JP5659540B2 - Steel plate surface defect inspection method and apparatus - Google Patents

Description

  The present invention relates to a steel sheet surface defect inspection method and apparatus in a production line for steel sheet materials such as slabs, thick plates, hot rolled steel sheets, and the like, and in particular, defect inspection for steel sheets having a large surface property variation for each steel sheet or within a steel sheet. It is related with what is suitable for.

  In the production line of steel plate materials (hereinafter referred to as “steel plates”) such as slabs, thick plates, hot rolled steel plates, etc., if any flaws or defects occur on the surface of the steel plates that are continuously conveyed and passed, the steel plates are rejected. Or, it is put on hold and treatment such as care is performed. For this reason, it is necessary to detect the surface defect of a steel plate, the steel plate on a conveyance line is visually monitored, and the presence or absence of a surface defect is judged.

  Thick steel plates have a maximum width of 5 m or more and a maximum length of 25 m or more. Therefore, it is difficult to monitor all of the steel plate surface (and back surface) being transported by a single inspector. The steel sheet surface inspection is performed by arranging.

  In addition, although it is possible to detect the presence or absence of defects on the surface of the steel sheet during inspection during transport, it is difficult to discriminate and judge in detail, so the steel sheet in which defects are detected is transferred to an offline position and stopped. In the state, the defect position and the defect state are confirmed, and necessary measures are taken.

  However, in the visual determination, the load on the inspector is high, and there is a possibility of missing a defect or a variation in defect detection level due to individual differences. Therefore, it is difficult to perform a uniform surface inspection on the entire surface of the steel sheet.

  In response to the above-mentioned problems of visual inspection, a CCD camera is used and the projection angle and reception angle for detecting irregularities on thick plates are also set. There is a technique for mitigating changes (for example, Patent Document 1).

JP 2002-303582 A

  However, since the technique disclosed in Patent Document 1 uses a monochrome CCD camera, there is a problem that it cannot be applied to detection of uneven color defects or surface variations in color tone. Moreover, since it is a general method for setting a fixed threshold value for discriminating between a normal part and an abnormal part, the surface properties (texture) of slabs, thick plates, hot rolled materials, etc. are rough, When the surface property variation in the steel plate is large, the technique of Patent Document 1 does not solve problems such as overdetection and non-detection.

  In the present invention, in view of the problems of these prior arts, it is possible to detect a defective portion (abnormal portion) while suppressing over-detection / non-detection with a steel sheet having a large surface texture variation including the color tone as an inspection target. An object of the present invention is to provide a steel sheet surface defect inspection method and apparatus.

  The above problems can be solved by the following invention.

  [1] The surface of the steel sheet irradiated with light is imaged by a color camera, and at least one value of average luminance, maximum luminance, minimum luminance, and luminance dispersion of each R / G / B component of the captured color image is calculated, A steel plate surface defect inspection method, wherein the calculated value is compared with a predetermined set value to determine the presence and / or extent of defects on the steel plate surface.

  [2] In the steel sheet surface defect inspection method according to [1], the color image is divided into a plurality of regions, and the average luminance, the maximum luminance, and the minimum luminance of each R / G / B component for each of the divided regions. Calculating at least one value of luminance dispersion, comparing the calculated value with the set value, or based on the difference between the divided areas of the calculated value A method for inspecting a surface defect of a steel sheet, characterized in that:

  [3] In the steel sheet surface defect inspection method according to any one of [1] or [2], the set value is set based on an evaluation result by evaluating a surface property of the steel sheet from each of the calculated values. A method for inspecting a surface defect of a steel sheet.

  [4] In the steel sheet surface defect inspection method according to any one of [2] and [3], in dividing into a plurality of regions, after dividing the region up to a predetermined division limit, A steel sheet surface defect inspection method characterized by performing region division corresponding to the surface property distribution of a steel plate by evaluating surface properties and reintegrating adjacent divided regions having equivalent surface properties.

[5] A light source that irradiates light on the surface of the steel sheet, a color camera that captures a color image of the steel sheet surface irradiated by the light source,
An image processing apparatus that calculates at least one value of average luminance, maximum luminance, minimum luminance, and luminance variance of each R / G / B component of the captured color image, and compares the calculated value with a predetermined set value. A steel sheet surface defect inspection apparatus comprising: a signal processing apparatus that determines the presence and / or extent of defects on the steel sheet surface or estimates a site where defects exist.

  According to the present invention, regarding a steel plate material having a large variation in surface properties such as a slab, a thick plate, and a hot rolled material, the surface of the material is color-imaged, and each R / G / B (red, green, blue) of the color image is obtained. Since the average luminance, maximum luminance, minimum luminance, and luminance variance of the component are calculated and it is determined whether or not the color image contains a defect, defects such as color unevenness and color tone It becomes possible to detect a defective portion (abnormal portion) while suppressing the influence of fluctuation and suppressing overdetection and non-detection.

  In addition, the captured color image is divided into a plurality of areas, and the average luminance, maximum luminance, minimum luminance, and luminance distribution are calculated for each area, and the values of each area are compared with the set values or the values of different areas. Since the presence / absence, degree, and location of the defect are estimated, it is possible to accurately detect the defect portion (abnormal portion) regardless of the change in surface properties.

  Furthermore, in the present invention, the estimation is performed on the image of the steel plate surface imaged in advance, and the determination result is also displayed in guidance along with the captured image, so that the inspector performs the inspection based on the image and guidance display, In addition to reducing the load on the inspector, it is also possible to reduce variations in the inspection level due to individual differences among the inspectors.

It is a figure which shows the example of application to the thick steel plate conveyance line of Embodiment 1. FIG. It is a figure which shows the process outline | summary of the steel plate surface defect inspection method which concerns on Embodiment 1. FIG. It is a flowchart which shows the example of an abnormal / defective part estimation procedure. It is a flowchart which shows the example of an abnormal / defective part estimation procedure. It is a flowchart which shows the example of an abnormal / defective part estimation procedure. It is a figure which shows the process outline | summary of the steel plate surface defect inspection method which concerns on Embodiment 2. FIG. It is a figure which shows the example of application to the thick steel plate conveyance line of Embodiment 2.

[Embodiment 1]
First, the inventors have focused on the fact that the conventional visual inspection performs surface property determination and defect presence / absence determination in consideration of the color tone of the steel material surface, etc. From the viewpoint of improving the color, a color image is used instead of a monochrome image.

  And when the steel sheet surface texture (texture) is relatively smooth and the fluctuation is small, the luminance distribution is concentrated in a narrow luminance range, and the R / G / B balance is almost equal (average luminance, maximum luminance, The minimum luminance value is substantially equal). However, if there is an abnormal part in the image area, that is, a part that fluctuates with respect to the texture of the steel sheet (a part with high brightness or a part with low brightness), the brightness distribution of the area It has a shape different from that of the region and the knowledge that the balance of R / G / B is also changed.

  Based on the knowledge, the properties of the entire steel sheet are evaluated from the color image data of the imaged steel sheet (characteristic values such as luminance data, luminance distribution, average luminance, maximum / minimum luminance of each R / G / B component) When it is determined that the fluctuation (surface texture fluctuation) of the steel sheet image is outside the predetermined range and the possibility of the presence of an abnormality / defect is high, the steel sheet image is divided into a plurality of areas, and each divided area or area The present invention was conceived by repeating the evaluation between the two to estimate the abnormality / defect site in the steel sheet (image).

  FIG. 1 is a diagram illustrating an application example of the first embodiment to the thick steel plate conveyance line. In the figure, 1 is a light source, 3 is an imaging device which is a color line CCD camera, 4 is a signal processing device, 5 is an image display device, 7 is a transport roll, 8 is a steel plate, 9 is a PLG, 10 is an image processing unit, 11 Represents a determination processing unit.

  The steel plate conveyance line is composed of conveyance rolls 7 at regular intervals, and conveys the steel plate 8 on the conveyance line by the rotation of the roll. In this conveyance line, a thick steel plate having a maximum width of about 5 m and a maximum length of about 25 m is conveyed. And in order to image the steel plate surface, the light source 1 and the imaging device 3 are arrange | positioned on the line upper surface. When imaging the back surface of the steel plate, the light source 1 and the imaging device 3 may be similarly arranged on the lower surface of the line.

  The incident angle of the light source 1 is, for example, 15 °, and the light receiving angle of the imaging device 3 is, for example, 5 °. The diffuse reflection component shifted by 10 ° from the regular reflection position is received. This is because, at the regular reflection position, the image (brightness) fluctuation due to the influence of the steel sheet surface property becomes large, and the discriminability on the image of the abnormal part such as a defect is deteriorated. The incident angle of the light source and the light receiving angle of the camera are examples, and the angles of the light source and the camera may be appropriately set to optimum values according to the target line and the target material.

  The light source for projecting the steel plate surface is a linear light source in which optical fibers are arranged in a straight line (the linear arrangement direction is perpendicular to the transport direction), and the light is emitted from the fiber end and placed on the front surface. The light is condensed by the cylindrical lens (condensing direction is the conveying direction), and the surface of the steel sheet is irradiated with line-shaped light. In addition, as light sources, LEDs are arranged in a straight line and light is irradiated on the surface of the steel sheet through a cylindrical lens, or a plurality of projectors are arranged at the imaging position. You may choose according to.

  The imaging device 3 is a color line CCD camera, and sets a field of view at a location irradiated with light on the surface of the steel plate. Note that the irradiation range in the steel plate conveyance direction (the region where light is collected and irradiated) is set so that the visual field position of the camera is within the projection range within the plate width range of the steel plate to be inspected. If a single camera or light source cannot measure all of the width direction of the steel sheet (the direction orthogonal to the transport direction), a plurality of cameras and light sources may be installed in the width direction.

  Also, a PLG (pulse generator) 9 is installed in the line, a pulse is generated at every predetermined transport distance (for example, 0.25 mm, 0.5 mm, etc.), and imaging is performed in synchronization with the pulse (color line CCD camera). (Line scan), imaging is performed at a constant long pitch in the steel plate conveyance direction regardless of the conveyance speed.

  An imaging signal imaged by the imaging device 3 is input to the image processing unit 10 in the signal processing device 4, and two-dimensional steel plate image data is generated by arranging the signals in the transport direction. The generated image data is output to the determination processing unit 11 and the image display device 5 installed at the visual inspection position.

  The determination processing unit 11 processes the steel plate image data according to the procedure shown in FIGS. 2 to 5 described later, and estimates an abnormality / defect portion in the steel plate. The estimated position information of the abnormality / defect portion obtained by the determination processing unit 11 is output to the image display device 5 together with the image at the timing when the measurement location reaches the visual inspection position, and is used as support information for the visual inspection. Is done. When there is an abnormality / defect estimation part in the display image range, the outer frame display, blinking display, color change display, etc. of the estimation part are performed, and the attention of the inspector is raised to prevent the inspection from being overlooked. An alarm sound, announcement, etc. may be given.

  The image display device 5 is installed so that the front surface (upper surface) and the back surface (lower surface) can be seen. For example, a plurality of screens are provided, and a screen in which the captured image data of the entire width is reduced and divided is displayed, and defects are generated. It is good to display the image of the part which is easy to do (for example, the width edge part of a steel plate) at equal magnification. Further, the configuration, display position, display magnification, display timing, and the like of the image display device may be arbitrarily changed according to the operation conditions and surface properties. Further, the image processing unit 10 may perform differentiation processing for emphasizing the change point of the luminance signal of the image and display the image.

  FIG. 2 is a diagram illustrating a processing outline of the steel sheet surface defect inspection method according to the first embodiment. First, after performing the property evaluation (1) of the entire steel sheet, the steel sheet is divided into the longitudinal direction (2) -1 and the width direction (2) -2, and the normal part or the abnormal part is located at any position in the longitudinal direction or the width direction. A rough evaluation is performed. Then, after rough determination (3) of the abnormal part, the abnormal part is subdivided and defect / abnormal part estimation (4) is performed for each subdivision region.

  3 to 5 are flowcharts showing an example of an abnormality / defective part estimation procedure. Hereinafter, the processing procedure will be described in detail from FIG.

(1) Evaluation of the luminance distribution of each R / G / B component in the entire image area (I _a : average luminance, I _max : maximum luminance) from the distribution of luminance data of each R / G / B component in the entire color image area , I _min : minimum luminance, I _d : calculation of luminance dispersion). (Step S101)
The entire steel plate (image) is evaluated based on the difference between the evaluation results (calculated values) of the components and the evaluation results between the components. When the luminance variation (I _d ) of each component is small (No in step S102), it can be evaluated that the properties of the steel plate are relatively uniform, and when the luminance variation (I _d ) of each component is large (step) In 102 (Yes), it is determined that the property variation within the steel plate is large (abnormal portions and defects are present) (step S105).

Even when the variation (I _d ) is small, the absolute value of the deviation between the maximum value and the minimum value ((I _max -I _a ), (I _a -I _min )) is large with respect to the average value of each component. In (Yes in step S103), it is determined that a portion having an abnormal surface texture, that is, a defect or an abnormal portion exists in the entire steel sheet (step S105).

  A case where judgment criteria are provided for the calculation data of each component for the entire color image of the steel plate, and the properties of the steel plate are uniform and it can be evaluated that the possibility of the presence of an abnormality / defect portion is low (No in step S104). The evaluation is finished (step S106).

  Further, it is determined that an abnormal portion exists even when the difference in the calculated values between the R / G / B components is large (Yes in step S104).

  (2) As a result of the evaluation of the entire steel sheet, when it is determined that there is an abnormality / defect (step S105), an abnormal region is estimated (processing after step S201, see FIG. 4).

  The steel plate image is divided in the longitudinal direction (step S201), and luminance distribution evaluation (calculation of average luminance, maximum / minimum luminance, luminance variance) of each R / G / B component in each region is performed (step S202). Then, the difference between the divided regions is calculated for each R / G / B component with respect to the calculated average luminance, maximum / minimum luminance, and luminance variance (step S203).

  Next, based on the calculated difference, it is determined whether or not there is variation in the luminance distribution evaluation between the divided regions (step S204). If the evaluation results between the divided areas are different (Yes in step S204), the area having the largest number of times that the difference from the other divided areas is equal to or greater than a predetermined value is extracted (step S205), and the length including the defect and the abnormal part The area is determined (step S207).

  If the difference between the divided regions is not significantly different (No in Step S204), it is determined that only the specific longitudinal region is not abnormal and the specific portion in the width direction is abnormal in the longitudinal direction (Step S206). ).

  (3) Next, moving to FIG. 5, the image is divided in the width direction (step S <b> 301), and R / G / B components in each region are evaluated. Similar to the processing in (2), abnormalities in the steel sheet The width region that is considered to contain a defect is estimated (steps S302 to S307).

  Next, it is determined that there is an abnormality or defect in a region where the width direction divided region and the longitudinal direction divided region estimated so far have an abnormality / defect overlapped with each other (step S308).

  When it is necessary to divide the area again with respect to the area determined in step S308 (Yes in step S309), the process returns to step S201 and repeats the processes described in (2) to (3) above. Further refine the position.

  Note that the further division of the region may be appropriately determined from the required position resolution, and may be completed with only one calculation in some cases. In addition, the divided area to be set first may be appropriately determined in consideration of the calculation time.

[Embodiment 2]
In this embodiment, the surface properties of the steel sheet are obtained by evaluating the R / G / B component for the entire color image or for each divided region and obtaining a representative value (evaluation value) for the color images of the captured steel sheet front and back surfaces. Is determined, and an abnormality / defect in the steel sheet is determined by setting an appropriate set value corresponding to the surface property of the steel sheet.

  FIG. 6 is a diagram illustrating a processing outline of the steel sheet surface defect inspection method according to the second embodiment. The following description is made according to the figure.

  (1) A divided image of R / G / B data (R / G / B single color luminance) and a monochrome (black / white) image are generated from the steel plate surface image.

  (2) Each image data is divided into a plurality of predetermined areas. A width direction equal division, a constant width division, a longitudinal direction equal division, or a constant length division is performed.

  (3) For luminance data of R / G / B and monochrome (black / white), for example, the maximum luminance, the minimum luminance, the maximum distribution luminance, the average luminance, and the luminance variance are calculated as the luminance distribution evaluation.

  (4) The surface texture is determined using the luminance dispersion, the maximum luminance, and the minimum luminance. That is, if the luminance dispersion is larger than a preset specified value, it is determined that the surface property of the area is rough. When the luminance dispersion is small, it is determined that the surface property of the area is uniform. However, even when the variance of luminance is small, if the difference between the maximum luminance and the minimum luminance is large, it is determined that a rough surface portion or an abnormal portion exists in a relatively uniform region.

  If the luminance distribution evaluation results have the same tendency for all of R / G / B and monochrome (black / white), there is no color tone deviation on the steel sheet surface, and even if the surface texture is rough, unevenness, It can be judged as a variation due to kimono.

  Moreover, regarding R / G / B, when there is a difference in tendency in the luminance distribution evaluation results, it can be determined that there is a color tone bias on the steel sheet surface due to surface changes such as rust.

  (5) A set value (hereinafter referred to as a threshold value) related to the area is set based on the luminance evaluation result. The threshold is set for each of R / G / B and monochrome (black / white). Further, for example, the upper threshold value for detecting the high luminance part and the lower threshold value for detecting the low luminance part are set as follows.

Threshold setting example-1
Upper threshold = average luminance level + (maximum luminance−average luminance) × coefficient (setting constant)
Lower threshold = average luminance level− (average luminance−minimum luminance) × coefficient (setting constant)
The coefficient is changed depending on the surface properties.

Threshold setting example-2
Upper threshold = average luminance level + 2σ or average luminance level + 3σ
Lower threshold = average luminance level-2σ or average luminance level-3σ
Here, σ represents luminance dispersion, and a value that is added to or subtracted from the average luminance level is changed by 2σ or 3σ as shown in the above equation depending on the surface properties.

  (6) A threshold determination process is performed on the R / G / B and monochrome (black / white) data of each region based on the threshold setting. A portion having a higher luminance than the upper threshold and a portion having a lower luminance than the lower threshold are determined as abnormal portions.

  (7) The abnormal / defect portions are detected by integrating the abnormal portion determination (threshold determination) processing results for each of R / G / B and monochrome (black / white). Depending on the type of target defect, the type of steel sheet, etc., all or any one or combination of R / G / B and monochrome (black / white) is determined as a threshold / defect portion. Shall.

  FIG. 7 is a diagram illustrating an application example to the thick steel plate conveyance line of the second embodiment. In the figure, 2a represents a light source (lamp), 2b represents a light source (projector), and 12 represents an image recording apparatus, and the other symbols are the same as those in FIG. Here, the light source for projecting the surface of the steel sheet is composed of a light source (lamp) 2a and a light source (projector) 2b. The lamp section collects the lamp light and enters the optical fiber of the projecting section. In the light projecting unit, incident light is transmitted to the light projecting unit side by an optical fiber. On the light projecting unit side, optical fibers are arranged in a straight line, light is projected from the end of the fiber, projection light is collected by a cylindrical lens arranged on the front surface, and line light is projected onto an object. .

  In the present embodiment, a halogen lamp light source is used as the lamp unit, and two lamp units are connected to one light projecting unit to enable high-intensity light projection. A light projecting unit having a light projecting unit length of 1.8 m is used, and the lens and the installation position are adjusted so that the light projecting range of the object surface is 1.8 m × 25 mm.

  The light projection range (light projection width) is determined from the plate width range of the target steel plate, the light source, and the installation angle of the camera, and the imaging position of the camera is within the light projection range in the plate width range of the target steel plate. . When the plate width range is narrow, it is possible to narrow the light projection width (condensed) and increase the luminance.

  And the light source is enabling the light projection with respect to the full width of a conveyance line (steel plate) by installing three sets in a conveyance line (steel plate) width direction. In the present embodiment, a halogen lamp is used, but a metal halide light source or the like can also be used. In addition, although the light source configuration of the lamp unit and the light projecting unit is used, a linear light source using a high-intensity LED array can be used.

  The steel plate conveyance line 6 is constituted by a conveyance roll (diameter 400 mm × length 5000 mm) having a constant interval (for example, 1 m interval), and conveys the steel plate 8 on the conveyance line by the rotation of the roll. In the transfer line, a thick steel plate having a maximum width of 5250 mm and a maximum length of 26 m is transferred.

  In order to image the steel plate surface and the steel plate back surface, the light sources (lamps and projectors) 2a and 2b and the imaging device (color line CCD camera) 3 described above are arranged on the upper surface and the lower surface of the line. The light projection angle from the light source to the steel plate is 15 ° with respect to the vertical axis on both the upper and lower surfaces.

  The color line CCD camera is installed at an angle of 5 ° with respect to the vertical axis, and the angle is shifted by 10 ° from the regular reflection position with respect to the projection angle of 15 °. This is because, at the regular reflection position, image (brightness) fluctuations due to the influence of the steel sheet surface properties are large, and the discriminability on the image of abnormal parts such as defects is deteriorated. Since the captured image changes depending on the angle of the light source and the camera, it is necessary to optimally set the angle of the light source and the camera depending on the target line and the target material.

  A plurality of color line CCD cameras 3 are installed in the line (steel plate) width direction, and the light projection position is imaged by the light source. The color line CCD camera 3 is mounted with a lens having a focal length of 35 mm, and the distance to the steel plate is set so that the field of view of the camera is 550 mm (in this embodiment, the color line CCD camera with 2048 pixels). And the resolution per pixel of the camera is set to 2048 of 0.275 mm).

  By installing 10 color line CCD cameras in the line (steel plate) width direction, a field of view that is greater than the steel plate width is secured. The field of view of the camera is arranged so that the end wraps and the wrap portion images the same position on the steel plate, and the entire surface of the steel plate can be imaged without a gap.

  In the present embodiment, in order to perform imaging in synchronization with the conveyance speed (steel plate speed) of the conveyance line, a pulse is generated every conveyance distance of 0.275 mm by a PLG (pulse generator) 9 installed in the conveyance line, By performing imaging (line scanning of a color line CCD camera) in synchronization with this, imaging is performed at a constant length (0.275 mm) pitch in the longitudinal direction of the steel sheet regardless of the conveyance speed.

  An imaging signal from the color line CCD camera is input to the imaging processing unit 10 in the signal processing device 4 to generate two-dimensional steel plate image data. The generated image data is transmitted to the determination processing unit 11 and the image display device 5 installed at the inspection position.

  In the determination processing unit, the above-described processing of the steel plate image data is performed, the region division of the steel plate image, the R / G / B and monochrome (black / white) luminance distribution in the region are evaluated, and the surface property evaluation for each region. The threshold value is set, and the abnormality / defective part is detected by the threshold determination. The detected position information of the abnormality / defect portion obtained by the inspection / determination processing unit is transmitted to the image display device 5.

  In the present embodiment, the image display apparatus displays the steel plate image data received from the imaging processing unit and the estimated position information of the abnormality / defect portion received from the inspection / determination processing unit. When a steel plate color image is displayed and there is an abnormality / defect estimation part in the display image range, an outer frame display, blinking display, color change display, etc. of the estimation part are performed. Alarm sounds, announcements, etc. are used to alert the inspector and prevent missed inspections.

  As the image display device, four surfaces are provided for the front surface (upper surface) and the back surface (lower surface), and the captured image data of the entire width is reduced and divided and displayed on two of the four surfaces. Shows an approximately 500mm wide image of both edges of the line (to monitor the edge where defects are likely to occur). The configuration of the image display device can be arbitrarily changed, and the number of screens can be increased or decreased and the image display range can be changed.

  In addition, the display image can be displayed at an arbitrary magnification and reduced at an arbitrary magnification, and it is also possible to enlarge and display a portion estimated to be a defect of the steel sheet to confirm details. .

  In this embodiment, the display of image data on the image display device 5 installed at the inspection position starts when the imaged steel plate reaches the inspection position, and displays an image in synchronization with the conveyance speed of the steel plate. However, the image display itself can also be displayed at an arbitrary timing asynchronously with the steel plate conveyance.

  By installing the image recording apparatus 10, it is possible to store the captured steel sheet image, redisplay the captured and transported steel sheet image, and perform re-examination and confirmation. In addition, abnormalities / defects are displayed on the image display device when abnormalities / defects are detected, but when that part reaches the visual inspection position, the steel plate transport is decelerated and stopped, It is also possible to perform visual confirmation. Furthermore, at the time of abnormality / defect portion detection, pass / fail judgment, hold processing is performed, steel plate conveyance is terminated, re-inspection with image data in the image recording device is performed on the hold processed steel plate, and final treatment is performed. It is also possible to operate in the form of determining

  In this embodiment, the light source cameras are arranged on the front and back surfaces (upper and lower surfaces). However, it is also possible to install the light source cameras on the side surfaces and to capture and display the steel plate edge portion.

1 Light source 2a Light source (lamp)
2b Light source (sender)
3 Imaging device (color line CCD camera)
4 Signal Processing Device 5 Image Display Device 6 Steel Plate Transport Line 7 Transport Roll 8 Steel Plate 9 PLG
DESCRIPTION OF SYMBOLS 10 Image processing part 11 Determination processing part 12 Image recording apparatus

Claims (4)

The surface of the steel sheet irradiated with light is imaged with a color camera,
Calculated R / G / B components and monochrome average luminance of the captured color image, the maximum brightness, minimum brightness, the distributed value of intensity,
After comparing the calculated value with a predetermined set value to determine the presence and / or extent of defects on the steel sheet surface, or to estimate the site where the defects are present on the entire steel sheet surface,
When it is determined that there is an abnormality / defect, the image of the steel sheet surface is divided in the longitudinal direction and the width direction, and among the divided regions, the luminance distribution is evaluated in the region divided in the longitudinal direction. After determining the area where the defect exists, evaluate the luminance distribution of the area divided in the width direction to determine the area where the abnormality / defect exists,
A region where the region divided in the longitudinal direction and the region divided in the width direction, which is determined to have the abnormality / defect portion, is defined as a region where the abnormality / defect portion exists,
The area is divided into smaller areas in the longitudinal direction and the width direction, and the luminance distribution of each divided area is evaluated. By repeating this process, surface properties are judged, uneven parts, deposits, and color tone are present. A method for inspecting a surface defect of a steel sheet, wherein the position of an abnormality / defect including information on whether or not there is a defect is estimated. In the steel sheet surface defect inspection method according to claim 1,
The set value is
A steel sheet surface defect inspection method, wherein the surface properties of a steel sheet are evaluated from the calculated values and set based on the evaluation result. In the steel sheet surface defect inspection method according to claim 1 or 2,
When dividing into multiple areas,
After dividing the area up to a predetermined division limit, the surface texture is evaluated for each divided area, and the area division corresponding to the surface texture distribution of the steel sheet is performed by reintegrating adjacent divided areas having the same surface texture. A steel sheet surface defect inspection method characterized by the above. A light source for irradiating light on the surface of the steel sheet;
A color camera that captures a color image of the steel sheet surface irradiated by the light source;
An image processing apparatus for calculating R / G / B components and monochrome average luminance of the captured color image, the maximum brightness, minimum brightness, the distributed value of intensity,
A signal processing device that compares the calculated value with a predetermined set value to determine the presence and / or extent of defects on the surface of the steel sheet, or to estimate a site where defects exist;
A steel sheet surface defect inspection apparatus comprising:
The signal processing device includes:
After performing the property evaluation of the entire steel sheet surface,
When it is determined that there is an abnormality / defect, the image of the steel sheet surface is divided in the longitudinal direction and the width direction, and among the divided regions, the luminance distribution is evaluated in the region divided in the longitudinal direction. After determining the area where the defect exists, evaluate the luminance distribution of the area divided in the width direction to determine the area where the abnormality / defect exists,
A region where the region divided in the longitudinal direction and the region divided in the width direction, which is determined to have the abnormality / defect portion, is defined as a region where the abnormality / defect portion exists,
The area is divided into smaller areas in the longitudinal direction and the width direction, and the luminance distribution of each divided area is evaluated. By repeating this process, surface properties are judged, uneven parts, deposits, and color tone are present. A steel sheet surface defect inspection apparatus characterized by performing a process of estimating a position of an abnormality / defect including information on whether or not there is a defect.