Strip steel surface indentation defect detection system and method

Abstract

The invention discloses a strip steel surface indentation defect detection system, which comprises an image acquisition unit used for acquiring a strip steel surface image; the detection image processing unit is used for filtering and segmenting the surface image of the strip steel, identifying a suspicious target and outputting a defect detection result; the system server is used for completing interaction of various signals, messages and instructions and summarizing and post-processing of defect detection results; and the detection terminal is used for outputting and displaying the defect detection result. The invention further discloses a strip steel surface indentation defect detection method. Through the combination of the red strip-shaped light source, the blue strip-shaped light source and the single color linear array industrial camera, on-line detection of surface indentation defects can be realized under the condition that the normal operation rhythm of a unit is not influenced.

Description

Technical Field

[0001] The present invention relates to machine vision detection technology, and more specifically, to a strip steel surface indentation defect detection system and method. Background Art

[0002] The indentation defect on the strip surface is an important factor affecting the quality of the strip, especially for high-quality strip, which has very strict requirements for surface defects. The machine vision surface online inspection system is to detect the surface quality of the strip online and control the generation of surface defects. Its principle is to irradiate the light emitted by the light source to the surface of the strip, and collect the light reflected by the surface of the strip online through the industrial camera. Due to the difference in reflection properties between the defective area and the non-defective area, there is a difference in grayscale between the defective area and the non-defective area in the image collected by the industrial camera. The defective area can be detected and the defects can be classified and identified through the image processing algorithm.

[0003] The lighting method is the key to the machine vision surface online inspection system, which directly affects the contrast of defects in the image. There are two commonly used lighting methods in surface inspection technology, one is bright field lighting and the other is dark field lighting. The bright field lighting method is to place the industrial camera in the optical path of the reflected light, and the dark field lighting method is to place the industrial camera outside the optical path of the reflected light. If there is no defect on the surface of the strip, the intensity of the reflected light should be evenly distributed, but if there is a defect on the surface of the strip, the intensity of the reflected light will change in the defective area, and the defective area can be detected by the change of grayscale in the image.

[0004] In existing patent applications, for example, Chinese patent application No. 202110814180.7 proposes a strip surface defect detection method to improve the accuracy of strip surface defect detection in a small sample environment. Another example is Chinese patent application No. 201310127783.5, which proposes a strip surface defect detection method and system, which overcomes the impact of a narrow space on the surface quality detection system and improves the detection effect of each surface of the strip. However, the above two patent technologies cannot obtain and process three-dimensional information of strip surface defects. Summary of the invention

[0005] In view of the defects existing in the prior art, the purpose of the present invention is to provide a system and method for detecting surface indentation defects of strip steel. Through the combination of two red and blue strip light sources and a single color linear array industrial camera, online detection of surface indentation defects can be achieved without affecting the normal operating rhythm of the unit.

[0006] To achieve the above object, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a strip steel surface indentation defect detection system, comprising:

[0008] An image acquisition unit, used for acquiring an image of the strip surface;

[0009] A detection image processing unit is used to complete the filtering, segmentation, and identification of suspicious targets of the strip surface image, and output defect detection results;

[0010] System server, used to complete the interaction of various signals, telegrams, and instructions, as well as the aggregation and post-processing of the defect detection results;

[0011] The detection terminal is used to output and display the defect detection result.

[0012] Preferably, the image acquisition unit includes a first light source, a second light source, an industrial camera, a terminal box and a signal control card;

[0013] The lens of the industrial camera is arranged perpendicular to the surface of the steel strip;

[0014] The first light source and the second light source are respectively located on two sides of the industrial camera, and the first light source and the second light source have the same incident angle and the same distance from the surface of the steel strip;

[0015] The light emitted by the first light source and the second light source is irradiated on the same area of ​​the steel strip, and the industrial camera collects a color image of the area;

[0016] The first light source, the second light source, and the industrial camera are all connected to the terminal box, and the terminal box is connected to the signal control card.

[0017] Preferably, the industrial camera is a color linear array CCD industrial camera.

[0018] Preferably, the detection image processing unit is connected to the terminal box.

[0019] Preferably, the system server is connected to the image processing unit and the detection terminal respectively through a switch.

[0020] The second aspect of the present invention provides a method for detecting indentation defects on a steel strip surface, and the following steps are implemented using the steel strip surface indentation defect detection system provided by the first aspect of the present invention:

[0021] S1, the industrial camera collects a color image of the steel strip;

[0022] S2, performing channel separation on the color image to separate an R channel image I1 and a B channel image I2, which correspond to the first light source and the second light source respectively;

[0023] S3, calculating the gradient q of the surface of the strip steel 4 in the moving direction;

[0024] S4, calculating the surface depth information of the steel strip;

[0025] S5. Detecting surface indentation defects of the steel strip.

[0026] Preferably, the step S1 wherein the industrial camera collects the color image of the steel strip specifically includes:

[0027] The industrial camera is arranged in a direction perpendicular to the surface of the steel strip, the first light source is arranged on the side of the industrial camera that is in the same direction as the running direction of the steel strip, and the second light source is arranged on the side of the industrial camera that is opposite to the running direction of the steel strip;

[0028] The first light source and the second light source have the same incident angle and the same distance from the surface of the steel strip;

[0029] The light emitted by the first light source and the second light source is irradiated onto the same area of ​​the steel strip, and the industrial camera collects a color image of the area.

[0030] Preferably, the industrial camera is a color linear array CCD industrial camera, which separates the color image into an R channel image I1 and a B channel image I2.

[0031] Preferably, the step S3 of calculating the gradient q of the strip surface in the moving direction specifically includes:

[0032] The x-axis is established with the industrial camera acquisition direction and the strip movement direction as the y-axis. According to the right-hand screw rule, the z-axis is established. Then, according to the photometric stereo principle, the intensity I of the reflected light of the first light source and the second light source is:

[0033] I=Eξs·n (1)

[0034] In formula (1), E is the luminous intensity; ξ is the surface reflectivity; s is the unit vector of the incident direction of the light source, and n is the unit vector of the surface normal.

[0035] make:

[0036] s=[S x ,S y ,S z ] (2)

[0037]

[0038] Then formula (1) is expressed as:

[0039]

[0040] make:

[0041]

[0042]

[0043] The above step S4 calculates the surface depth information of the strip steel specifically including:

[0044] According to equations (5) and (6), equation (1) can be expressed as:

[0045]

[0046] For the first light source, we have:

[0047]

[0048] For the second light source, we have:

[0049]

[0050] Since the first light source and the second light source are strip light sources, the incident light fan is symmetrical about the xz plane, the illumination model is greatly simplified, and the light intensity is equal in the x direction of the strip, so:

[0051] S 1x =S 2x =0 (10)

[0052] E1=E2 (11)

[0053] Since the first light source and the second light source are at the same height from the strip surface, therefore:

[0054] S 1z =S 2z =S z (12)

[0055] Since the incident angles of the first light source 1 and the second light source 2 are the same and are located on both sides of the normal line of the strip surface, therefore:

[0056] S 1y =-S 2y =S y (13)

[0057] Substituting into equations (3) and (4), we have:

[0058]

[0059]

[0060] Divide equations (14) and (15), and according to E1=E2, we get:

[0061]

[0062] From formula (16), we can get:

[0063]

[0064] Since the surface of the strip is a plane as a whole, the three-dimensional information of the surface indentation of the strip is solved with q and the horizontal base plane as the optimization target:

[0065]

[0066] Where z(x,y) is the surface depth of the strip, its local three-dimensional shape is measured by photometric stereo, and the overall shape conforms to the horizontal base surface assumption; η is a coefficient used to adjust the optimization weight;

[0067] The color image captured by the industrial camera is separated into an R channel image I1 and a B channel image I2, and the three-dimensional information of the indentation defect on the surface of the strip is calculated using equations (17) and (18) for the R channel image I1 and the B channel image I2.

[0068] The present invention provides a strip surface indentation defect detection system and method. When the strip surface indentation defect is illuminated by multiple light sources, light incident at different angles will irradiate the strip surface indentation defect, forming shadows in different directions. The image collected by the industrial camera under the illumination of different light sources is reconstructed in three dimensions through photometric stereo technology to detect the strip surface indentation defect. The present invention performs three-dimensional reconstruction on the strip surface indentation defect under the illumination of multiple light sources through photometric stereo technology to restore the three-dimensional information of the strip surface indentation defect. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a schematic diagram of the framework structure of the strip steel surface indentation defect detection system of the present invention;

[0070] Figure 2 It is a schematic diagram of the dark field illumination arrangement of the first light source, the second light source, and the industrial camera in the strip steel surface indentation defect detection system of the present invention;

[0071] Figure 3 It is a schematic diagram of the bright and dark field illumination arrangement of the first light source, the second light source, and the industrial camera in the strip surface indentation defect detection system of the present invention;

[0072] Figure 4 It is a schematic flow chart of the method for detecting indentation defects on the surface of strip steel of the present invention. DETAILED DESCRIPTION

[0073] In order to better understand the above technical solution of the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0074] Combination Figure 1 and Figure 2 As shown, the present invention also provides a strip steel surface indentation defect detection system, which is composed of multiple sub-modules based on the C / S system architecture, including:

[0075] An image acquisition unit 11 is used to acquire the surface image of the steel strip 4, so as to effectively acquire the full-width surface image of the steel strip 4;

[0076] The detection image processing unit 7 is used to complete the filtering, segmentation, and identification of suspicious targets of the surface image of the steel strip 4, and upload the defect detection results to the system server 8;

[0077] The system server 8 is used to complete the interaction of various signals, telegrams, and instructions, as well as the aggregation and post-processing of defect detection results through the switch 9;

[0078] The detection terminal 10 is used to output and display the defect detection results to the user, who can view the detection results and perform data query and report statistics.

[0079] The image acquisition unit 11 includes a first light source 1, a second light source 2, an industrial camera 3, a terminal box 5 and a signal control card 6;

[0080] The lens of the industrial camera 3 is arranged perpendicular to the surface (upper surface, lower surface) of the steel strip 4;

[0081] The first light source 1 and the second light source 2 are respectively located on two sides of the industrial camera 3, and the first light source 1 and the second light source 2 have the same incident angle and the same distance from the surface of the strip steel 4;

[0082] The emitted light of the first light source 1 and the second light source 2 is irradiated on the same area of ​​the steel strip 4, and the industrial camera 3 collects a color image of the area;

[0083] The first light source 1 , the second light source 2 , and the industrial camera 3 are all connected to a terminal box 5 , and the terminal box 5 is respectively connected to a signal control card 6 and a detection image processing unit 7 .

[0084] Industrial camera 3 is a color linear array CCD industrial camera with three rows of pixels, R, G, and B, which are responsible for capturing the energy of red, green, and blue light respectively. Therefore, the images collected by industrial camera 3 can be separated into channels to obtain images of the three channels of R, G, and B, which correspond to the images under the illumination of red, green, and blue light sources respectively.

[0085] The industrial camera 3 is placed in a direction perpendicular to the surface of the steel strip 4 and also perpendicular to the moving direction of the steel strip 4. Two strip light sources (the first light source 1 and the second light source 2) with red and blue luminous colors are installed in the width direction of the steel strip 4. The luminous intensity of the blue light source is equal to that of the red light source. They are installed on both sides of the industrial camera 3 in the running direction of the steel strip 4, respectively, with the same incident angle α and the same height from the surface of the steel strip 4, forming a dark field lighting mode with the industrial camera 3.

[0086] Combination Figures 1 to 4 As shown, the present invention also provides a method for detecting indentation defects on the surface of a steel strip. When the indentation defects on the surface of the steel strip 4 are illuminated by multiple light sources, light incident at different angles irradiates the indentation defects on the surface of the steel strip 4, forming shadows in different directions. By using photometric stereo technology, the images collected by the industrial camera 3 under the illumination of different light sources are three-dimensionally reconstructed to detect the indentation defects on the surface of the steel strip 4. The following steps are implemented by using the indentation defect detection system on the surface of the steel strip 4 of the present invention:

[0087] S1, industrial camera 3 collects color images of steel strip 4;

[0088] S2, performing channel separation on the color image to separate an R channel image I1 and a B channel image I2, which correspond to the first light source 1 and the second light source 2 respectively;

[0089] S3, calculating the gradient q of the surface of the strip steel 4 in the moving direction;

[0090] S4, calculating the surface depth information of the steel strip 4;

[0091] S5. Detect surface indentation defects of the steel strip 4.

[0092] The above step S1 of collecting the color image of the steel strip by the industrial camera specifically includes:

[0093] The industrial camera 3 is arranged in a direction perpendicular to the surface of the steel strip 4, the first light source 1 is arranged on the side of the industrial camera 3 that is in the same direction v as the steel strip 4, and the second light source 2 is arranged on the side of the industrial camera 3 that is opposite to the direction v of the steel strip 4;

[0094] The first light source 1 and the second light source 2 have the same incident angle α and the same distance from the surface of the steel strip 4;

[0095] The emitted light from the first light source 1 and the second light source 2 irradiates the same area of ​​the steel strip 4, and the industrial camera 3 collects a color image of the area.

[0096] The industrial camera 3 is a color linear array CCD industrial camera, which separates the color image into an R channel image I1 and a B channel image I2, which correspond to the reflected light images of the first light source 1 and the second light source 2, respectively.

[0097] The above step S3 calculates the gradient q of the surface of the steel strip 4 in the moving direction, specifically including:

[0098] The x-axis is established with the industrial camera acquisition direction and the strip movement direction as the y-axis. According to the right-hand screw rule, the z-axis is established. According to the photometric stereo principle, the intensity I of the reflected light of the first light source 1 and the second light source 2 is:

[0099] I=Eξs·n (1)

[0100] In formula (1), E is the luminous intensity; ξ is the surface reflectivity; s is the unit vector of the incident direction of the light source, and n is the unit vector of the surface normal.

[0101] make:

[0102] s=[S x ,S y ,S z ] (2)

[0103]

[0104] Then formula (1) is expressed as:

[0105]

[0106] make:

[0107]

[0108]

[0109] The above step S4 calculates the surface depth information of the steel strip 4 specifically including:

[0110] According to equations (5) and (6), equation (1) can be expressed as:

[0111]

[0112] For the first light source 1, we have:

[0113]

[0114] For the second light source 2, we have:

[0115]

[0116] Since the first light source 1 and the second light source 2 are strip light sources, the incident light fan is bilaterally symmetrical about the xz plane, the illumination model is greatly simplified, and the light intensity in the x direction of the strip steel 4 is equal, therefore:

[0117] S 1x =S 2x =0 (10)

[0118] E1=E2 (11)

[0119] Since the first light source 1 and the second light source 2 are at the same height from the surface of the strip steel 4, therefore:

[0120] S 1z =S 2z =S z (12)

[0121] Since the incident angles of the first light source 1 and the second light source 2 are the same and are located on both sides of the normal line of the surface of the steel strip 4, therefore:

[0122] S 1y =-S 2y =S y (13)

[0123] Substituting into equations (3) and (4), we have:

[0124]

[0125]

[0126] Divide equations (14) and (15), and according to E1=E2, we get:

[0127]

[0128] From formula (16), we can get:

[0129]

[0130] Since the surface of the steel strip 4 is a plane as a whole, the three-dimensional information of the surface indentation of the steel strip 4 is solved with q and the horizontal base plane as the optimization targets:

[0131]

[0132] Where z is the surface depth of the strip 4, its local three-dimensional shape is measured by the photometric stereo method, and the overall shape conforms to the horizontal base surface assumption; η is a coefficient used to adjust the optimization weight;

[0133] The color image captured by the industrial camera 3 is separated into an R channel image I1 and a B channel image I2. The three-dimensional information of the surface indentation defect of the strip steel 4 is calculated using equations (17) and (18) for the R channel image I1 and the B channel image I2.

[0134] Example 1

[0135] like Figure 2As shown, the industrial camera 3 is placed in a direction perpendicular to the surface of the steel strip 4 and also perpendicular to the moving direction of the steel strip 4. Two strip light sources (first light source 1 and second light source 2) with red and blue luminous colors are installed in the width direction of the steel strip 4. The luminous intensity of the blue light source is equal to that of the red light source. They are installed on both sides of the industrial camera 3 in the running direction of the steel strip 4, respectively, with the same incident angle α and the same height from the surface of the steel strip 4, forming a dark field lighting mode with the industrial camera 3.

[0136] according to Figure 4 In the process shown, when the indentation defect on the surface of the steel strip 4 is illuminated by multiple light sources, light incident at different angles will form shadows in different directions. After the image collected by the industrial camera 3 is channel-separated, images at different incident angles will be obtained. Through the photometric stereo technology, the images collected by the industrial camera 3 under different light sources are 3D reconstructed to detect the indentation defect on the surface of the steel strip 4.

[0137] This example uses dark field illumination to effectively reconstruct the indentation defects on the strip surface in three dimensions, thereby improving the detection rate of indentation defects on the strip surface.

[0138] Example 2

[0139] like Figure 3 As shown, two strip light sources (first light source 1 and second light source 2) with red and blue luminous colors are installed in the width direction of the steel strip 4. The luminous intensity of the blue light source is equal to that of the red light source, and their incident angles α are equal, and their heights from the surface of the steel strip 4 are also equal. The industrial camera 3 is placed on the reflected light path of the first light source 1, forming a bright field illumination mode with the first light source 1, and forming a dark field illumination mode with the second light source 2.

[0140] according to Figure 4 In the process shown, when the indentation defect on the surface of the steel strip 4 is illuminated by multiple light sources, light incident at different angles will form shadows in different directions. After the image collected by the industrial camera 3 is channel-separated, images at different incident angles will be obtained. Through the photometric stereo technology, the images collected by the industrial camera 3 under different light sources are 3D reconstructed to detect the indentation defect on the surface of the steel strip 4.

[0141] This example uses a combination of bright and dark field illumination to effectively reconstruct the indentation defects on the strip surface in three dimensions, thereby improving the detection rate of indentation defects on the strip surface.

[0142] Those skilled in the art should recognize that the above embodiments are only used to illustrate the present invention, and are not intended to limit the present invention. As long as they are within the spirit of the present invention, any changes or modifications to the above embodiments will fall within the scope of the claims of the present invention.

Claims

1. A strip steel surface indentation defect detection system, characterized in that: include: An image acquisition unit, used for acquiring an image of the strip surface; A detection image processing unit is used to complete the filtering, segmentation, and identification of suspicious targets of the strip surface image, and output defect detection results; System server, used to complete the interaction of various signals, telegrams, and instructions, as well as the aggregation and post-processing of the defect detection results; The detection terminal is used to output and display the defect detection result.

2. The strip steel surface indentation defect detection system according to claim 1, characterized in that: The image acquisition unit includes a first light source, a second light source, an industrial camera, a terminal box and a signal control card; The lens of the industrial camera is arranged perpendicular to the surface of the steel strip; The first light source and the second light source are respectively located on two sides of the industrial camera, and the first light source and the second light source have the same incident angle and the same distance from the surface of the steel strip; The light emitted by the first light source and the second light source is irradiated on the same area of ​​the steel strip, and the industrial camera collects a color image of the area; The first light source, the second light source, and the industrial camera are all connected to the terminal box, and the terminal box is connected to the signal control card.

3. The strip steel surface indentation defect detection system according to claim 2, characterized in that: The industrial camera is a color linear array CCD industrial camera.

4. The strip steel surface indentation defect detection system according to claim 2, characterized in that: The detection image processing unit is connected to the terminal box.

5. The strip surface indentation defect detection system according to claim 4 is characterized in that: The system server is connected to the image processing unit and the detection terminal respectively through a switch.

6. A method for detecting indentation defects on a strip steel surface, characterized in that: The following steps are implemented using the strip surface indentation defect detection system as described in any one of claims 1 to 5: S1, the industrial camera collects a color image of the steel strip; S2, performing channel separation on the color image to separate an R channel image I1 and a B channel image I2, which correspond to the first light source and the second light source respectively; S3, calculating the gradient q of the strip surface in the moving direction; S4, calculating the surface depth information of the steel strip; S5. Detecting surface indentation defects of the steel strip.

7. The method for detecting indentation defects on the surface of a steel strip according to claim 6, characterized in that: The step S1 wherein the industrial camera collects the color image of the steel strip specifically includes: The industrial camera is arranged in a direction perpendicular to the surface of the steel strip, the first light source is arranged on the side of the industrial camera that is in the same direction as the running direction of the steel strip, and the second light source is arranged on the side of the industrial camera that is opposite to the running direction of the steel strip; The first light source and the second light source have the same incident angle and the same distance from the surface of the steel strip; The light emitted by the first light source and the second light source is irradiated onto the same area of ​​the steel strip, and the industrial camera collects a color image of the area.

8. The method for detecting indentation defects on a steel strip surface according to claim 7, characterized in that: The industrial camera is a color linear array CCD industrial camera, which separates the color image into an R channel image I1 and a B channel image I2.

9. The method for detecting indentation defects on the surface of a steel strip according to claim 8, characterized in that: The step S3 of calculating the gradient q of the strip surface in the moving direction specifically includes: The x-axis is established with the industrial camera acquisition direction and the strip movement direction as the y-axis. According to the right-hand screw rule, the z-axis is established. Then, according to the photometric stereo principle, the intensity I of the reflected light of the first light source and the second light source is: I=Eξs·n (1) In formula (1), E is the luminous intensity; ξ is the surface reflectivity; s is the unit vector of the incident direction of the light source, and n is the unit vector of the surface normal. make: s=[S x , S y , S z ] (2) Then formula (1) is expressed as: make: The above step S4 calculates the surface depth information of the strip steel specifically including: According to equations (5) and (6), equation (1) can be expressed as: For the first light source, we have: For the second light source, we have: Since the first light source and the second light source are strip light sources, the incident light fan is symmetrical about the xz plane, the illumination model is greatly simplified, and the light intensity is equal in the x direction of the strip, so: S 1x =S 2x =0 (10) E1=E2 (11) Since the first light source and the second light source are at the same height from the strip surface, therefore: S 1z =S 2z =S z (12) Since the incident angles of the first light source and the second light source are the same and are located on both sides of the normal line of the strip surface, therefore: S 1y =-S 2y =S y (13) Substituting into equations (3) and (4), we have: Divide equations (14) and (15), and according to E1=E2, we get: From formula (16), we can get: Since the surface of the strip is a plane as a whole, the three-dimensional information of the surface indentation of the strip is solved with q and the horizontal base plane as the optimization target: Where z is the surface depth of the strip 4, its local three-dimensional shape is measured by the photometric stereo method, and the overall shape conforms to the horizontal base surface assumption; η is a coefficient used to adjust the optimization weight; The color image captured by the industrial camera is separated into R channel image I1 and B channel image I2. The three-dimensional information of the indentation defect on the strip surface is calculated using equations (17) and (18) for the R channel image I1 and the B channel image I2.

Images