CN116297493A - Method for improving steel plate defect detection accuracy and camera fault tolerance rate - Google Patents
Method for improving steel plate defect detection accuracy and camera fault tolerance rate Download PDFInfo
- Publication number
- CN116297493A CN116297493A CN202310309587.3A CN202310309587A CN116297493A CN 116297493 A CN116297493 A CN 116297493A CN 202310309587 A CN202310309587 A CN 202310309587A CN 116297493 A CN116297493 A CN 116297493A
- Authority
- CN
- China
- Prior art keywords
- defect
- defects
- camera
- detection
- steel plate
- 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
- G01N21/8851—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)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a method for improving the defect detection accuracy of a steel plate and the fault tolerance of cameras, which comprises the steps that a group of cameras are respectively arranged above and below the steel plate, and the capturing ranges of the cameras are overlapped; firstly, performing defect hot state detection, and then performing defect cold state detection: collecting two-dimensional image characteristics and color information of defects through a color camera, recording positions of defect targets on a plate, recording similarity of the defects, judging whether periodicity exists, combining an upstream thermal state inspection mother board number with a downstream table inspection system, and adding information interaction and reminding functions; since the defects and the pseudo defects have different color and shape characteristics, the defect types can be judged by the collected defect forms, periodicity, colors and positions. The invention has the advantages of improving the accuracy of defect discrimination by maximum efficiency, improving the fault tolerance of camera faults, reducing the labor intensity of workers, reducing potential safety hazards and ensuring accurate data.
Description
Technical Field
The invention belongs to the technical field of steel plate surface defect detection, and particularly relates to a method for improving steel plate defect detection accuracy and camera fault tolerance.
Background
The detection of the surface defects of the medium steel plate is generally carried out manually by domestic iron and steel enterprises, and the detection mode often needs to consume more manpower resources and has the problem that the simultaneous double-sided detection cannot be carried out. In order to improve the work and the safety environment of workers, lighten the labor intensity of the workers, improve the stability of a steel plate surface detection system, ensure the accuracy and promote the improvement of the intelligent level of steel plate production and manufacture, the design of the steel plate surface detection system with better functions and higher efficiency is necessary.
In order to solve the problems, a steel plate surface detection system is developed to carry out double-sided artificial intelligent detection, but in actual production, defects or non-defects are judged to have a plurality of similar points only from two-dimensional black-and-white images, the outline of non-defects such as water marks, iron scales and the like is similar to that of inclusion defects, and all the iron scales, heavy scales and bubble-shaped cracks have serrated edges. Similar forms are prone to confusion during classification, causing false positives, non-defects that are misidentified as defects are also referred to as "false defects", and excessive false positives have been a problem with two-dimensional image-based surface detection methods in use.
Disclosure of Invention
The invention aims to solve the problems of inaccurate detection data and high error recognition rate of the existing steel plate defects, and provides a method for improving the detection accuracy of the steel plate defects and the fault tolerance rate of camera faults, which can improve the accuracy of defect discrimination at maximum efficiency, improve the fault tolerance rate of camera faults, reduce the labor intensity of workers, reduce potential safety hazards and ensure that the data are accurate.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
a method for improving the detection accuracy of the defects of a steel plate and the fault tolerance of a camera comprises the following specific steps:
(1) Device arrangement: a group of cameras are respectively arranged above and below the steel plate, the lenses of the cameras are arranged opposite to the upper surface and the lower surface of the steel plate, each three cameras are in a group, and the capturing range of the left camera and the right camera is L 1 And L 2 And L is 1 And L 2 Corresponding steel plate areas are connected, and the capturing range of the middle camera is L 1 +L 2 The corresponding steel plate area is just the sum of the steel plate areas captured by the left camera and the right camera;
(2) Defect thermal state detection: the detection equipment acquires the defect information of the steel plate and informs a downstream table detection system of the information;
(3) Defect cold state detection:
a. collecting two-dimensional image characteristics and color information of the defects through a color camera;
b. recording the position of the defect target on the board;
c. recording the similarity of defects, and judging whether periodicity exists or not;
d. the upstream thermal state inspection mother board number is combined with the downstream meter inspection system, and the information interaction and reminding functions are added, so that the defect classification accuracy is improved;
(4) And a defect judging module: since the defects and the pseudo defects have different color and shape characteristics, the defect types can be judged by the collected defect forms, periodicity, colors and positions.
Further, in the step (2), the defect thermal state detection adopts a black-and-white camera, and the system is checked and recorded in advance.
Further, in the step (3), the defect cold state detection adopts a color camera, and a white light source is used to optimize the light path.
Further, in the step (3), for periodic defects such as indentations, the similarity of the shapes of the indentations at different positions is analyzed by combining the length information recorded by the counter, that is, the defect map obtained by target detection is expressed as a vector by using the similarity of cosin, and the similarity of different defect maps is represented by calculating the cosine distance between the vectors;
further, in the step (3), for defects with position distribution characteristics, such as cracks, different detection thresholds are set on different camera corresponding algorithms to solve the defects, that is, the detection threshold of the longitudinal crack algorithm of the middle camera is higher than that of the two-side cameras, and the detection threshold of the bubble-shaped crack algorithm of the two-side cameras is higher than that of the middle camera.
Further, in the step (4), the defect detection adopts a voting mechanism, and only when the defects detected by two adjacent cameras are regarded as defects, the false detection can be greatly reduced.
Further, in the step (4), the defect mainly includes an indentation, a crack, a scratch, a pressed-in iron oxide scale and an inclusion, the two-dimensional image of the indentation is characterized by an irregular block shape, the color of the indentation is gray, the two-dimensional image of the crack is characterized by a chapped shape and a meandering strip shape, the color of the crack is black, the two-dimensional image of the scratch is characterized by a straight line shape and a parallel line shape, the color of the crack is a metal bright color, the two-dimensional image of the pressed-in iron oxide scale is characterized by a block shape and an edge meandering shape, the color of the pressed-in iron oxide scale is black, and the two-dimensional image of the inclusion is characterized by an irregular block shape, the color of the inclusion is white or yellow and red dot shape.
According to the technical scheme, the automatic and intelligent functions of the steel plate surface defect detection system are realized through a machine vision technology, the accuracy of the steel plate surface defect detection system is improved through increasing defect color information, improving the overlapping area shot by a camera and judging the periodicity of the defect position, the labor intensity of workers and potential safety hazards are reduced, the method is suitable for the field environment of complex working conditions, and particularly the fault tolerance rate of the camera can be improved.
Drawings
FIG. 1 is a diagram of a camera arrangement of the present invention;
FIG. 2 is a flowchart of a method for improving the accuracy of detecting defects of a steel plate and the fault tolerance of a camera according to the present invention.
Description of the embodiments
Examples
In order that the present invention may become more apparent, a method for improving the accuracy of detecting defects in steel plates and the fault tolerance of cameras according to the present invention will be further described with reference to the accompanying drawings.
A method for improving the detection accuracy of the defects of a steel plate and the fault tolerance of a camera comprises the following specific steps:
(1) Device arrangement: as shown in fig. 1, a group of cameras are respectively arranged above and below the steel plate 1, the lenses of the cameras are arranged opposite to the upper and lower surfaces of the steel plate, and the camera a 1 、B 1 And C 1 Camera A as a group 2 、B 2 And C 2 A group of;
for detection of the upper surface of the steel plate, left and right cameras A 1 And C 1 The capture ranges of (a) are respectively L 1 And L 2 And L is 1 And L 2 Corresponding steel plate areas are connected, and an intermediate camera B 1 Is L 1 +L 2 The corresponding steel plate area is just the sum of the steel plate areas captured by the left camera and the right camera, and the working principle of the detection camera on the lower surface of the steel plate is the same as that of the detection camera on the upper surface of the steel plate;
(2) Defect thermal state detection: the detection equipment acquires the defect information of the steel plate, informs a downstream table detection system of the defect information, adopts a black-and-white camera for defect thermal state detection, and performs pre-inspection and entry into the system;
(3) Defect cold state detection:
a. a color camera is adopted, a white light source is used for optimizing a light path, and two-dimensional image characteristics and color information of defects are collected through the color camera;
b. recording the position of the defect target on the board;
c. recording the similarity of defects, and judging whether periodicity exists or not;
d. the upstream thermal state inspection mother board number is combined with the downstream meter inspection system, and the information interaction and reminding functions are added, so that the defect classification accuracy is improved;
(4) And a defect judging module: since the defects and the pseudo defects have different color and shape characteristics, the defect types can be judged by the collected defect forms, periodicity, colors and positions.
In this embodiment, for the periodical defect of the indentation, the length information recorded by the counter is combined, the shape similarity of the indentation at different positions is analyzed, that is, the defect map obtained by the target detection is expressed as a vector by using the cosine similarity, and the similarity of different defect maps is represented by calculating the cosine distance between the vectors.
In this embodiment, for the defect with the position distribution feature, such as the crack, different detection thresholds are set on different camera corresponding algorithms, for example, the detection threshold of the longitudinal crack algorithm of the middle camera is higher than that of the two-side cameras, and the detection threshold of the bubble-shaped crack algorithm of the two-side cameras is higher than that of the middle camera.
In this embodiment, the defect detection adopts a voting mechanism, and only when two adjacent cameras detect defects, the defects are considered to be defects, so that false detection can be greatly reduced.
The working principle of the invention is as follows: (1) Color information exists on the surface defect of the steel plate, and the color information is shown in the following table; (2) The defect detection accuracy of the overlapping area observed by the camera is higher than that of the non-overlapping area observed by the camera, so that a mode of overlapping shooting by the camera is adopted; (3) And the detection information of the hot state meter detection and the cold state meter detection is comprehensively utilized to carry out information interaction, so that the defect detection accuracy is further ensured.
As can be seen from the above table, the defects mainly include an indentation, a crack, a scratch, a pressed scale and inclusions, the two-dimensional image of the indentation is characterized by an irregular block shape, the color of the indentation is gray, the two-dimensional image of the crack is characterized by a chapped shape and a meandering strip shape, the color of the crack is black, the two-dimensional image of the scratch is characterized by a straight line shape and a parallel line shape, the color of the crack is a metallic bright color, the two-dimensional image of the pressed scale is characterized by a block shape and edge meandering, the color of the pressed scale is black, and the two-dimensional image of the inclusions is characterized by an irregular block shape, the color of the inclusions is white or yellow and red dot shape. It can be seen that the defects are distinguished from the two-dimensional image features and colors corresponding to the pseudo defects, and the types of the defects can be accurately judged by comparing the defect shapes and colors of the defects.
The invention interacts the defect data with the downstream through the defect discovered by the upstream thermal state inspection, thus greatly improving the accuracy of defect inspection and facilitating the daily operation of workers. When the same linear camera builds images, the same defect is located at the top or bottom edge of the images, so that the defect is distributed on two adjacent images, and when a defect data set is expanded, a complete defect image truncating and truncating method can be adopted. The invention can realize the automatic and intelligent functions of the steel plate surface detection system through the machine vision technology on the basis of double-sided artificial intelligent detection, ensure the accuracy of defect discrimination, reduce the labor intensity of workers and the potential safety hazard, optimize the camera shooting scheme, improve the accuracy of defect discrimination with maximum efficiency and improve the fault tolerance of camera faults.
In addition to the embodiments described above, other embodiments of the invention are possible. All technical schemes formed by equivalent substitution or equivalent transformation fall within the protection scope of the invention.
Claims (7)
1. A method for improving the detection accuracy of the defects of a steel plate and the fault tolerance of a camera is characterized by comprising the following specific steps:
(1) Device arrangement: a group of cameras are respectively arranged above and below the steel plate, the lenses of the cameras are arranged opposite to the upper surface and the lower surface of the steel plate, each three cameras are in a group, and the capturing range of the left camera and the right camera is L 1 And L 2 And L is 1 And L 2 Corresponding steel plate areas are connected, and the capturing range of the middle camera is L 1 +L 2 The corresponding steel plate area is just the sum of the steel plate areas captured by the left camera and the right camera;
(2) Defect thermal state detection: the detection equipment acquires the defect information of the steel plate and informs a downstream table detection system of the information;
(3) Defect cold state detection:
a. collecting two-dimensional image characteristics and color information of the defects through a color camera;
b. recording the position of the defect target on the board;
c. recording the similarity of defects, and judging whether periodicity exists or not;
d. combining an upstream thermal state inspection motherboard number with a downstream meter inspection system, and adding information interaction and reminding functions;
(4) And a defect judging module: since the defects and the pseudo defects have different color and shape characteristics, the defect types can be judged by the collected defect forms, periodicity, colors and positions.
2. The method for improving the accuracy of detecting defects of a steel plate and the fault tolerance of a camera according to claim 1, wherein the method comprises the following steps:
in the step (2), a black-and-white camera is adopted for defect thermal state detection, and the defect thermal state detection is checked in advance and recorded into a system.
3. The method for improving the defect detection accuracy and the fault tolerance of a camera according to claim 1 or 2, wherein:
in the step (3), a color camera is used for defect cold state detection, and a white light source is used to optimize the light path.
4. The method for improving the accuracy of detecting defects in steel plates and the fault tolerance of cameras according to claim 1 or 2,
in the step (3), for periodic defects such as indentations, the similarity of the shapes of the indentations at different positions is analyzed by combining the length information recorded by the counter, namely, the defect map obtained by target detection is expressed as a vector by utilizing the similarity of cosine, and the similarity of different defect maps is represented by calculating the cosine distance between the vectors.
5. The method for improving the defect detection accuracy and the fault tolerance of a camera according to claim 1 or 2, wherein:
in the step (3), for defects with position distribution characteristics, such as cracks, different detection thresholds are set on different camera corresponding algorithms to solve the defects, namely, the detection threshold of a longitudinal crack algorithm of a middle camera is higher than that of two side cameras, and the detection threshold of a bubble-shaped crack algorithm of the two side cameras is higher than that of the middle camera.
6. The method for improving the defect detection accuracy and the fault tolerance of a camera according to claim 1 or 2, wherein:
in the step (4), the defect detection adopts a voting mechanism, and only the defects detected by two adjacent cameras are considered as defects, so that false detection can be greatly reduced.
7. The method for improving the defect detection accuracy and the fault tolerance of a camera according to claim 1 or 2, wherein:
in the step (4), the defects mainly comprise indentations, cracks, scratches, pressed iron oxide sheets and inclusions, wherein the two-dimensional image of the indentations is characterized by irregular blocks, the colors of the two-dimensional image of the indentations are gray, the two-dimensional image of the cracks is characterized by chapped and bent strips, the colors of the two-dimensional image of the cracks are black, the two-dimensional image of the scratches is characterized by straight lines and parallel lines, the colors of the two-dimensional image of the scratches are metallic bright colors, the two-dimensional image of the pressed iron oxide sheets is characterized by blocks and edge bends, the colors of the two-dimensional image of the pressed iron oxide sheets are black, and the colors of the two-dimensional image of the inclusions are irregular blocks, and the colors of the two-dimensional image of the inclusions are white or yellow and red dots.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310309587.3A CN116297493A (en) | 2023-03-28 | 2023-03-28 | Method for improving steel plate defect detection accuracy and camera fault tolerance rate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310309587.3A CN116297493A (en) | 2023-03-28 | 2023-03-28 | Method for improving steel plate defect detection accuracy and camera fault tolerance rate |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116297493A true CN116297493A (en) | 2023-06-23 |
Family
ID=86816666
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310309587.3A Pending CN116297493A (en) | 2023-03-28 | 2023-03-28 | Method for improving steel plate defect detection accuracy and camera fault tolerance rate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116297493A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117007611A (en) * | 2023-09-28 | 2023-11-07 | 杭州百子尖科技股份有限公司 | Method, device, equipment and medium for detecting periodic defects of sheet material |
-
2023
- 2023-03-28 CN CN202310309587.3A patent/CN116297493A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117007611A (en) * | 2023-09-28 | 2023-11-07 | 杭州百子尖科技股份有限公司 | Method, device, equipment and medium for detecting periodic defects of sheet material |
| CN117007611B (en) * | 2023-09-28 | 2024-01-09 | 杭州百子尖科技股份有限公司 | Method, device, equipment and medium for detecting periodic defects of sheet material |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115170576B (en) | Aluminum pipe surface defect detection method based on machine vision | |
| CN107705301B (en) | Highway marking damage detection method based on unmanned aerial vehicle aerial highway image | |
| CN109636815A (en) | A kind of metal plate and belt Product labelling information identifying method based on computer vision | |
| CN111814686A (en) | Vision-based power transmission line identification and foreign matter invasion online detection method | |
| CN102426649A (en) | Simple steel seal digital automatic identification method with high accuracy rate | |
| CN116297493A (en) | Method for improving steel plate defect detection accuracy and camera fault tolerance rate | |
| CN109816645B (en) | Automatic detection method for steel coil loosening | |
| CN114677362B (en) | Surface defect detection method based on improved YOLOv5 | |
| CN112326671A (en) | Metal plate surface defect detection method based on machine vision | |
| CN110390306A (en) | Detection method, vehicle and the computer readable storage medium of right angle parking stall | |
| Wei et al. | Beamlet transform based pavement image crack detection | |
| CN109389165A (en) | Oil level gauge for transformer recognition methods based on crusing robot | |
| CN104483243A (en) | Adhered-rice detection and segmentation method, device and system | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN112561875A (en) | Photovoltaic cell panel coarse grid detection method based on artificial intelligence | |
| CN110321855A (en) | A kind of greasy weather detection prior-warning device | |
| CN110084807B (en) | Detection method for false detection of welding seam flaw detection negative image | |
| CN116485764A (en) | Structural surface defect identification method, system, terminal and medium | |
| CN111783672A (en) | Image feature identification method for improving bridge dynamic displacement precision | |
| Lu et al. | Traffic light recognition | |
| Zhang et al. | A license plate recognition system based on tamura texture in complex conditions | |
| CN113077465A (en) | Pavement crack detection and extraction device and detection method thereof | |
| CN110610474A (en) | Solar panel defect real-time detection method based on infrared image | |
| CN103473566A (en) | Multi-scale-model-based vehicle detection method | |
| CN111507177A (en) | Identification method and device for metering turnover cabinet |
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 |