) | Workpiece, defect detection method based on machine vision | |
| CN115100203B (en) | Method for detecting quality of steel bar polishing and rust removal | |
| CN115063423B (en) | Self-adaptive identification method for cold and hot cracks of mechanical castings based on computer vision | |
| US20060029257A1 (en) | Apparatus for determining a surface condition of an object | |
| CN115375686B (en) | Glass edge flaw detection method based on image processing | |
| CN112037203A (en) | Side surface defect detection method and system based on complex workpiece outer contour registration | |
| CN112767359B (en) | Method and system for detecting corner points of steel plate under complex background | |
| CN115965623B (en) | Surface flaw detection method and system in transformer production | |
| CN113340909B (en) | Glue line defect detection method based on machine vision | |
| CN116503397B (en) | In-vehicle transmission belt defect detection method based on image data | |
| CN115100191A (en) | Metal casting defect identification method based on industrial detection | |
| CN115330796B (en) | Copper wire tinning defect identification method | |
| CN115684176A (en) | Online visual inspection system for film surface defects | |
| CN112669295A (en) | Lithium battery pole piece defect detection method based on secondary threshold segmentation theory | |
| CN116468687A (en) | Scratch defect detection method and device, storage medium and electronic equipment | |
| CN110426395B (en) | Method and device for detecting surface of solar EL battery silicon wafer | |
| CN115274486A (en) | Semiconductor surface defect identification method | |
| CN116612112B (en) | Visual inspection method for surface defects of bucket | |
| CN115797353A (en) | Intelligent detection system and method for quality of cold-rolled strip steel | |
| CN115423816A (en) | Metal surface grinding quality detection method | |
| CN114998354A (en) | Steel belt roll mark detection method based on graph filtering | |
| CN115115624A (en) | Rolling damage detection method for anti-corrosion coating of cable bridge |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: A Quality Intelligent Detection System and Method for Cold Rolled Strip Steel Effective date of registration: 20231129 Granted publication date: 20230509 Pledgee: China Postal Savings Bank Co.,Ltd. Jiaxiang County sub branch Pledgor: Shandong Qiangang Metal Technology Co.,Ltd. Registration number: Y2023980068224 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right |