A multispectral fusion high-speed detection device and method for surface defects of a steel strip

Abstract

The application provides a kind of steel band surface defect multispectral fusion high-speed detection device and detection method, belong to steel band surface defect detection technical field, including foundation and main support, the main support is door type structure, across the both sides of steel band production equipment, bottom is connected with foundation using embedded bolt.The application is irradiated to steel band surface by LED array light source respectively using white, ultraviolet light and red light source, and the image of steel band surface after being irradiated by different light sources is collected using CCD camera respectively, so that when image recognition is carried out, the image of white light source is identified according to the brightness of reflected light, the defect of different depth is identified, the image of ultraviolet light source is identified by shortening wavelength to break through diffraction limit, the resolution and contrast of surface defect detection are improved;The image of red light source enhances the color development effect of small scratch on metal surface;Through the image recognition of different light colors, the detection and recognition accuracy can be effectively improved, and the influence of sundries and reflection on detection result is reduced.

Description

Technical Field

[0001] This invention relates to a multispectral fusion high-speed detection device and method, specifically a multispectral fusion high-speed detection device and method for steel strip surface defects, belonging to the field of steel strip surface defect detection technology. Background Technology

[0002] Surface defect detection of steel strip is a crucial step in the steel strip production process. It primarily utilizes optical, electromagnetic, or acoustic technologies to identify and assess defects such as cracks, scratches, inclusions, and oxide scale on the steel strip surface. The accuracy and efficiency of this detection directly impact product quality and production efficiency. Traditional methods for steel strip surface defect detection largely rely on manual visual inspection or machine vision technology. Manual visual inspection is heavily influenced by subjective factors such as the inspector's experience and fatigue, resulting in low efficiency and high rates of missed or false detections. Traditional machine vision technology is highly sensitive to environmental conditions; water stains, oil, and other impurities on the steel strip surface are easily mistaken for defects, leading to a high false alarm rate. Furthermore, steel strip is a highly reflective material, and high light reflection can interfere with defect identification.

[0003] A Chinese patent entitled "Surface Inspection System under a Surface Inspection Instrument" (publication number CN215263087U) discloses a surface inspection technology that uses an axial flow fan to blow away foreign objects such as cold water, iron oxide scale, oil stains, and dust that fall on the steel strip, preventing them from falling onto the optical glass and interfering with the inspection results. However, this inspection system cannot directly clean the surface of the steel strip, resulting in residual water stains, oil stains, oil mist, and other debris on the surface, which can also interfere with the inspection results. Furthermore, the lack of a light source setting leads to low inspection accuracy.

[0004] Chinese patent titled "Nickel Alloy Steel Strip Surface Defect Detection Device" (publication number CN117848958B) discloses a steel strip surface defect detection technology. It uses a first driving component to control the reciprocating movement of the insertion shaft in the horizontal direction to achieve rapid replacement of the cotton sleeve, thereby improving the efficiency of detecting surface defects on the strip. However, although the device has the advantage of good cleaning effect, its operating cost is too high, and it requires continuous drying of the cotton sleeve. Moreover, it does not include a light source.

[0005] To address this, a high-speed multispectral fusion detection device and method for detecting surface defects in steel strips are proposed. Summary of the Invention

[0006] In view of this, the present invention provides a multispectral fusion high-speed detection device and method for steel strip surface defects, so as to solve or alleviate the technical problems existing in the prior art, or at least provide a beneficial option.

[0007] The technical solution of this invention is implemented as follows: A multispectral fusion high-speed detection device for surface defects of steel strip includes a foundation and a main support. The main support is a portal structure that spans both sides of the steel strip production equipment. The bottom is connected to the foundation by pre-embedded bolts. A light source support and a camera support are installed on one side of the main support. An LED array light source is detachably installed on one side of the light source support. A CCD camera is installed on the movable end of the camera support. An oil mist blowing device is installed at the bottom of one side of the main support. The oil mist blowing device includes a main blowing assembly, which consists of an oil mist blowing support and a main nozzle. The main nozzle is connected to one end of the oil mist blowing support.

[0008] The LED array light source includes a high-brightness white LED strip light source, an ultraviolet array LED light source, and / or a red LED strip light source. The light source beads are arranged in a single row or multiple rows in a straight line, covering the width of the steel strip.

[0009] The oil mist purging bracket and the main nozzle are used to blow compressed air onto the surface of the steel strip, and the output port of the main nozzle is a slit, which is inclined downward at 30° to the surface of the steel strip and faces the detection surface of the steel strip.

[0010] More preferably, the bottom flange of the CCD camera is connected to the moving end of the camera bracket, and the flange and the housing of the CCD camera are integrally die-cast from aluminum alloy. The CCD camera is a high-precision 2K-8K camera with 4096-8192 pixels per line and a scanning frequency of 10000-50000 lines / second. The CCD cameras can be arranged in single or multiple units. A single CCD camera is centered along the width of the steel strip and is suitable for surface inspection of steel strips with a width of less than 400mm. Multiple CCD cameras are arranged side by side along the width of the steel strip and are suitable for surface inspection of steel strips with a width of more than 400mm. The overlap detection area of ​​multiple CCD cameras is 5-50mm.

[0011] More preferably, the camera bracket is a composite bracket made of aluminum alloy profiles, including a main vertical lifting rod and a horizontal adjusting arm;

[0012] The horizontal adjustment arm is installed at the output end of the main body's vertical lifting rod, and an angle adjustment gimbal is hinged to the end of the horizontal adjustment arm. The CCD camera flange is connected to one side of the angle adjustment gimbal. The vertical lifting rod is used to adjust the height of the CCD camera from 0.5 to 2m, the horizontal adjustment arm is used to adjust the horizontal extension of the CCD camera from 0 to 1m, and the angle adjustment gimbal is used to achieve 360° horizontal rotation and ±30° vertical tilt of the CCD camera.

[0013] In a further preferred embodiment, a dovetail groove is provided between the light source bracket and the LED array light source, and the back of the LED array light source is an aluminum alloy shell with a dovetail track, which is slidably connected to the inner sidewall of the dovetail groove and fixedly connected to the light source bracket with bolts.

[0014] More preferably, the wavelength range of the ultraviolet array LED light source is 200-400nm, and the wavelength range of the red LED strip light source is 620-750nm.

[0015] More preferably, the oil mist purging device further includes auxiliary purging components, which consist of two components, each installed 100mm directly above the CCD camera with its nozzle aligned with the lens; and 100mm directly above the LED array light source detection end with its nozzle aligned with the light source detection surface.

[0016] More preferably, the slit length of the main nozzle is 30-60mm, the inlet pressure is 0.2-0.3Mpa, and the air source is dry compressed air with a dew point below -20℃.

[0017] Further preferably, it also includes a computer system, which is used to observe the images acquired by the CCD camera in real time, and has a built-in image processing and recognition system to identify surface defect information of the steel strip based on the images acquired by the CCD camera, and to statistically locate the defects in combination with the steel strip information.

[0018] A high-speed multispectral fusion detection method for surface defects in steel strips includes the following steps:

[0019] S1. Obtain basic information about the steel strip and set the scanning frequency and width of the CCD camera according to the speed and width of the steel strip;

[0020] S2. Start the oil mist blowing device to clean the surface of the steel strip of debris and oil mist;

[0021] S3. Start the LED array light source, adjust the angle and brightness, and calibrate the parameters of the CCD camera;

[0022] The S4 and CCD cameras acquire images of the steel strip surface at a set frequency and feed them back to the computer system for real-time observation.

[0023] S5. The computer system classifies and stores images according to the basic information of the steel strip, combined with time and light color.

[0024] S6. The image is preprocessed by the image processing and recognition system, and the preprocessed image is compared with the feature map. The defect image is stored and the defect location is marked based on the time information.

[0025] S7. If there are no similar defect features, proceed to S8; if there are similar features, proceed to S9.

[0026] S8. Continue to acquire the next frame image, repeat S4-S7, and execute S10 when the detection is complete.

[0027] S9. Automatically mark the defect location, identify the suspected type, collect defect data, and provide personnel prompts. Then repeat S4-S7. Once the inspection is complete, execute S10.

[0028] S10. Generate an inspection report (including full information on defects / steel strips) and archive and store the report in conjunction with the push direction.

[0029] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions:

[0030] I. This invention uses an oil mist blowing device to blow compressed air onto the surface of the steel strip, thereby cleaning the surface of the steel strip by blowing, avoiding interference from debris and oil mist with the test results. Compared with the traditional cleaning method that relies on cotton covers, the cost of using compressed air blowing is lower, and it saves the step of drying the cotton covers, effectively improving the continuous operation capability of the testing device.

[0031] II. This invention uses an LED array light source to illuminate the surface of a steel strip with white, ultraviolet, and red light sources respectively, and uses a CCD camera to capture images of the steel strip surface after illumination by different light sources. During image recognition, the white light source image identifies defects of different depths based on the brightness of reflected light; the ultraviolet light source image, by shortening the wavelength to overcome the diffraction limit, improves the resolution and contrast of surface defect detection, and the surface cleanliness of the steel strip can be identified based on the fluorescence reaction of the steel strip surface; the red light source image enhances the color rendering effect of minor scratches on the metal surface. By fusing images of different light colors, the detection and recognition accuracy can be effectively improved, and the influence of impurities and reflections on the detection results can be reduced.

[0032] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1This is a structural diagram of Embodiment 1 of the present invention;

[0035] Figure 2 This is a schematic flowchart of the detection method of the present invention;

[0036] Figure 3 This is a structural diagram of Embodiment 2 of the present invention.

[0037] Reference numerals in the attached diagram: 1. Foundation; 2. Main support; 3. Light source support; 4. Dovetail groove; 5. LED array light source; 6. Camera support; 7. CCD camera; 8. Oil mist purging support; 9. Main nozzle; 10. Strip. Detailed Implementation

[0038] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0039] It is important to note that terms such as "first," "second," "symmetric," and "array" are used only to distinguish between descriptive and positional descriptions and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, features specified with terms such as "first" or "symmetric" may explicitly or implicitly include one or more of that feature; similarly, when the quantity of certain features is not limited by words such as "two" or "three," it should be noted that such features also explicitly or implicitly include one or more features.

[0040] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0041] Example 1

[0042] like Figures 1-2 As shown, this embodiment of the invention provides a high-speed multispectral fusion detection device for steel strip surface defects to meet the single-sided inspection requirements. The device includes a foundation 1 and a main support 2. The main support 2 is a portal frame structure that spans both sides of the steel strip production equipment. The bottom is connected to the foundation 1 by pre-embedded bolts. A light source support 3 and a camera support 6 are installed on one side of the main support 2. An LED array light source 5 is detachably installed on one side of the light source support 3. A CCD camera 7 is installed on the movable end of the camera support 6. An oil mist blowing device is installed at the bottom of one side of the main support 2. The oil mist blowing device includes a main blowing assembly, which consists of an oil mist blowing support 8 and a main nozzle 9. The main nozzle 9 is connected to one end of the oil mist blowing support 8.

[0043] Among them, the LED array light source 5 has a high-brightness white LED strip light source, an ultraviolet array LED light source and a red LED strip light source. The light source beads are arranged in multiple rows in a straight line, covering the width of the steel strip.

[0044] Among them, the oil mist purging bracket 8 and the main nozzle 9 are used to blow compressed air onto the surface of the steel strip, and the output port of the main nozzle 9 is a slit, which is inclined downward at 30° to the surface of the steel strip and faces the detection surface of the steel strip.

[0045] In one embodiment, the bottom flange of the CCD camera 7 is connected to the moving end of the camera bracket 6. The flange and the housing of the CCD camera 7 are integrally die-cast from aluminum alloy. The CCD camera 7 is arranged in a single unit, with each CCD camera 7 centered along the width of the steel strip. It is suitable for surface inspection of steel strips with a width of less than 400mm.

[0046] The CCD camera 7 can be equipped with a linear CCD camera, with 4096 pixels per line and a scanning frequency of 50,000 lines per second, enabling real-time dynamic capture of the surface of a high-speed moving steel strip. Its lens uses an industrial-grade high-resolution fixed-focus lens, and the focal length can be adjusted according to the width of the steel strip being inspected and the installation distance to ensure clear and distortion-free imaging.

[0047] The camera bracket 6 can be installed independently. The strip 10 spans both sides of the production equipment and is directly connected to the concrete foundation 1 through pre-embedded bolts. Whether the camera bracket 6 is installed on one side of the main bracket 2 or installed independently, it must be set independently relative to the slitting equipment to prevent image shaking caused by the vibration of the production equipment.

[0048] In one embodiment, the camera bracket 6 is an aluminum alloy profile combined bracket, including a main vertical lifting rod and a horizontal adjusting arm;

[0049] The horizontal adjustment arm is installed at the output end of the main vertical lifting rod, and the end of the horizontal adjustment arm is hinged to an angle adjustment gimbal. The CCD camera 7 is flanged and connected to one side of the angle adjustment gimbal. The vertical lifting rod is used to adjust the height of the CCD camera 7 from 0.5 to 2m, the horizontal adjustment arm is used to adjust the horizontal extension of the CCD camera 7 from 0 to 1m, and the angle adjustment gimbal is used to achieve 360° horizontal rotation and ±30° vertical tilt of the CCD camera 7.

[0050] The vertical lifting rod has a built-in screw slide adjustment structure, which is used to adjust the overall height of the horizontal adjustment arm. The horizontal adjustment arm has multiple telescopic slots, which are used to adjust the horizontal extension of the angle adjustment gimbal. The angle adjustment gimbal is used to adjust the CCD camera 7 at multiple angles.

[0051] In one embodiment, a dovetail groove 4 is provided between the light source bracket 3 and the LED array light source 5. The back of the LED array light source 5 is an aluminum alloy shell with a dovetail track, which is slidably connected to the inner sidewall of the dovetail groove 4 and fixedly connected to the light source bracket 3 with bolts.

[0052] By using the dovetail groove 4 in conjunction with the dovetail track to connect the light source bracket 3 and the LED array light source 5, the entire LED array light source 5 can be pulled out and replaced during actual production.

[0053] The light source bracket 3 can also be installed independently, spanning across the strip 10 and both sides of the production equipment, and is directly connected to the foundation 1 by pre-embedded bolts.

[0054] In one embodiment, the wavelength range of the ultraviolet array LED light source is 200-400nm, preferably 250-350nm, and the wavelength range of the red LED strip light source is 620-750nm, preferably 630-680nm.

[0055] By utilizing ultraviolet light sources and shortening the wavelength to overcome the diffraction limit, the resolution and contrast of surface defect detection are improved. The cleanliness of the steel strip surface can be identified based on the fluorescence reaction of organic contaminants such as residual oil, coolant, and lubricant. Red light sources enhance the color rendering of minor scratches on the metal surface. In addition to white, ultraviolet, and red light sources, a blue LED strip light source with a wavelength range of 450-495nm can be added. This light source effectively highlights the oxidation color differences on the steel strip surface, helping inspectors to more clearly distinguish color changes caused by different degrees of oxidation. Simultaneously, through the synergistic effect of multispectral light sources, the system can fuse images from different wavelengths and combine them with image algorithms to comprehensively analyze various defect features. This enables accurate identification and classification of various defects on the steel strip surface, such as cracks, pits, scratches, and contamination, significantly improving the comprehensiveness and accuracy of the detection. In actual inspection, multiple light source combinations can be flexibly selected based on the characteristics of different steel strip materials, production processes, and defect types to adapt to diverse inspection needs.

[0056] In one embodiment, the oil mist purging device further includes two auxiliary purging components, which are respectively installed 100mm directly above the CCD camera 7 with the nozzles aligned with the lens; and 100mm directly above the light source detection end of the LED array light source 5 with the nozzles aligned with the light source detection surface. The slit length of the main nozzle 9 is 30-60mm, the air inlet pressure is 0.2-0.3Mpa, and the air source is dry compressed air with a dew point below -20℃.

[0057] The main nozzle 9 is connected to the diversion valve via a high-pressure PU air pipe. The diversion valve is used to supply compressed air to the main nozzle 9 and the nozzle of the auxiliary purging assembly respectively. The main nozzle 9 and the auxiliary purging assembly are both mounted on the steel belt conveyor frame, camera bracket 6 and light source bracket via fixed seats / brackets.

[0058] In one embodiment, a computer system is also included. The computer system is used to observe the images acquired by the CCD camera 7 in real time and has a built-in image processing and recognition system. Based on the images acquired by the CCD camera 7, the system identifies surface defect information of the steel strip and performs statistical location of the defects in combination with the steel strip information.

[0059] The computer system supports real-time image observation, acquisition frequency setting, and storage functions. The image processing and recognition system uses artificial intelligence to process images, identify surface defect information of the steel strip, and statistically locate the defects based on the acquired steel strip information for subsequent processing.

[0060] In the initial stage of use, various defect images are manually identified and marked to form standard defect images. In subsequent use, the system compares the captured photos with the standard defect images, marks images with similar features, and prompts the operator.

[0061] A high-speed multispectral fusion detection method for surface defects in steel strips includes the following steps:

[0062] S1. Obtain basic information about the steel strip and set the scanning frequency and width of the CCD camera 7 according to the speed and width of the steel strip;

[0063] Basic information includes the steel strip coil number, width, thickness, weight, speed, etc.

[0064] S2. Start the oil mist blowing device to clean the surface of the steel strip of debris and oil mist. The main and auxiliary blowing components work simultaneously.

[0065] S3. Start the LED array light source 5, adjust its angle and brightness, and calibrate the parameters of the CCD camera 7;

[0066] S4 and CCD camera 7 acquire images of the steel strip surface at a set frequency and feed them back to the computer system for real-time observation;

[0067] S5. The computer system classifies and stores images according to the basic information of the steel strip (roll number) combined with time and light color;

[0068] S6. The image is preprocessed (noise reduction, contrast enhancement, and correction) through an image processing and recognition system, and the preprocessed image is compared with a feature map. The defect image is stored, and the defect location is marked based on time information.

[0069] S7. If there are no similar defect features, proceed to S8; if there are similar features, proceed to S9.

[0070] S8. Continue to acquire the next frame image, repeat S4-S7, and execute S10 when the detection is complete.

[0071] S9. Automatically mark the defect location, identify the suspected type, collect defect data, and provide personnel prompts. Then repeat S4-S7. Once the inspection is complete, execute S10.

[0072] S10. Generate an inspection report containing full information on defects / steel strips, and archive and store the report in conjunction with the push direction.

[0073] Example 2

[0074] like Figures 1-3 As shown, this embodiment of the invention provides a high-speed multispectral fusion detection device for surface defects of steel strips, addressing the need for double-sided inspection. Figure 3 As shown in the diagram, the front bright-field detection station, the front dark-field detection station, the back bright-field detection station, and the back dark-field detection station all include a foundation 1 and a main support 2. The main support 2 is a portal frame structure, spanning both sides of the steel strip production equipment. Its bottom is connected to the foundation 1 using pre-embedded bolts. A light source support 3 and a camera support 6 are installed on one side of the main support 2. An LED array light source 5 is detachably installed on one side of the light source support 3. Figure 3 The LED light source in the camera bracket 6 is equipped with a CCD camera 7 at the movable end of the camera bracket 6. An oil mist blowing device is installed at the bottom of one side of the main bracket 2. The oil mist blowing device includes a main blowing assembly, which consists of an oil mist blowing bracket 8 and a main nozzle 9. The main nozzle 9 is connected to one end of the oil mist blowing bracket 8.

[0075] Among them, the LED array light source 5 has a high-brightness white LED strip light source, an ultraviolet array LED light source, or a red LED strip light source. The light source beads are arranged in a single row in a straight line, covering the width of the steel strip. Multiple LED array light sources 5 with different colors are set along the length of the steel strip.

[0076] Among them, the oil mist purging bracket 8 and the main nozzle 9 are used to blow compressed air onto the surface of the steel strip, and the output port of the main nozzle 9 is a slit, which is inclined downward at 30° to the surface of the steel strip and faces the detection surface of the steel strip.

[0077] In one embodiment, the bottom flange of the CCD camera 7 is connected to the moving end of the camera bracket 6. The flange and the housing of the CCD camera 7 are integrally die-cast from aluminum alloy. The CCD camera 7 is a high-precision 8K camera with 8192 pixels per line and a scanning frequency of 10000 lines / second. Multiple CCD cameras 7 are arranged side by side along the width of the steel strip, which is suitable for surface inspection of steel strips with a width of 400mm or more. The overlapping detection area of ​​multiple CCD cameras 7 is 5-50mm.

[0078] The camera bracket 6 can be installed independently. The strip 10 spans both sides of the production equipment and is directly connected to the concrete foundation 1 through pre-embedded bolts. Whether the camera bracket 6 is installed on one side of the main bracket 2 or installed independently, it must be set independently relative to the slitting equipment to prevent image shaking caused by the vibration of the production equipment.

[0079] In one embodiment, the camera bracket 6 is an aluminum alloy profile combined bracket, including a main vertical lifting rod and a horizontal adjusting arm;

[0080] The horizontal adjustment arm is installed at the output end of the main vertical lifting rod, and the end of the horizontal adjustment arm is hinged to an angle adjustment gimbal. The CCD camera 7 is flanged and connected to one side of the angle adjustment gimbal. The vertical lifting rod is used to adjust the height of the CCD camera 7 from 0.5 to 2m, the horizontal adjustment arm is used to adjust the horizontal extension of the CCD camera 7 from 0 to 1m, and the angle adjustment gimbal is used to achieve 360° horizontal rotation and ±30° vertical tilt of the CCD camera 7.

[0081] The vertical lifting rod has a built-in screw slide adjustment structure, which is used to adjust the overall height of the horizontal adjustment arm. The horizontal adjustment arm has multiple telescopic slots, which are used to adjust the horizontal extension of the angle adjustment gimbal. The angle adjustment gimbal is used to adjust the CCD camera 7 at multiple angles.

[0082] In one embodiment, a dovetail groove 4 is provided between the light source bracket 3 and the LED array light source 5. The back of the LED array light source 5 is an aluminum alloy shell with a dovetail track, which is slidably connected to the inner sidewall of the dovetail groove 4 and fixedly connected to the light source bracket 3 with bolts.

[0083] By using the dovetail groove 4 in conjunction with the dovetail track to connect the light source bracket 3 and the LED array light source 5, the entire LED array light source 5 can be pulled out and replaced during actual production.

[0084] The light source bracket 3 can also be installed independently, spanning across the strip 10 and both sides of the production equipment, and is directly connected to the foundation 1 by pre-embedded bolts.

[0085] In one embodiment, the wavelength range of the ultraviolet array LED light source is 200-400nm, and the wavelength range of the red LED strip light source is 620-750nm.

[0086] By using ultraviolet light sources to break through the diffraction limit by shortening the wavelength, the resolution and contrast of surface defect detection can be improved. The cleanliness of the steel strip surface can be identified based on the fluorescence reaction of organic contaminants such as residual oil, coolant, and lubricant on the steel strip surface. By using red light sources to enhance the color rendering effect of tiny scratches on the metal surface, the image can be improved.

[0087] In one embodiment, the oil mist purging device further includes two auxiliary purging components, which are respectively installed 100mm directly above the CCD camera 7 with the nozzles aligned with the lens; and 100mm directly above the light source detection end of the LED array light source 5 with the nozzles aligned with the light source detection surface. The slit length of the main nozzle 9 is 30-60mm, the air inlet pressure is 0.2-0.3Mpa, and the air source is dry compressed air with a dew point below -20℃.

[0088] The main nozzle 9 is connected to the diversion valve via a high-pressure PU air pipe. The diversion valve is used to supply compressed air to the main nozzle 9 and the nozzle of the auxiliary purging assembly respectively. The main nozzle 9 and the auxiliary purging assembly are both mounted on the steel belt conveyor frame, camera bracket 6 and light source bracket via fixed seats / brackets.

[0089] In one embodiment, a computer system is also included. The computer system is used to observe the images acquired by the CCD camera 7 in real time and has a built-in image processing and recognition system. Based on the images acquired by the CCD camera 7, the system identifies surface defect information of the steel strip and performs statistical location of the defects in combination with the steel strip information.

[0090] The computer system supports real-time image observation, acquisition frequency setting, and storage functions. The image processing and recognition system uses artificial intelligence to process images, identify surface defect information of the steel strip, and statistically locate the defects based on the acquired steel strip information for subsequent processing.

[0091] In the initial stage of use, various defect images are manually identified and marked to form standard defect images. In subsequent use, the system compares the captured photos with the standard defect images, marks images with similar features, and prompts the operator.

[0092] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A high-speed multispectral fusion detection device for surface defects of steel strip, comprising a foundation (1) and a main support (2), characterized in that, The main support (2) is a portal structure that spans both sides of the steel strip production equipment. The bottom is connected to the foundation (1) by pre-embedded bolts. A light source support (3) and a camera support (6) are installed on one side of the main support (2). An LED array light source (5) is detachably installed on one side of the light source support (3). A CCD camera (7) is installed on the movable end of the camera support (6). An oil mist blowing device is installed at the bottom of one side of the main support (2). The oil mist blowing device includes a main blowing assembly, which consists of an oil mist blowing bracket (8) and a main nozzle (9). The main nozzle (9) is connected to one end of the oil mist blowing bracket (8). The LED array light source (5) has a high-brightness white LED strip light source, an ultraviolet array LED light source and / or a red LED strip light source. The light source beads are arranged in a single row or multiple rows in a straight line, covering the width of the steel strip. The oil mist purging bracket (8) and the main nozzle (9) are used to blow compressed air onto the surface of the steel strip, and the output port of the main nozzle (9) is a slit, which is inclined downward at 30° to the surface of the steel strip and faces the steel strip detection surface.

2. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: The bottom flange of the CCD camera (7) is connected to the moving end of the camera bracket (6). The flange and the outer shell of the CCD camera (7) are integrally die-cast from aluminum alloy. The CCD camera (7) is a high-precision 2k-8k camera with 4096-8192 pixels per line and a scanning frequency of 10000-50000 lines / second. The CCD camera (7) can be arranged in single or multiple units. A single CCD camera (7) is arranged in the center along the width direction of the steel strip and is suitable for surface detection of steel strips with a width of less than 400mm. Multiple CCD cameras (7) are arranged side by side along the width direction of the steel strip and are suitable for surface detection of steel strips with a width of more than 400mm. The overlapping detection area of ​​multiple CCD cameras (7) is 5-50mm.

3. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: The camera bracket (6) is a combination bracket made of aluminum alloy profiles, including a main vertical lifting rod and a horizontal adjusting arm; The horizontal adjustment arm is installed at the output end of the main body vertical lifting rod, and the end of the horizontal adjustment arm is hinged to an angle adjustment gimbal. The CCD camera (7) is flanged and connected to one side of the angle adjustment gimbal. The vertical lifting rod is used to realize the 0.5-2m height adjustment of the CCD camera (7), the horizontal adjustment arm is used to realize the 0-1m horizontal extension and retraction of the CCD camera (7), and the angle adjustment gimbal is used to realize the 360° horizontal rotation and ±30° vertical pitch of the CCD camera (7).

4. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: A dovetail groove (4) is provided between the light source bracket (3) and the LED array light source (5). The back of the LED array light source (5) is an aluminum alloy shell with a dovetail track, which is slidably connected to the inner wall of the dovetail groove (4) and fixedly connected to the light source bracket (3) with bolts.

5. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: The wavelength range of the ultraviolet array LED light source is 200-400nm, and the wavelength range of the red LED strip light source is 620-750nm.

6. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: The oil mist purging device also includes auxiliary purging components, which consist of two components, respectively installed 100mm above the CCD camera (7), with the nozzles aligned with the lens; and 100mm above the light source detection end of the LED array light source (5), with the nozzles aligned with the light source detection surface.

7. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: The slit length of the main nozzle (9) is 30-60mm, the air inlet pressure is 0.2-0.3Mpa, and the air source is dry compressed air with a dew point below -20℃.

8. The high-speed multispectral fusion detection device for steel strip surface defects according to claim 1, characterized in that: It also includes a computer system, which is used to observe the images acquired by the CCD camera (7) in real time and has a built-in image processing and recognition system. Based on the images acquired by the CCD camera (7), it identifies the surface defect information of the steel strip and performs statistical positioning of the defects in combination with the steel strip information.

9. A high-speed multispectral fusion detection method for steel strip surface defects, using the high-speed multispectral fusion detection device for steel strip surface defects as described in any one of claims 1-8, characterized in that: Includes the following steps: S1. Obtain basic information about the steel strip and set the scanning frequency and width of the CCD camera (7) according to the speed and width of the steel strip; S2. Start the oil mist blowing device to clean the surface of the steel strip of debris and oil mist; S3. Start the LED array light source (5), adjust the angle and brightness, and calibrate the parameters of the CCD camera (7); S4, CCD camera (7) acquires images of the steel strip surface at a set frequency and feeds them back to the computer system for real-time observation; S5. The computer system classifies and stores images according to the basic information of the steel strip, combined with time and light color. S6. The image is preprocessed by the image processing and recognition system, and the preprocessed image is compared with the feature map. The defect image is stored and the defect location is marked based on the time information. S7. If there are no similar defect features, proceed to S8; if there are similar features, proceed to S9. S8. Continue to acquire the next frame image, repeat S4-S7, and execute S10 when the detection is complete. S9. Automatically mark the defect location, identify the suspected type, collect defect data, and provide personnel prompts. Then repeat S4-S7. Once the inspection is complete, execute S10. S10. Generate an inspection report (including full information on defects / steel strips) and archive and store the report in conjunction with the push direction.

Images