CN113628189B - Rapid strip steel scratch defect detection method based on image recognition - Google Patents
Rapid strip steel scratch defect detection method based on image recognition Download PDFInfo
- Publication number
- CN113628189B CN113628189B CN202110918572.8A CN202110918572A CN113628189B CN 113628189 B CN113628189 B CN 113628189B CN 202110918572 A CN202110918572 A CN 202110918572A CN 113628189 B CN113628189 B CN 113628189B
- Authority
- CN
- China
- Prior art keywords
- sub
- image
- block
- gray
- blocks
- 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.)
- Active
Links
- 230000007547 defect Effects 0.000 title claims abstract description 58
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 23
- 239000010959 steel Substances 0.000 title claims abstract description 23
- 238000001514 detection method Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000009499 grossing Methods 0.000 claims abstract description 16
- 230000009466 transformation Effects 0.000 claims abstract description 11
- 230000002708 enhancing effect Effects 0.000 claims abstract description 4
- 230000002950 deficient Effects 0.000 claims description 35
- 238000005286 illumination Methods 0.000 claims description 5
- 238000013178 mathematical model Methods 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims description 3
- 239000006185 dispersion Substances 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 230000000903 blocking effect Effects 0.000 abstract 1
- 238000004519 manufacturing process Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 2
- 238000011897 real-time detection Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 208000003464 asthenopia Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000002932 luster Substances 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a rapid strip steel scratch defect detection method based on image recognition, which comprises the following steps: step 1, obtaining a high-definition image I 1 The method comprises the steps of carrying out a first treatment on the surface of the Step 2, for image I 1 Image I using improved mean shift smoothing algorithm 1 Performing background smoothing, enhancing the detail part of the scratch edge, constructing a nonlinear transformation function to improve the contrast of the image, and obtaining an image I 3 The method comprises the steps of carrying out a first treatment on the surface of the Step 3, for image I 3 Carrying out blocking processing, carrying out number sequence marking on the sub-block images, and counting the total number of gray levels, the gray average value, the deviation coefficient SK and the gray level distance D of the sub-blocks by using the gray level histogram of the sub-block images according to the sequence; step 4, judging whether each sub-block contains scratch defects, counting sub-block images containing the defects, and splicing the sub-block images with the scratches to obtain an image I 4 . The method solves the problems of overlarge algorithm complexity, unstable algorithm and poor robustness of the existing algorithm.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence, and relates to a rapid strip steel scratch defect detection method based on image recognition.
Background
With the development of technology, the high-end industries such as automobiles, aerospace, mechatronics and the like have higher and higher requirements on the surface quality of the strip steel. However, the surface quality is mainly caused by surface defects of products, wherein scratches are one of the common defects of the steel rolling surface. In the production process, scratches on the surface of the strip steel are usually long and narrow along the running direction of the strip steel, and are usually two types, namely high-temperature area scratches and normal-temperature area scratches, and the colors caused by the two scratches are different, wherein the former is bluish or black, and the latter is grey-white or metallic luster. The severity of scratches will affect the production and use of downstream users, and therefore how to detect and control the quality of strip has become a hot spot problem that strip manufacturers are highly demanding to solve.
At present, the detection means of scratches mainly comprise manual uncoiling detection and detection based on machine vision. However, due to long-time work and the influence of noisy production environment, visual fatigue can occur to human eyes, the condition of missed detection can not be avoided, and the detection effect is poor. In order to solve the problem, a scratch detection mode based on image processing is proposed to improve the efficiency of defect detection, however, due to the influence of production environment, the acquired image is clear, the defect characteristics are not obvious, great difficulty is brought to the subsequent segmentation result, the complexity of an algorithm is multiplied, and the problems of unstable algorithm and poor robustness are caused, so that real-time detection is difficult to realize.
Disclosure of Invention
The invention aims to provide a rapid strip steel scratch defect detection method based on image recognition, which solves the problems of excessive algorithm complexity, unstable algorithm and poor robustness in the prior art.
The technical scheme adopted by the invention is that the rapid strip steel scratch defect detection method based on image recognition is implemented according to the following steps:
step 1, obtaining a detection image,
under the high-brightness LED illumination condition, the upper and lower surfaces of the strip steel are acquired in real time by adopting an industrial camera to obtain a high-definition image I 1 ；
Step 2, for image I 1 The pretreatment is carried out,
image I using improved mean shift smoothing algorithm 1 Performing background smoothing, enhancing the detail part of the scratch edge, constructing a nonlinear transformation function to improve the contrast of the image, and obtaining an image I 3 ；
Step 3, for image I 3 The sub-block image is marked according to the sequence from top to bottom and the number and the serial number of the sub-block, and then the sub-block image is marked according to the sequence from top to bottomThe number of the illumination sequentially uses the gray level histogram of the sub-block image to count the total gray level { l } of the sub-block 1 ,l 2 ,...,l 64 Average gray scale { m } 1 ,m 2 ,...,m 64 -a bias coefficient SK, a gray level distance D;
step 4, judging whether each sub-block contains scratch defects or not by using the bias coefficient SK and the gray level distance D, counting sub-block images containing the defects, and splicing the sub-block images with the scratches to obtain an image I 4 The preparation method comprises the steps of.
The method has the beneficial effects of solving the problems that the acquired image is unclear, the defect characteristics are not obvious and the scratch defect detection is difficult to realize in real time due to the influence of the production environment. The invention carries out background smoothing on the image through an improved mean shift smoothing algorithm, enhances the defect characteristics of the image, omits a segmentation algorithm, directly detects the subblocks by adopting a method based on block gray level statistics, judges whether the subblocks contain defects, and then splices the subblocks, thereby avoiding the conditions of shadow, uneven illumination, different contrast, burst noise, complex background and the like in the effectively processed image and providing a new real-time detection idea for detecting the scratch defects of the strip steel.
Drawings
FIG. 1 is a flow chart of an embodiment of the method of the present invention;
FIG. 2 is a flow chart of an embodiment of improved mean shift smoothing in the method of the present invention;
FIG. 3 is a graph of a nonlinear transformation function constructed in the method of the present invention;
FIG. 4a is an example strip surface image I 1 FIG. 4b is an effect image I with improved mean shift filtering 2 ；
FIG. 5 is a graph of a mathematical model of a computational feature in the method of the present invention;
FIG. 6a is a block non-defective subgraph and its gray level histogram in the method of the present invention, and FIG. 6b is a block defective subgraph and its gray level histogram in the method of the present invention;
FIG. 7a is a preprocessed image I in the method of the present invention 3 FIG. 7b shows the partitioning result of the present inventionDefect detection result image I of method embodiment 4 。
Detailed Description
The invention will be described in detail below with reference to the drawings and the detailed description.
Referring to fig. 1, the detection method of the scratch defect degree of the strip steel based on the invention is implemented according to the following steps:
step 1, obtaining a detection image,
under the high-brightness LED illumination condition, the industrial camera is adopted to collect the upper and lower surfaces of the strip steel in real time, so that the high-definition image I of the surface of the strip steel can be obtained in the severe steel mill environment 1 ；
Step 2, for the image I obtained in step 1 1 The pretreatment is carried out,
image I using improved mean shift smoothing algorithm 1 Performing background smoothing, enhancing the detail part of the scratch edge, constructing a nonlinear transformation function to improve the image contrast, performing a mean shift smoothing algorithm flow as shown in figure 2,
2.1 Image I) 1 The pixel points in (a) include two types of information: namely, a coordinate space and a color space, coordinate value x of each pixel point is calculated s And pixel gray value x r Together form a feature vector x= (X) s ,x r ) Setting a positive number threshold sigma of each iteration smoothing, and selecting an image I 1 The first pixel point x at the upper left starts to carry out a smoothing algorithm;
2.2 The pixel point x of the current iteration is recorded as an iteration initial point x 0 The method comprises the steps of carrying out a first treatment on the surface of the Giving an initial bandwidth h at the first iteration, taking a value of 10, and then calculating the spatial bandwidth h according to the following formula for each iteration s Value range bandwidth h r And pixel x neighborhood mean, expressed as follows:
wherein ,Ts 、T r Represents the magnification and reduction coefficient for a fixed value h, x i For pixel x neighborhood pixels, i=1, 2,..n, n is the total number of pixels in the sliding window (bandwidth) range;
2.3 Let M (x) be the functional formula of the mean shift iteration, its expression is:
wherein ,for the coordinate value of the pixel point in the window, +.>Is the gray value of the pixel point in the window,
when the absolute value M (x) -x absolute value is larger than sigma, assigning M (x) to the pixel point x, and returning to the step 2.2) to continue iteration;
stopping iteration when the I M (x) -x I is less than or equal to sigma, and giving the pixel gray value of the iterative pixel point x at the moment to an initial value x 0 ；
2.4 Continuing with image I) 1 Repeating the step 2.2) and the step 2.3) for the next pixel point until the image I is traversed 1 All the pixel points are obtained to obtain a mean shift smooth image I 2 As shown in fig. 4 b;
2.5 To improve the mean-shift smoothed image I 2 A nonlinear transformation function is constructed as shown in formula (5), the function graph is shown in figure 3, the purpose of which is to keep the low gray value unchanged (i.e. background gray value), stretch the high gray value (i.e. scratch defect gray value), further promote the gray value close to 1 area (i.e. background gray area), and make the background area and defect area have higher contrast ratio, namelyThe following gray histogram feature value extraction lays a foundation more accurately, and the expression of the nonlinear transformation function is as follows:
wherein f (x, y) is the mean shift smoothed image I 2 F (x, y) is a result graph after linear transformation, and is marked as an image I 3 ；
Step 3, for the image I obtained in step 2 3 The sub-block image is divided into sub-block images of 8 x 8 as shown in fig. 7a, the sub-block images are marked with the number sequence numbers of the sub-blocks in the sequence from top to bottom and from left to right, and the gray level total number { l of the sub-blocks is counted by using the gray level histogram of the sub-block images in sequence according to the sequence numbers 1 ,l 2 ,...,l 64 Average gray scale { m } 1 ,m 2 ,...,m 64 The specific processes of the }, the bias coefficient SK and the gray level distance D are as follows:
3.1 Determining an image I) 3 Size, block order is to divide image I 3 Sequentially partitioning from top to bottom and marking sub-blocks with sequence numbers { B } 1 ,B 2 ,...,B 64 -a }; in the embodiment, since the scratch defect is 0.3 mm at most and occupies about 32 pixels on the image, the image of 512 x 512 is divided into sub-block images with the size of 64 x 64, so that the scratch area occupation ratio in the occupied sub-block is increased;
3.2 For each sub-block image, a gray histogram H (k) and a gray average value m= { m are obtained 1 ,m 2 ,...,m 64 Total number of gray levels { l } contained in sub-block 1 ,l 2 ,...,l 64 The specific steps of } are: searching sub-block images according to 0 to 255 gray levels in sequence, counting the total number containing gray levels as l and counting the number N of pixel points of each gray level k The expression of the gray histogram H (k) is:
where N is the total number of pixels of the sub-block image, k=0, 1,.. k The expression of the sub-block gray level average value m is:
3.3 As can be seen from fig. 6a and 6b, the shape of the defective sub-block image and the defective sub-block image on the gray level histogram have a significant difference, the gray level histogram of the defective sub-block image is represented in a symmetrical shape, and the gray level histogram of the defective sub-block image is represented in a right-hand shape; because the pixel value of the whole scratch defect is larger than that of the background, the scratch is counted to be on the right on the histogram, the difference is represented by a bias coefficient SK of the gray level histogram, and the bias coefficient SK is expressed as follows:
wherein N is the total number of pixels of the sub-block image, m is the sub-block gray level average value,
the judgment standard for primarily screening the suspected subblock defect image according to the gray level histogram is as follows:
a) When the gray histogram bias coefficient SK of the sub-block image is smaller than 0, the gray histogram shape is left, and the sub-block image has no scratch defect;
b) When the gray histogram bias coefficient SK of the sub-block image is equal to 0, the gray histogram is symmetrical, the sub-block image is uniform, when the scratch length is smaller, the counted image is likely to be symmetrical, and the sub-block image is classified as a sub-block with suspected scratch defects;
c) When the gray histogram deviation coefficient SK of the sub-block image is larger than 0, the larger the deviation coefficient is, the larger the deviation degree of the gray histogram to the right is, and the sub-block image is classified as a sub-block with suspected scratch defects;
3.4 A gray level distance D of the gray histogram of the sub-block image is calculated to detect the sub-block of the suspected scratch defect,
referring to fig. 6a and 6b, gray levels of non-defective sub-block images are concentrated, and background portions of sub-blocks containing scratch defects are concentrated, and the gray levels of the scratch portions are higher, and the gray level of the scratch portions are dispersed to a greater extent in the whole, so that the degree of dispersion is described by using a gray level distance, and a threshold T is determined for defective sub-block images i Judging whether the suspected sub-block image contains sub-block defect image or not (processing the image for multiple times to determine a threshold T) i Empirical values of (c), the specific judgment mode is:
wherein, the expression of the gray level distance D of the defective sub-block is:
wherein m is the average value of sub-block gray level, k is the gray level corresponding to the average value m, d i The gray level size in the sub-block, l is the total number of gray levels in the sub-block;
step 4, judging whether each sub-block contains scratch defects or not by using the bias coefficient SK and the gray level distance D obtained in the step 3, counting sub-block images containing the defects, and splicing the sub-block images with scratches to obtain an image I 4 As shown in fig. 7 b.
Counting the total number of blocks row occupied by the scratch area in the horizontal direction, the total number of blocks colum occupied in the vertical direction and the area ratio occupied by the number of blocks containing scratches, and judging the severity degree of the scratch defects according to industry standards, wherein the specific process is as follows:
4.1 Since the direction of the scratch defect is varied, the surroundings of the scratch defect are often affected when the defect occurs, thus counting all defective sub-blocks in the image I 3 Occupied in the vertical direction and in the horizontal directionThe maximum influence range of scratches can be shown by the number of blocks, and the image I 3 A mathematical model diagram of the block calculation feature is shown in fig. 5.
Because the sub-block images are numbered and arranged in the order from top to bottom and from left to right, 8 rows and 8 columns of square grids are arranged, as shown in the embodiment of fig. 5; when counting image I 3 When the number of the defective sub-blocks in the vertical direction is increased, judging whether each row contains the defective sub-blocks or not according to the defective sub-block marks judged in the step 3, and then accumulating the rows containing the defective sub-blocks to be represented by row; when counting image I 3 When the number of the defective sub-blocks in the horizontal direction is increased, judging whether each column contains the defective sub-blocks according to the defect sub-block marks judged in the step 3.3) and the step 3.4), then accumulating the columns containing the defective sub-blocks to be represented by column, and finally counting the proportion ratio of the defective sub-blocks to the total number of the sub-blocks;
4.2 Marking according to the sequence of the sub-block marking sequence numbers, splicing column by column, only keeping the defective sub-block, assigning the gray value of the non-defective sub-block area to 0 as shown in formula (11), so as to quickly mark the defective part, and the detection result is shown in fig. 7 b:
in the formula,Bi (x, y) is a sub-block image, i is a block marking sequence number;
4.3 Dividing the steel plate into severity according to the obtained characteristic quantity row, column and ratio and according to the numerical value of the characteristic quantity, judging different defect degrees of different users, and generating a defect inspection result image I 4 The preparation method comprises the steps of.
Claims (4)
1. The rapid strip steel scratch defect detection method based on image recognition is characterized by comprising the following steps of:
step 1, obtaining a detection image,
under the condition of high-brightness LED illumination, the industrial camera is adopted to collect the upper and lower surfaces of the strip steel in real time,obtaining a high definition image I 1 ；
Step 2, for image I 1 The pretreatment is carried out,
image I using improved mean shift smoothing algorithm 1 Performing background smoothing, enhancing the detail part of the scratch edge, constructing a nonlinear transformation function to improve the contrast of the image, and obtaining an image I 3 The specific process is as follows:
2.1 Image I) 1 The pixel x in (1) includes two types of information: namely, a coordinate space and a color space, coordinate value x of each pixel point is calculated s And pixel gray value x r Together form a feature vector x= (X) s ,x r ) Setting a positive number threshold sigma of each iteration smoothing, and selecting an image I 1 The first pixel point x at the upper left starts to carry out a smoothing algorithm;
2.2 The pixel point x of the current iteration is recorded as an iteration initial point x 0 The method comprises the steps of carrying out a first treatment on the surface of the Giving an initial bandwidth h at the first iteration, taking a value of 10, and then calculating the spatial bandwidth h according to the following formula for each iteration s Value range bandwidth h r And pixel x neighborhood mean, expressed as follows:
wherein ,Ts 、T r Represents the magnification and reduction coefficient for a fixed value h, x i For pixel x neighborhood pixels, i=1, 2,..n, n is the total number of pixels in the sliding window range;
2.3 Let M (x) be the functional formula of the mean shift iteration, its expression is:
wherein ,for the coordinate value of the pixel point in the window, +.>Is the gray value of the pixel point in the window,
when the absolute value M (x) -x absolute value is larger than sigma, assigning M (x) to the pixel point x, and returning to the step 2.2) to continue iteration;
stopping iteration when the I M (x) -x I is less than or equal to sigma, and assigning the pixel gray value of the pixel point x at the center of the iteration to the initial value x 0 ；
2.4 Continuing with image I) 1 Repeating the step 2.2) and the step 2.3) for the next pixel point until the image I is traversed 1 All the pixel points are obtained to obtain a mean shift smooth image I 2 ；
2.5 To improve the mean-shift smoothed image I 2 Constructing a nonlinear transformation function such as equation (5), the expression of the nonlinear transformation function being as follows:
wherein f (x, y) is the mean shift smoothed image I 2 F (x, y) is a result graph after linear transformation, and is marked as an image I 3 ；
Step 3, for image I 3 The sub-block image is marked according to the number sequence of the sub-blocks from top to bottom and from left to right, and the gray level total number { l } of the sub-blocks is counted by using the gray level histogram of the sub-block image according to the sequence 1 ,l 2 ,...,l 64 Average gray scale { m } 1 ,m 2 ,...,m 64 }, off-state coefficients SK and gray levelsA distance D;
step 4, judging whether each sub-block contains scratch defects or not by using the bias coefficient SK and the gray level distance D, counting sub-block images containing the defects, and splicing the sub-block images with the scratches to obtain an image I 4 The preparation method comprises the steps of.
2. The method for detecting the scratch defect of the rapid strip steel based on the image recognition according to claim 1, wherein in the step 3, the specific process is as follows:
3.1 Determining an image I) 3 Size, block order is to divide image I 3 Sequentially partitioning from top to bottom and marking sub-blocks with sequence numbers { B } 1 ,B 2 ,...,B 64 }；
3.2 For each sub-block image, a gray histogram H (k) and a gray average value m= { m are obtained 1 ,m 2 ,...,m 64 Total number of gray levels { l } contained in sub-block 1 ,l 2 ,...,l 64 The specific steps of } are: searching sub-block images according to 0 to 255 gray levels in sequence, counting the total number containing gray levels as l and counting the number N of pixel points of each gray level k The expression of the gray histogram H (k) is:
where N is the total number of pixels of the sub-block image, k=0, 1,.. k The expression of the sub-block gray level average value m is:
3.3 Since the pixel value of the whole scratch defect is larger than the background, the scratch is counted to be on the right of the whole histogram, the difference is represented by a bias coefficient SK of the gray level histogram, and the bias coefficient SK is expressed as follows:
wherein N is the total number of pixels of the sub-block image, m is the sub-block gray level average value,
primarily screening out suspected subblock defect images according to the gray level histogram;
3.4 A gray level distance D of the gray histogram of the sub-block image is calculated to detect the sub-block of the suspected scratch defect,
describing the degree of dispersion by gray scale distance, determining a threshold T for a defective sub-block image i Judging whether the suspected sub-block image contains a sub-block defect image or not, wherein the specific judging mode is as follows:
wherein, the expression of the gray level distance D of the defective sub-block is:
wherein m is the average value of sub-block gray level, k is the gray level corresponding to the average value m, d i And l is the total number of gray levels in the sub-block.
3. The method for detecting a scratch defect of a rapid strip steel based on image recognition according to claim 2, wherein in the step 3, the judgment standard for preliminarily screening out the suspected subblock defect image according to the gray level histogram is as follows:
a) When the gray histogram bias coefficient SK of the sub-block image is smaller than 0, the gray histogram shape is left, and the sub-block image has no scratch defect;
b) When the gray histogram bias coefficient SK of the sub-block image is equal to 0, the gray histogram is symmetrical, the sub-block image is uniform, when the scratch length is smaller, the counted image is likely to be symmetrical, and the sub-block image is classified as a sub-block with suspected scratch defects;
c) When the gray histogram bias coefficient SK of the sub-block image is larger than 0, the larger the bias coefficient is, the larger the deviation degree of the gray histogram to the right is, and the sub-block image is classified as a sub-block with suspected scratch defects.
4. The method for detecting the scratch defect of the rapid strip steel based on the image recognition according to claim 1, wherein in the step 4, the specific process is as follows:
counting the total number of blocks row occupied by the scratch area in the horizontal direction, the total number of blocks colum occupied in the vertical direction and the area ratio occupied by the number of blocks containing scratches, and judging the severity degree of the scratch defects according to industry standards, wherein the specific process is as follows:
4.1 Statistics of all defective sub-blocks in image I 3 The maximum influence range of scratches can be shown by the number of blocks occupied in the vertical direction and the horizontal direction, and an image I is obtained 3 The mathematical model of the feature is computed in blocks,
when counting image I 3 When the number of the defective sub-blocks in the vertical direction is increased, judging whether each row contains the defective sub-blocks or not according to the defective sub-block marks judged in the step 3, and then accumulating the rows containing the defective sub-blocks to be represented by row; when counting image I 3 When the number of the defective sub-blocks in the horizontal direction is increased, judging whether each column contains the defective sub-blocks according to the defect sub-block marks judged in the step 3.3) and the step 3.4), then accumulating the columns containing the defective sub-blocks to be represented by column, and finally counting the proportion ratio of the defective sub-blocks to the total number of the sub-blocks;
4.2 Marking according to the sequence of the sub-block marking sequence numbers, splicing column by column, only keeping the defective sub-block, and assigning the gray value of the non-defective sub-block area to be 0 as shown in a formula (11), so as to quickly mark out the defective part, wherein the detection result is as follows:
in the formula,Bi (x, y) is a sub-block image, i is a block marking sequence number;
4.3 Dividing the steel plate into severity according to the obtained characteristic quantity row, column and ratio and according to the numerical value of the characteristic quantity, judging different defect degrees of different users, and generating a defect inspection result image I 4 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110918572.8A CN113628189B (en) | 2021-08-11 | 2021-08-11 | Rapid strip steel scratch defect detection method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110918572.8A CN113628189B (en) | 2021-08-11 | 2021-08-11 | Rapid strip steel scratch defect detection method based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113628189A CN113628189A (en) | 2021-11-09 |
| CN113628189B true CN113628189B (en) | 2023-10-24 |
Family
ID=78384340
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110918572.8A Active CN113628189B (en) | 2021-08-11 | 2021-08-11 | Rapid strip steel scratch defect detection method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113628189B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114136992A (en) * | 2021-11-25 | 2022-03-04 | 陕西金之力电子科技有限公司 | Novel real-time online detection system and method for surface defects of strip steel |
| CN113934190B (en) * | 2021-12-17 | 2022-03-22 | 武汉春田纸品包装有限公司 | Corrugated board production quality regulation and control method based on machine vision |
| CN114693652B (en) * | 2022-04-01 | 2023-08-25 | 深圳市科宝图像技术有限公司 | Fabric Defect Detection Method Based on Gaussian Mixture Model |
| CN115049736B (en) * | 2022-08-15 | 2022-10-25 | 聊城市正晟电缆有限公司 | Method for accurately positioning surface defects of cable |
| CN115049668B (en) * | 2022-08-16 | 2022-11-29 | 江苏众联管业有限公司 | Steel strip roll mark identification method based on feature extraction |
| CN115239702B (en) * | 2022-09-19 | 2024-02-27 | 佛山市官塑科技有限公司 | Plastic film defect analysis processing method |
| CN116342589B (en) * | 2023-05-23 | 2023-08-22 | 之江实验室 | Cross-field scratch defect continuity detection method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101056465B1 (en) * | 2010-09-10 | 2011-08-12 | 주식회사 에코마이스터 | Apparatus of detecting scratch of wheel and wheel inspection system using the same |
| CN103743749A (en) * | 2014-01-03 | 2014-04-23 | 苏州吉视电子科技有限公司 | Device and method for detecting surface quality of flaky parts |
| CN106157323A (en) * | 2016-08-30 | 2016-11-23 | 西安工程大学 | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine |
| CN106204614A (en) * | 2016-07-21 | 2016-12-07 | 湘潭大学 | A kind of workpiece appearance defects detection method based on machine vision |
| CN109490316A (en) * | 2018-11-30 | 2019-03-19 | 熵智科技(深圳)有限公司 | A kind of surface defects detection algorithm based on machine vision |
| CN113034464A (en) * | 2021-03-23 | 2021-06-25 | 昆明理工大学 | Method for detecting defects of liquid crystal display in real time under multiple backgrounds |
-
2021
- 2021-08-11 CN CN202110918572.8A patent/CN113628189B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101056465B1 (en) * | 2010-09-10 | 2011-08-12 | 주식회사 에코마이스터 | Apparatus of detecting scratch of wheel and wheel inspection system using the same |
| CN103743749A (en) * | 2014-01-03 | 2014-04-23 | 苏州吉视电子科技有限公司 | Device and method for detecting surface quality of flaky parts |
| CN106204614A (en) * | 2016-07-21 | 2016-12-07 | 湘潭大学 | A kind of workpiece appearance defects detection method based on machine vision |
| CN106157323A (en) * | 2016-08-30 | 2016-11-23 | 西安工程大学 | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine |
| CN109490316A (en) * | 2018-11-30 | 2019-03-19 | 熵智科技(深圳)有限公司 | A kind of surface defects detection algorithm based on machine vision |
| CN113034464A (en) * | 2021-03-23 | 2021-06-25 | 昆明理工大学 | Method for detecting defects of liquid crystal display in real time under multiple backgrounds |
Non-Patent Citations (1)
| Title |
|---|
| 基于表面缺陷检测的金属图像识别;张立新;《技术研讨与交流》;第121-123页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113628189A (en) | 2021-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113628189B (en) | Rapid strip steel scratch defect detection method based on image recognition | |
| CN114937055B (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
| Zhu et al. | Yarn-dyed fabric defect detection based on autocorrelation function and GLCM | |
| CN115345885A (en) | Method for detecting appearance quality of metal fitness equipment | |
| CN115100221B (en) | Glass defect segmentation method | |
| CN115239704B (en) | Accurate detection and repair method for wood surface defects | |
| CN107490582B (en) | Assembly line workpiece detection system | |
| CN113554629A (en) | Strip steel red rust defect detection method based on artificial intelligence | |
| CN115100206B (en) | Printing defect identification method for textile with periodic pattern | |
| CN109215009B (en) | Continuous casting billet surface image defect detection method based on deep convolution neural network | |
| CN109472788B (en) | Method for detecting flaw on surface of airplane rivet | |
| CN105069778B (en) | Based on the industrial products detection method of surface flaw that target signature notable figure builds | |
| CN114881915A (en) | Symmetry-based mobile phone glass cover plate window area defect detection method | |
| CN114897908B (en) | Machine vision-based method and system for analyzing defects of selective laser powder spreading sintering surface | |
| CN115063407B (en) | Scratch and crack identification method for annular copper gasket | |
| CN109543686B (en) | Character recognition preprocessing binarization method based on self-adaptive multi-threshold | |
| CN111724376B (en) | Paper disease detection method based on texture feature analysis | |
| CN114926387A (en) | Weld defect detection method and device based on background estimation and edge gradient suppression | |
| CN115272336A (en) | Metal part defect accurate detection method based on gradient vector | |
| CN115272350A (en) | Method for detecting production quality of computer PCB mainboard | |
| CN114998346B (en) | Waterproof cloth quality data processing and identifying method | |
| CN110827272B (en) | Tire X-ray image defect detection method based on image processing | |
| CN113724238B (en) | Ceramic tile color difference detection and classification method based on feature point neighborhood color analysis | |
| CN114612333A (en) | Microscopic image-based microencapsulated algae cell statistical method | |
| CN117152151B (en) | Motor shell quality detection method based on machine vision |
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 |