CN117538332B - Machine vision-based steel low-power shrinkage cavity defect identification and evaluation method - Google Patents
Machine vision-based steel low-power shrinkage cavity defect identification and evaluation method Download PDFInfo
- Publication number
- CN117538332B CN117538332B CN202311487568.6A CN202311487568A CN117538332B CN 117538332 B CN117538332 B CN 117538332B CN 202311487568 A CN202311487568 A CN 202311487568A CN 117538332 B CN117538332 B CN 117538332B
- Authority
- CN
- China
- Prior art keywords
- low
- power
- defect
- image
- power image
- 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 89
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 24
- 239000010959 steel Substances 0.000 title claims abstract description 24
- 238000011156 evaluation Methods 0.000 title claims abstract description 21
- 238000009749 continuous casting Methods 0.000 claims abstract description 36
- 238000001514 detection method Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000013135 deep learning Methods 0.000 claims abstract description 7
- 238000000861 blow drying Methods 0.000 claims abstract description 4
- 239000003518 caustics Substances 0.000 claims abstract description 4
- 238000011010 flushing procedure Methods 0.000 claims abstract description 4
- 238000003754 machining Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 18
- 238000000034 method Methods 0.000 claims description 16
- 230000000694 effects Effects 0.000 claims description 13
- 238000013441 quality evaluation Methods 0.000 claims description 11
- 230000009466 transformation Effects 0.000 claims description 10
- 238000013527 convolutional neural network Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000003708 edge detection Methods 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000003746 surface roughness Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 7
- 230000008520 organization Effects 0.000 abstract description 6
- 238000011161 development Methods 0.000 abstract description 4
- 230000006872 improvement Effects 0.000 abstract description 3
- 238000011160 research Methods 0.000 abstract description 2
- 238000013461 design Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 3
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005554 pickling Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000005204 segregation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
- G01N1/32—Polishing; Etching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8883—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges involving the calculation of gauges, generating models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- 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
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biochemistry (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a machine vision-based steel low-power shrinkage cavity defect identification and evaluation method, which relates to the technical field of image processing and comprises the following steps: step 1: machining a continuous casting round billet sample, ensuring the roughness of a detection surface, corroding by using a corrosive agent, and flushing and blow-drying after corroding; step 2: collecting a low-power image of a blow-dried continuous casting round billet sample; step 3: identifying defects in a center region of the low-power image based on deep learning; step 4: for any defect, identifying the major axis and the minor axis of the defect, and judging the shrinkage defect; step 5: automatic grading is performed based on shrinkage defects. The invention can realize standardization of steel low-power shrinkage cavity detection through development of a low-power digital detection system, provide more accurate detection data for production, provide more abundant low-power organization information and better research and improvement service for production process.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a machine vision-based steel low-power shrinkage cavity defect identification and evaluation method.
Background
The low-power sample mainly comprises continuous casting square billets, continuous casting round billets, round bars and the like, the low-power quality is evaluated by means of traditional pickling and manual grading, and detection items generally comprise general porosity, center porosity, shrinkage cavity, internal cracks, segregation and the like. Because the production equipment and the process of each enterprise have the characteristics of the enterprise, the low-power organization state of the product is also different to a certain extent, the low-power organization state of the product cannot be completely corresponding to a standard map basically, many details cannot be quantified, a detector needs to understand the standard or convert the map to comprehensively evaluate the low-power organization level, and the detection is greatly influenced by human factors.
Especially, the shrinkage cavity defect is that the 1.0-3.0 level map in the YB/T153 standard map gives round (or approximate round) shrinkage cavity, and the 4.0 level is long strip shape, but in actual products, the shrinkage cavity shape is usually irregular cavity, and the condition that the shrinkage cavity is long strip-shaped and has a plurality of cavities is provided, so that the detection standard is more easily influenced by subjective experience and is not easy to unify.
Disclosure of Invention
The invention aims at overcoming the defects of the prior art, and designs a steel low-power shrinkage cavity defect identification and evaluation method based on machine vision, which can realize standardization of steel low-power shrinkage cavity detection through development of a low-power digital detection system, provide more accurate detection data for production, provide more abundant low-power organization information, and better study and improvement service of production process.
A machine vision-based steel low-shrinkage cavity defect identification and evaluation method comprises the following steps:
Step 1: machining a continuous casting round billet sample, ensuring the roughness of a detection surface, corroding by using a corrosive agent, and flushing and blow-drying after corroding;
step 2: collecting a low-power image of a blow-dried continuous casting round billet sample;
step 3: identifying defects in a center region of the low-power image based on deep learning;
Step 4: for any defect, identifying a long axis and a short axis of the defect, wherein the long axis corresponds to the actual length of the continuous casting round billet sample and is a, the short axis corresponds to the actual length of the continuous casting round billet sample and is b, a is larger than b, and when the a and the b meet the following conditions, marking the defect as a shrinkage defect:
step 5: automatic grading is performed based on shrinkage defects.
Preferably, the central area of step 3 is an area within 5% of the diameter of the continuous casting round billet in the low power image.
Identifying a region within a 5% diameter range of a continuous casting round billet in a low-power image comprises the following steps:
step 3-1: drawing a continuous casting round billet boundary line in the low-power image;
Step 3-2: finding out the two farthest points on the boundary line, and marking the two farthest points as A and B;
step 3-3: connecting the point A with the point B, finding out the midpoint of the AB line segment, and marking the midpoint as a circle center O;
step 3-4: dividing the line segment AO into 5 segments averagely, and finding out a segmentation point C closest to the O point on the line segment AO;
step 3-5: and drawing a circle by taking O as a circle center and OC as a radius, wherein the space surrounded by the circle is the central area of the low-power image.
Further, a and b satisfy the following formulas, respectively:
a=laRAB/lAB
b=lbBAB/lAB
Where l a represents the length of the long axis of the defect on the low-power image, l b represents the length of the short axis of the defect on the low-power image, R AB represents the diameter of the continuous casting round billet sample, and l AB represents the length of the line segment AB on the low-power image.
Preferably, the defect of the low-power tissue image is identified based on deep learning, and the method specifically comprises the following steps of:
step 3.1: collecting a plurality of low-power tissue image samples containing defects, and marking the defects;
step 3.2: preprocessing the low-power tissue image sample, scaling the low-power tissue image sample to a fixed size, and carrying out normalization processing;
Step 3.3: constructing a convolutional neural network model, and training the preprocessed low-power tissue image sample by using the convolutional neural network;
step 3.4: using another group of low-power tissue image samples as test data to evaluate the precision of the convolutional neural network model, adding the low-power tissue image samples when the precision does not meet the requirement, and repeating the steps 3.1-3.4;
step 3.5: and preprocessing the collected low-power tissue image, and reasoning and labeling defects of the low-power tissue image based on ResNeSt network models.
Preferably, step 1 detects a surface roughness Ra < 0.1um.
Further, identifying the major and minor axes of the defect includes operating to: firstly, the farthest two points P and Q on the defect boundary are found out, and a line segment PQ is used as the long axis of the defect; finding out the midpoint of the PQ, and marking the midpoint as a T point; a straight line perpendicular to PQ and passing through the T point is made, the intersection of the straight line and the defect boundary is noted J and K, and the line segment JK serves as the short axis of the defect.
Further, the automatic rating formula is:
Wherein T represents a rank score, n represents the number of shrinkage defects in the center region of the low-power image, a i represents the actual length of the continuous casting round billet sample corresponding to the long axis of the ith defect, and b i represents the actual length of the continuous casting round billet sample corresponding to the short axis of the ith defect.
Preferably, in step 2, local image enhancement processing is performed on the low-power image by using an LGE model transformation function, where the LGE model transformation function is:
Wherein f (u, v) represents the gray value of the point of the low-power image (u, v), u is the abscissa, v is the ordinate, g (u, v) represents the gray value of the point of the low-power image (u, v) after enhancement, T LGE [ ] represents the LGE model transformation, gm represents the global average gray value, m (u, v) represents the average gray value in the window, sigma (u, v) represents the local standard deviation after windowing, parameter a is the low-power image enhancement effect index factor, parameter b is the low-power image enhancement effect score factor, parameter c * is the low-power image enhancement effect inverse multiplier factor, parameter k is the low-power image enhancement effect forward multiplier factor, and parameters a, b, c * and k are all empirical values.
Preferably, step 2 adopts a low-power image quality evaluation function to evaluate the quality of the low-power image, if the evaluation result is smaller than the threshold value, step 3 is continued, if the evaluation result is larger than the threshold value, the low-power image of the blow-dried continuous casting round billet sample is collected again,
Wherein, the low-power image quality evaluation function is:
Where F () represents a low-power image quality evaluation function, edge (Ie) represents the number of edge pixels obtained by the first-order differential edge detection operator, E (Ie) represents the low-power image edge intensity, ie represents the enhanced low-power image, H (Ie) represents the low-power image entropy value, M I is the low-power image maximum abscissa number, and N I is the low-power image maximum ordinate number.
Compared with the prior art, the invention has the beneficial effects that:
The invention designs a machine vision-based steel low-power shrinkage cavity defect identification and evaluation method, which can realize standardization of steel low-power shrinkage cavity detection through development of a low-power digital detection system, provide more accurate detection data for production, provide more abundant low-power organization information and better research and improvement service of a production process.
1. The invention does not need to consider the difference of shrinkage cavity shapes, greatly simplifies the design difficulty of the scheme and is convenient for unification of standards.
2. The invention creatively provides the method for evaluating the defects by using the major axis and the minor axis of the defects as key indexes, and designs a corresponding automatic grading scheme, thereby having strong pertinence and high precision.
3. The invention adopts the LGE model transformation function to carry out local image enhancement processing on the low-power image, carries out control on the quality of the low-power image based on the low-power image quality evaluation function, and improves the processing effect of the low-power image.
Drawings
FIG. 1 is a representation of a low magnification tissue shrinkage cavity image and its corresponding defect in accordance with the present invention.
Detailed Description
The invention further provides a machine vision-based steel low shrinkage cavity defect identification and evaluation method by combining the drawings and a specific implementation method.
The invention provides a machine vision-based steel macroscopic aperture digital detection method, which aims to integrate sample preparation conveying, image acquisition, defect identification and data arrangement functions, improve the automation, intellectualization, precision and safety degree of a detection process, construct a high-efficiency, accurate and objective steel sample macroscopic tissue quality evaluation system, and provide technical support for short flow, automation and intellectualization development of steel industry product quality detection.
A method for identifying and evaluating steel low-shrinkage cavity defects based on machine vision is shown in fig. 1, and comprises the following steps:
Step 1: machining a continuous casting round billet sample, ensuring the roughness of a detection surface, corroding by using a corrosive agent, and flushing and blow-drying after corroding;
step 2: collecting a low-power image of a blow-dried continuous casting round billet sample;
step 3: identifying defects in a center region of the low-power image based on deep learning;
Step 4: for any defect, identifying a long axis and a short axis of the defect, wherein the long axis corresponds to the actual length of the continuous casting round billet sample and is a, the short axis corresponds to the actual length of the continuous casting round billet sample and is b, a is larger than b, and when the a and the b meet the following conditions, marking the defect as a shrinkage defect:
step 5: automatic grading is performed based on shrinkage defects. Where a and b are both in mm.
Preferably, the central area of step 3 is an area within 5% of the diameter of the continuous casting round billet in the low power image.
Identifying a region within a 5% diameter range of a continuous casting round billet in a low-power image comprises the following steps:
step 3-1: drawing a continuous casting round billet boundary line in the low-power image;
Step 3-2: finding out the two farthest points on the boundary line, and marking the two farthest points as A and B;
step 3-3: connecting the point A with the point B, finding out the midpoint of the AB line segment, and marking the midpoint as a circle center O;
step 3-4: dividing the line segment AO into 5 segments averagely, and finding out a segmentation point C closest to the O point on the line segment AO;
step 3-5: and drawing a circle by taking O as a circle center and OC as a radius, wherein the space surrounded by the circle is the central area of the low-power image.
Further, a and b satisfy the following formulas, respectively:
a=laBAB/lAB
b=lbRAB/lAB
Where l a represents the length of the long axis of the defect on the low-power image, l b represents the length of the short axis of the defect on the low-power image, R AB represents the diameter of the continuous casting round billet sample, and l AB represents the length of the line segment AB on the low-power image.
Preferably, the defect of the low-power tissue image is identified based on deep learning, and the method specifically comprises the following steps of:
step 3.1: collecting a plurality of low-power tissue image samples containing defects, and marking the defects;
step 3.2: preprocessing the low-power tissue image sample, scaling the low-power tissue image sample to a fixed size, and carrying out normalization processing;
Step 3.3: constructing a convolutional neural network model, and training the preprocessed low-power tissue image sample by using the convolutional neural network;
step 3.4: using another group of low-power tissue image samples as test data to evaluate the precision of the convolutional neural network model, adding the low-power tissue image samples when the precision does not meet the requirement, and repeating the steps 3.1-3.4;
step 3.5: and preprocessing the collected low-power tissue image, and reasoning and labeling defects of the low-power tissue image based on ResNeSt network models.
Preferably, step 1 detects a surface roughness Ra < 0.1um.
Further, identifying the major and minor axes of the defect includes operating to: firstly, the farthest two points P and Q on the defect boundary are found out, and a line segment PQ is used as the long axis of the defect; finding out the midpoint of the PQ, and marking the midpoint as a T point; a straight line perpendicular to PQ and passing through the T point is made, the intersection of the straight line and the defect boundary is noted J and K, and the line segment JK serves as the short axis of the defect.
Further, the automatic rating formula is:
Wherein T represents a rank score, n represents the number of shrinkage defects in the center region of the low-power image, a i represents the actual length of the continuous casting round billet sample corresponding to the long axis of the ith defect, and b i represents the actual length of the continuous casting round billet sample corresponding to the short axis of the ith defect.
Preferably, in step 2, local image enhancement processing is performed on the low-power image by using an LGE model transformation function, where the LGE model transformation function is:
Wherein f (u, v) represents the gray value of the point of the low-power image (u, v), u is the abscissa, v is the ordinate, g (u, v) represents the gray value of the point of the low-power image (u, v) after enhancement, T LGE [ ] represents the LGE model transformation, gm represents the global average gray value, m (u, v) represents the average gray value in the window, sigma (u, v) represents the local standard deviation after windowing, parameter a is the low-power image enhancement effect index factor, parameter b is the low-power image enhancement effect score factor, parameter c * is the low-power image enhancement effect inverse multiplier factor, parameter k is the low-power image enhancement effect forward multiplier factor, and parameters a, b, c * and k are all empirical values.
Preferably, step 2 adopts a low-power image quality evaluation function to evaluate the quality of the low-power image, if the evaluation result is smaller than the threshold value, step 3 is continued, if the evaluation result is larger than the threshold value, the low-power image of the blow-dried continuous casting round billet sample is collected again,
Wherein, the low-power image quality evaluation function is:
Where F () represents a low-power image quality evaluation function, edge (Ie) represents the number of edge pixels obtained by the first-order differential edge detection operator, E (Ie) represents the low-power image edge intensity, ie represents the enhanced low-power image, H (Ie) represents the low-power image entropy value, M I is the low-power image maximum abscissa number, and N I is the low-power image maximum ordinate number.
The present invention provides a method for identifying and evaluating defects of low shrinkage cavities of steel based on machine vision, and the above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement the same, and not to limit the protection scope of the present invention. All equivalent changes or modifications made in accordance with the spirit of the present invention should be construed to be included in the scope of the present invention.
Claims (8)
1. A machine vision-based steel low-shrinkage cavity defect identification and evaluation method is characterized by comprising the following steps:
Step 1: machining a continuous casting round billet sample, ensuring the roughness of a detection surface, corroding by using a corrosive agent, and flushing and blow-drying after corroding;
step 2: collecting a low-power image of a blow-dried continuous casting round billet sample;
step 3: identifying defects in a center region of the low-power image based on deep learning;
Step 4: for any defect, identifying a long axis and a short axis of the defect, wherein the long axis corresponds to the actual length of the continuous casting round billet sample and is a, the short axis corresponds to the actual length of the continuous casting round billet sample and is b, a is larger than b, and when the a and the b meet the following conditions, marking the defect as a shrinkage defect:
Step 5: automatically grading based on shrinkage cavity defects;
the defect of the low-power tissue image is identified based on deep learning, and the method specifically comprises the following steps of:
step 3.1: collecting a plurality of low-power tissue image samples containing defects, and marking the defects;
step 3.2: preprocessing the low-power tissue image sample, scaling the low-power tissue image sample to a fixed size, and carrying out normalization processing;
Step 3.3: constructing a convolutional neural network model, and training the preprocessed low-power tissue image sample by using the convolutional neural network;
step 3.4: using another group of low-power tissue image samples as test data to evaluate the precision of the convolutional neural network model, adding the low-power tissue image samples when the precision does not meet the requirement, and repeating the steps 3.1-3.4;
step 3.5: preprocessing the collected low-power tissue image, and reasoning and labeling defects of the low-power tissue image based on ResNeSt network models;
The automatic rating formula is:
Wherein T represents a rank score, n represents the number of shrinkage defects in the center region of the low-power image, a i represents the actual length of the continuous casting round billet sample corresponding to the long axis of the ith defect, and b i represents the actual length of the continuous casting round billet sample corresponding to the short axis of the ith defect.
2. The machine vision-based steel product low shrinkage cavity defect identification and evaluation method as claimed in claim 1, wherein the method is characterized by comprising the following steps:
and 3, the central area in the step is an area in the range of 5% of the diameter of the continuous casting round billet around the circle center in the low-power image.
3. The machine vision-based steel product low shrinkage cavity defect identification and evaluation method as claimed in claim 2, wherein the method is characterized by comprising the following steps:
Identifying the center region of the low power image includes the steps of:
step 3-1: drawing a continuous casting round billet boundary line in the low-power image;
Step 3-2: finding out the two farthest points on the boundary line, and marking the two farthest points as A and B;
step 3-3: connecting the point A with the point B, finding out the midpoint of the AB line segment, and marking the midpoint as a circle center O;
step 3-4: dividing the line segment AO into 5 segments averagely, and finding out a segmentation point C closest to the O point on the line segment AO;
step 3-5: and drawing a circle by taking O as a circle center and OC as a radius, wherein the space surrounded by the circle is the central area of the low-power image.
4. The method for identifying and evaluating the defects of the steel low shrinkage cavity based on the machine vision according to claim 3, wherein a and b respectively satisfy the following formulas:
a=laRAB/lAB
b=lbRAB/lAB
Where l a represents the length of the long axis of the defect on the low-power image, l b represents the length of the short axis of the defect on the low-power image, R AB represents the diameter of the continuous casting round billet sample, and l AB represents the length of the line segment AB on the low-power image.
5. The machine vision-based steel product low shrinkage cavity defect identification and evaluation method as claimed in claim 1, wherein the method is characterized by comprising the following steps:
and step 1, detecting the surface roughness Ra < 0.1um.
6. The machine vision-based steel product low shrinkage cavity defect identification and evaluation method as set forth in claim 4, wherein the method is characterized in that:
identifying the major and minor axes of the defect includes operating to: firstly, the farthest two points P and Q on the defect boundary are found out, and a line segment PQ is used as the long axis of the defect; finding out the midpoint of the PQ, and marking the midpoint as a T point; a straight line perpendicular to PQ and passing through the T point is made, the intersection of the straight line and the defect boundary is noted J and K, and the line segment JK serves as the short axis of the defect.
7. The machine vision-based steel product low shrinkage cavity defect identification and evaluation method as claimed in claim 1, wherein the method is characterized by comprising the following steps:
step 2, local image enhancement processing is carried out on the low-power image by adopting an LGE model transformation function, wherein the LGE model transformation function is as follows:
Wherein f (u, v) represents the gray value of the point of the low-power image (u, v), u is the abscissa, v is the ordinate, g (u, v) represents the gray value of the point of the low-power image (u, v) after enhancement, T LGE [ ] represents the LGE model transformation, gm represents the global average gray value, m (u, v) represents the average gray value in the window, sigma (u, v) represents the local standard deviation after windowing, parameter a is the low-power image enhancement effect index factor, parameter b is the low-power image enhancement effect score factor, parameter c * is the low-power image enhancement effect inverse multiplier factor, parameter k is the low-power image enhancement effect forward multiplier factor, and parameters a, b, c * and k are all empirical values.
8. The machine vision-based steel product low shrinkage cavity defect identification and evaluation method as claimed in claim 1, wherein the method is characterized by comprising the following steps:
Step 2, adopting a low-power image quality evaluation function to evaluate the quality of the low-power image, continuing to step 3 if the evaluation result is smaller than the threshold value, re-collecting the low-power image of the blow-dried continuous casting round billet sample if the evaluation result is larger than the threshold value,
Wherein, the low-power image quality evaluation function is:
Where F () represents a low-power image quality evaluation function, edge (Ie) represents the number of edge pixels obtained by the first-order differential edge detection operator, E (Ie) represents the low-power image edge intensity, ie represents the enhanced low-power image, H (Ie) represents the low-power image entropy value, M I is the low-power image maximum abscissa number, and N I is the low-power image maximum ordinate number.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311487568.6A CN117538332B (en) | 2023-11-09 | 2023-11-09 | Machine vision-based steel low-power shrinkage cavity defect identification and evaluation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311487568.6A CN117538332B (en) | 2023-11-09 | 2023-11-09 | Machine vision-based steel low-power shrinkage cavity defect identification and evaluation method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117538332A CN117538332A (en) | 2024-02-09 |
CN117538332B true CN117538332B (en) | 2024-05-03 |
Family
ID=89781800
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311487568.6A Active CN117538332B (en) | 2023-11-09 | 2023-11-09 | Machine vision-based steel low-power shrinkage cavity defect identification and evaluation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117538332B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118135334B (en) * | 2024-04-30 | 2024-07-23 | 华东交通大学 | Method and system for identifying faults of catenary hanger |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109191459A (en) * | 2018-09-30 | 2019-01-11 | 东北大学 | The automatic identification and ranking method of continuous casting billet macrostructure center segregation defect |
CN112669307A (en) * | 2021-01-06 | 2021-04-16 | 大冶特殊钢有限公司 | Low-power acid etching defect automatic identification and rating method and system based on machine vision |
CN115541601A (en) * | 2022-11-30 | 2022-12-30 | 江苏永钢集团有限公司 | Steel low-power digital detection method based on machine vision |
CN116309598A (en) * | 2023-05-23 | 2023-06-23 | 江苏永钢集团有限公司 | Intelligent grading method for steel continuous casting billet macrostructure cracks based on machine vision |
-
2023
- 2023-11-09 CN CN202311487568.6A patent/CN117538332B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109191459A (en) * | 2018-09-30 | 2019-01-11 | 东北大学 | The automatic identification and ranking method of continuous casting billet macrostructure center segregation defect |
CN112669307A (en) * | 2021-01-06 | 2021-04-16 | 大冶特殊钢有限公司 | Low-power acid etching defect automatic identification and rating method and system based on machine vision |
WO2022117118A1 (en) * | 2021-01-06 | 2022-06-09 | 大冶特殊钢有限公司 | Machine vision-based automatic identification and rating method and system for low-magnification acid etching defect |
CN115541601A (en) * | 2022-11-30 | 2022-12-30 | 江苏永钢集团有限公司 | Steel low-power digital detection method based on machine vision |
CN116309598A (en) * | 2023-05-23 | 2023-06-23 | 江苏永钢集团有限公司 | Intelligent grading method for steel continuous casting billet macrostructure cracks based on machine vision |
Also Published As
Publication number | Publication date |
---|---|
CN117538332A (en) | 2024-02-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN117538332B (en) | Machine vision-based steel low-power shrinkage cavity defect identification and evaluation method | |
CN115082467B (en) | Building material welding surface defect detection method based on computer vision | |
CN115082418B (en) | Precise identification method for automobile parts | |
CN114354639B (en) | Weld defect real-time detection method and system based on 3D point cloud | |
CN109472761B (en) | Cell counting method and system based on fluorescence image | |
CN106251361A (en) | A kind of rail surface defects image self-adapting division method | |
CN108596038B (en) | Method for identifying red blood cells in excrement by combining morphological segmentation and neural network | |
CN111179273A (en) | Method and system for automatically segmenting leucocyte nucleoplasm based on deep learning | |
CN109215009B (en) | Continuous casting billet surface image defect detection method based on deep convolution neural network | |
CN107832801B (en) | Model construction method for cell image classification | |
CN114067114B (en) | Area nuclear segmentation counting method based on algae cell morphology | |
CN113487533B (en) | Part assembly quality digital detection system and method based on machine learning | |
US20230184703A1 (en) | Quantitative statistical characterization method of micron-level second phase in aluminum alloy based on deep learning | |
CN109740254B (en) | Ship diesel engine abrasive particle type identification method based on information fusion | |
CN113962994B (en) | Method for detecting cleanliness of lock pin on three-connecting-rod based on image processing | |
CN115797361B (en) | Aluminum template surface defect detection method | |
CN110287456A (en) | Large coil rolling surface defect analysis method based on data mining | |
CN118097310B (en) | Method for digitally detecting concrete surface defects | |
CN115272323A (en) | Data intelligent regulation and control acquisition method for traffic engineering pavement quality detection | |
CN114155226A (en) | Micro defect edge calculation method | |
CN116309598B (en) | Intelligent grading method for steel continuous casting billet macrostructure cracks based on machine vision | |
CN112288000A (en) | Intelligent ferrographic image identification method based on support vector machine | |
CN116958964A (en) | Data knowledge dual-drive diatom image classification system and method | |
CN113989209B (en) | Power line foreign matter detection method based on Faster R-CNN | |
CN113298148A (en) | Ecological environment evaluation-oriented unbalanced data resampling method |
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 |