CN115494074A - Online detection method for surface defects of continuous casting slab - Google Patents
Online detection method for surface defects of continuous casting slab Download PDFInfo
- Publication number
- CN115494074A CN115494074A CN202211288689.3A CN202211288689A CN115494074A CN 115494074 A CN115494074 A CN 115494074A CN 202211288689 A CN202211288689 A CN 202211288689A CN 115494074 A CN115494074 A CN 115494074A
- Authority
- CN
- China
- Prior art keywords
- defect
- continuous casting
- casting slab
- defects
- detection method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- 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
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
-
- 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
Abstract
The invention discloses an online detection method for surface defects of a continuous casting slab, which adopts an imaging system consisting of a high-speed area array industrial camera, an LED stroboscopic light source, a linear array industrial camera and a linear array high-brightness light source to continuously and synchronously shoot the surface of the continuous casting slab to form a high-definition image to be acquired; detecting high-definition images containing defects by a vision processor, and obtaining defect types identified by initial artificial intelligence through a deep learning model; combining the defect merging, periodic calculation and expert post-processing rules of the server unit to obtain a final defect classification result, storing and alarming for output; the human-computer interface displays corresponding information. The method overcomes the defects of the traditional continuous casting slab surface defect detection, accurately positions the defects by using line-surface combination, analyzes the defect characteristics through deep learning, improves the defect identification precision, can accurately, timely and effectively detect and identify the concerned defects, and improves the quality of the continuous casting slab.
Description
Technical Field
The invention relates to the technical field of detection, in particular to an online detection method for surface defects of a continuous casting slab.
Background
The real-time online detection of the surface defects of the continuous casting slabs is crucial to the quality of the continuous casting slabs, most of the existing surface defect detection systems only adopt an area-array scanning camera to detect the defects, and after pictures are taken on site, detection personnel judge whether the defects exist or not in a mode of looking through human eyes and estimate the positions of the defects; some fields adopt an image recognition technology, and some fields adopt a mode of taking a plurality of groups of pictures by a face scanning camera and then splicing. The above method has the problem of inaccurate defect positioning, the detected defects need to manually select the characteristics in advance, and the quality of the characteristic selection has a great influence on the defect identification effect, so that the possibility of missing detection or false detection exists, the detection efficiency and precision of the surface defects of the continuous casting slab are seriously influenced, and the improvement of the quality of the continuous casting slab is hindered.
Disclosure of Invention
The invention aims to solve the technical problem of providing an online detection method for the surface defects of the continuous casting slabs, which overcomes the defects of the traditional continuous casting slab surface defect detection, accurately positions the defects by using line-surface combination, analyzes the defect characteristics by deep learning, improves the defect identification precision, can accurately, timely and effectively detect and identify the concerned defects, and improves the quality of the continuous casting slabs.
In order to solve the technical problem, the online detection method for the surface defects of the continuous casting slab comprises the following steps:
the method comprises the following steps that firstly, an imaging system consisting of a high-speed area array industrial camera, an LED stroboscopic light source, a linear array industrial camera and a linear array high-brightness light source is adopted to continuously and synchronously shoot the surface of a continuous casting plate blank to form a high-definition image to be acquired;
secondly, the high-definition images are transmitted to a visual processor, the visual processor analyzes and processes the high-definition images, high-definition images with defects are detected, and image segmentation and artificial intelligence recognition based on a deep learning model are carried out on the high-definition images with the defects to obtain defect types of initial artificial intelligence recognition;
thirdly, combining the identified defect categories with defect merging, periodic calculation and expert post-processing rules of the server unit to obtain a final defect classification result, and storing a database, storing a high-definition image and outputting an alarm according to the defect classification result;
and fourthly, displaying the defect type, the high-definition image, the defect position and the defect grade information for an operator through an HMI (human machine interface).
Furthermore, the imaging system is deployed in the upper surface space and the lower surface space of the continuous casting slab conveying roller way area, and after the continuous casting slab enters the detection area, the imaging system is triggered to carry out high-definition image acquisition on the upper surface and the lower surface of the continuous casting slab based on the trigger signal of the material tracking sensor and the continuous casting slab speed signal.
Further, the HMI human-machine interface comprises an online mode and an offline mode, wherein the online mode is used for displaying production information and latest defect data, including a material list, a defect real object diagram, a defect simulation diagram and unit speed information, the default state of the online mode is a tracking mode, a pause mode can be selected in the tracking mode, and at the moment, only data are received without refreshing an interface, so that a user can select to view interested data in the list; the off-line mode is used for providing a searching function of historical production information and defect information, and the terminal program can acquire a historical defect information record from a database of the detection server and download a defect image to the local for caching and displaying.
Further, the deep learning model acquires high-definition image data with defects of a new continuous casting slab from a database, performs multiple convolution calculations on the high-definition image data to extract defect characteristics, performs defect detection frame generation and classification calculation processes according to the extracted defect characteristics, uploads the position of the defect detection frame and defect category information to the database, and completes defect detection based on the deep learning model.
Further, the imaging system carries out preprocessing on the acquired high-definition images, including filtering, corrosion and expansion operator processing, extracts the edges of the two sides of the continuous casting slab after eliminating the interference influence of external impurities, and measures and obtains the width of the continuous casting slab.
Furthermore, the high-speed area array industrial camera and the linear array industrial camera are provided with infrared filters for filtering infrared wavelength light generated by the surface thermal radiation of the continuous casting slab.
The method for on-line detection of the surface defects of the continuous casting slab adopts the technical scheme, namely, the method adopts an imaging system consisting of a high-speed area-array industrial camera, an LED stroboscopic light source, a linear array industrial camera and a linear array high-brightness light source to continuously and synchronously shoot the surface of the continuous casting slab to form a high-definition image to be acquired; detecting high-definition images containing defects by a vision processor, and obtaining defect types identified by initial artificial intelligence through a deep learning model; combining the defect merging, the periodic calculation and the expert post-processing rule of the server unit to obtain a final defect classification result, and storing and alarming for output; the human-computer interface displays corresponding information. The method overcomes the defects of the traditional continuous casting slab surface defect detection, accurately positions the defects by using line-surface combination, analyzes the defect characteristics through deep learning, improves the defect identification precision, can accurately, timely and effectively detect and identify the concerned defects, and improves the quality of the continuous casting slab.
Drawings
The invention is described in further detail below with reference to the following figures and embodiments:
FIG. 1 is a flow chart of the online detection method for the surface defects of the continuous casting slabs.
Detailed Description
Embodiment example as shown in fig. 1, the method for on-line detection of surface defects of a continuous casting slab according to the present invention comprises the steps of:
the method comprises the following steps that firstly, an imaging system consisting of a high-speed area array industrial camera, an LED stroboscopic light source, a linear array industrial camera and a linear array high-brightness light source is adopted to continuously and synchronously shoot the surface of a continuous casting plate blank to form a high-definition image to be collected;
secondly, the high-definition images are transmitted to a visual processor, the visual processor analyzes and processes the high-definition images, high-definition images with defects are detected, and image segmentation and artificial intelligence recognition based on a deep learning model are carried out on the high-definition images with the defects to obtain defect types of initial artificial intelligence recognition;
thirdly, combining the identified defect categories with defect merging, periodic calculation and expert post-processing rules of the server unit to obtain a final defect classification result, and storing a database, storing a high-definition image and outputting an alarm according to the defect classification result;
and fourthly, displaying the defect type, the high-definition image, the defect position and the defect grade information for an operator through an HMI (human machine interface).
Preferably, the imaging system is deployed in the upper surface space and the lower surface space of the continuous casting slab conveying roller way area, and after the continuous casting slab enters the detection area, the imaging system is triggered to carry out high-definition image acquisition on the upper surface and the lower surface of the continuous casting slab based on the trigger signal of the material tracking sensor and the speed signal of the continuous casting slab.
Preferably, the HMI human-machine interface comprises an online mode and an offline mode, wherein the online mode is used for displaying production information and latest defect data, and comprises a material list, a defect real object diagram, a defect simulation diagram and unit speed information, the online mode is in a tracking mode in a default state, a pause mode can be selected in the tracking mode, and at the moment, only data is received but the interface is not refreshed, so that a user can select to view interested data in the list; the off-line mode is used for providing a searching function of historical production information and defect information, and the terminal program can acquire a historical defect information record from a database of the detection server and download a defect image to the local for caching and displaying.
Preferably, the deep learning model acquires high-definition image data of a new continuous casting slab with defects from the database, performs multiple convolution calculations on the high-definition image data to extract defect characteristics, performs defect detection frame generation and classification calculation processes according to the extracted defect characteristics, uploads the position of the defect detection frame and defect category information to the database, and completes defect detection based on the deep learning model.
The method has defect judgment knowledge through mass defect picture training, the deep learning model becomes an artificial intelligence model with defect judgment knowledge, the method has the capability of automatically classifying and judging defects, and the method can detect various continuous casting slab defects in a full-coverage real-time manner. In addition, the error result can be periodically corrected through man-machine interaction, and the correction result is fed back to the deep learning model, so that the artificial intelligence model can strengthen expert knowledge, and the defect judgment effect is improved.
Preferably, the imaging system carries out pretreatment on the acquired high-definition images, including filtering, corrosion and expansion operator treatment, extracts the edges of the two sides of the continuous casting slab after eliminating the interference influence of external impurities, and measures and obtains the width of the continuous casting slab.
The corrosion and expansion mainly consider that the measured value is not the actual plate width due to the fact that pits, impurities and the like exist at the edge of a continuous casting slab, so that the pits need to be expanded to be flush with the edge when high-definition images are processed, the impurities protruding out of the edge are corroded, and the actual data of the plate width needed on site can be measured.
Preferably, the high-speed area-array industrial camera and the line-array industrial camera are provided with infrared filters for filtering infrared wavelength light generated by thermal radiation on the surface of the continuous casting slab.
The method is based on machine vision AI detection technology, obtains a target image in a non-contact mode through an area array and a linear array image sensor, processes image data through a special algorithm, obtains the characteristics of the target image, analyzes and identifies the characteristics and judges the defects of the continuous casting slab, wherein the defect detection classification adopts a deep learning mode based on GPU operation, can overcome the influence of the complex state of the surface of the continuous casting slab on the defect identification, can accurately, timely and effectively detect and identify the concerned defects, and has the function of measuring the width of the continuous casting slab.
In the method, the imaging system and the corresponding electromechanical complete hardware design are subject to the high-temperature red hot surface of the continuous casting slab, and can stably work in a severe working condition environment. The method comprehensively applies the multidisciplinary knowledge of optics, electromechanics, mode recognition, AI detection and the like, and is advanced equipment for intelligent application of machine vision AI detection in the steel industry.
The method adopts an imaging system design combining a high-speed area array industrial camera and a linear array industrial camera to ensure that high resolution and clear full-width image data are acquired; realizing real-time full-width full-length non-contact detection of the continuous casting slab to identify the surface defects of the continuous casting slab and carry out grade division; the concerned defects can be accurately, timely and effectively detected and identified, and the quality of the continuous casting slab is improved.
Claims (6)
1. An on-line detection method for surface defects of continuous casting slabs is characterized by comprising the following steps:
the method comprises the following steps that firstly, an imaging system consisting of a high-speed area array industrial camera, an LED stroboscopic light source, a linear array industrial camera and a linear array high-brightness light source is adopted to continuously and synchronously shoot the surface of a continuous casting plate blank to form a high-definition image to be collected;
secondly, the high-definition images are transmitted to a vision processor, the vision processor analyzes and processes the high-definition images, high-definition images with defects are detected, and image segmentation and artificial intelligence identification based on a deep learning model are carried out on the high-definition images with the defects to obtain defect types of initial artificial intelligence identification;
step three, combining the identified defect types with defect merging, periodic calculation and expert post-processing rules of the server unit to obtain a final defect classification result, and performing database storage, high-definition image storage and alarm output on the defect classification result;
and fourthly, displaying the defect type, the high-definition image, the defect position and the defect grade information for an operator through an HMI (human machine interface).
2. The on-line detection method for the surface defects of the continuous casting slab as set forth in claim 1, wherein: the imaging system is arranged in the upper surface space and the lower surface space of the continuous casting slab conveying roller way area, and after the continuous casting slab enters the detection area, the imaging system is triggered to acquire high-definition images of the upper surface and the lower surface of the continuous casting slab based on the trigger signal of the material tracking sensor and the speed signal of the continuous casting slab.
3. The on-line detection method for the surface defects of the continuous casting slab as set forth in claim 1, wherein: the HMI human-computer interface comprises an online mode and an offline mode, wherein the online mode is used for displaying production information and latest defect data, and comprises a material list, a defect real object diagram, a defect simulation diagram and unit speed information, the default state of the online mode is a tracking mode, a pause mode can be selected in the tracking mode, and at the moment, only data is received, but the interface is not refreshed, so that a user can select and view interested data in the list; the off-line mode is used for providing a searching function of historical production information and defect information, and the terminal program can acquire a historical defect information record from a database of the detection server and download a defect image to the local for caching and displaying.
4. The on-line detection method for the surface defects of the continuous casting slab as set forth in claim 1, wherein: the deep learning model acquires high-definition image data of new continuous casting plate blanks with defects from a database, performs multiple convolution calculations on the high-definition image data to extract defect characteristics, performs defect detection frame generation and classification operation processes according to the extracted defect characteristics, uploads the position and defect category information of the defect detection frame to the database, and completes defect detection based on the deep learning model.
5. The on-line detection method for the surface defects of the continuous casting slab according to claim 1, characterized in that: and the imaging system is used for preprocessing the acquired high-definition image, including filtering, corrosion and expansion operator processing, extracting the edges of the two sides of the continuous casting slab after eliminating the interference influence of external impurities, and measuring and obtaining the width of the continuous casting slab.
6. The on-line detection method for the surface defects of the continuous casting slab as set forth in claim 1, wherein: the high-speed area array industrial camera and the linear array industrial camera are provided with infrared filters for filtering infrared wavelength light generated by the surface thermal radiation of the continuous casting slab.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211288689.3A CN115494074A (en) | 2022-10-20 | 2022-10-20 | Online detection method for surface defects of continuous casting slab |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211288689.3A CN115494074A (en) | 2022-10-20 | 2022-10-20 | Online detection method for surface defects of continuous casting slab |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115494074A true CN115494074A (en) | 2022-12-20 |
Family
ID=84474453
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211288689.3A Pending CN115494074A (en) | 2022-10-20 | 2022-10-20 | Online detection method for surface defects of continuous casting slab |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115494074A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116429170A (en) * | 2023-03-18 | 2023-07-14 | 宝钢工程技术集团有限公司 | Quality detection method for plate blank |
-
2022
- 2022-10-20 CN CN202211288689.3A patent/CN115494074A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116429170A (en) * | 2023-03-18 | 2023-07-14 | 宝钢工程技术集团有限公司 | Quality detection method for plate blank |
| CN116429170B (en) * | 2023-03-18 | 2024-02-20 | 宝钢工程技术集团有限公司 | Quality detection method for plate blank |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yun et al. | Automated defect inspection system for metal surfaces based on deep learning and data augmentation | |
| CN104914111B (en) | A kind of steel strip surface defect online intelligent recognition detecting system and its detection method | |
| US7495758B2 (en) | Apparatus and methods for two-dimensional and three-dimensional inspection of a workpiece | |
| CN110619620B (en) | Method, device and system for positioning abnormity causing surface defects and electronic equipment | |
| CN104992449A (en) | Information identification and surface defect on-line detection method based on machine visual sense | |
| CN102529019B (en) | Method for mould detection and protection as well as part detection and picking | |
| CN110991360B (en) | Robot inspection point position intelligent configuration method based on visual algorithm | |
| CN111591715A (en) | Belt longitudinal tearing detection method and device | |
| EP1867979A1 (en) | Method and apparatus for recognizing repeating patterns | |
| CN115494074A (en) | Online detection method for surface defects of continuous casting slab | |
| CN108364306B (en) | Visual real-time detection method for high-speed periodic motion | |
| CN111915649A (en) | Strip steel moving target tracking method under shielding condition | |
| CN110873718A (en) | Steel plate surface defect detection system and method based on machine vision | |
| Wu et al. | Design of online surface inspection system of hot rolled strips | |
| CN1320335C (en) | Dynamic detection method based on image recognition for emery wheel appearance | |
| Kähler et al. | Anomaly detection for industrial surface inspection: Application in maintenance of aircraft components | |
| Hashmi et al. | Computer-vision based visual inspection and crack detection of railroad tracks | |
| CN113333329A (en) | Cigarette defect detection system based on deep learning | |
| Yang et al. | Design of drug and wine bottlecap defect detection system based on machine vision | |
| KR20210122429A (en) | Method and System for Artificial Intelligence based Quality Inspection in Manufacturing Process using Machine Vision Deep Learning | |
| Zhu et al. | Machine vision based on pipe joint surface defect detection and identification | |
| Appio et al. | Automatic surface inspection in steel products ensures safe, cost-efficient and timely defect detection in production | |
| Bharat et al. | Surface finish evaluation using curvelet transforms based machine vision system | |
| CN110674754A (en) | Online intermittent hollow filter stick visual defect detection and identification system | |
| CN111340802A (en) | Intelligent cloth inspecting machine capable of automatically inspecting cloth by adopting artificial intelligence technology |
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 |