KR20130017421A - Apparatus and method for detecting strip defects in strip casting process - Google Patents

Abstract

PURPOSE: A slab underside inspecting system and an inspecting method are provided to detect periodic defects generated on a surface of a casting roll, thereby preventing additional damage to the casting roll. CONSTITUTION: A slab underside inspecting system is as follows. Original images of a surface of a slab are generated by photographing the slab generated by a casting roll(S10). Defects existing on the surface of the slab are detected by image-processing the original images(S20). The size of the detected defects is measured(S30). A kind of the detected defects is determined(S40). [Reference numerals] (AA) Start; (BB) End; (S10) Original image obtaining step; (S21) Boundary detection; (S22) Color value comparison; (S31) Horizontal length measurement; (S32) Longitudinal length measurement; (S40) Defect type determining step

Description

Apparatus and method for detecting strip defects in strip casting process

The present invention relates to a slab lower surface inspection system and an inspection method, and more particularly, a slab lower surface inspection system and inspection method for detecting defects of the slab by imaging a slab manufactured by a sheet casting process and image processing the captured image. Is about

Strip casting is a technology that is more advanced than thin slab, and it is a technology that can manufacture cast steel directly from molten steel without making slabs. It can omit the hot rolling process and mass production in a short time. Of course, it is known as a casting method that can significantly reduce the equipment investment costs, production costs and operating costs.

In this casting method, molten steel is supplied between a pair of casting rolls through a nozzle to form a molten steel pool formed by a pair of casting rolls and an edge dam, and then the contact between the casting roll and the molten steel is rotated while the casting roll is rotated. The molten steel is rapidly solidified by heat leakage inside the rolls to cast the cast steel.

However, in such a thin plate casting process, a scull may be mixed in the slab and pressed together, and defects due to various causes may occur on the surface of the slab, such as foreign material indentation.

As described above, if a defect occurs in the cast steel, the quality of the cast steel may be deteriorated. In order to prevent the quality of the cast steel from being deteriorated, the slab may be subjected to edge trimming and subsequent rolling. The treatment step can be carried out. However, in order to efficiently perform the subsequent process, accurate information about the size and location of defects generated in the cast steel is required.

Conventionally, this has been visually determined, or the subsequent process has been carried out with an arbitrary value obtained statistically, and therefore, it is difficult to expect an efficient subsequent process.

The present invention is to provide a cast body lower surface inspection system capable of inspecting the lower surface of the cast steel to detect defects present in the cast steel.

The present invention also provides a method for inspecting the lower surface of the cast steel that can detect the defects present in the cast steel by inspecting the lower surface of the cast steel.

According to the slab lower surface inspection method according to an embodiment of the present invention, in the thin plate casting process, the original photographing step of generating the original image of the surface of the cast by photographing the cast produced by the casting roll; A defect detection step of detecting a defect existing on the surface of the cast steel in the original image by performing image processing on the original image; A defect size measuring step of measuring a size of the detected defect; And a defect type determination step of determining the type of the detected defect.

The defect detecting step may include a boundary for detecting a widthwise edge of the slab by extracting a pixel having a change in gray-level of the pixel greater than or equal to a preset range from among pixels constituting the original image. Detection process; And a color value comparison process of detecting a defect present on a surface of the cast by extracting a pixel whose color value RGB of the pixel is out of a preset range among pixels constituting the original image. Can be.

The defect size measuring step may include: a horizontal length measuring process of measuring the number of pixels the defect has in the width direction of the cast piece to generate a horizontal length of the coupling; And a longitudinal length measuring process of generating the longitudinal length of the defect by measuring the number of pixels the defect has in the advancing direction of the slab and the tension applied to the slab.

Wherein the horizontal length measurement process,

Lx: the width of the defect, Ax: the number of pixels in the width direction of the defect, S: the number of pixels in the width direction of the cast, and W: the width in the width direction of the cast.

Wherein the vertical length measurement process,

Ly: length of defect, Ax: number of pixels in the width direction of the defect, Ay: number of pixels in the direction of progress of the defect, Lx: width of the defect, k: proportional constant.

Here, in the longitudinal length measurement process, the value of k when the magnitude of tension applied to the slab is maximum and minimum is set in advance, and the proportional constant k is determined by using linear interpolation for the other magnitude of tension. Can be.

In the determining of the defect type, the detected type of defect may be determined using an artificial neural network (ANN).

Here, the defect type determination step may generate an alarm when the defect is a periodic defect caused by a surface defect of the casting roll.

In accordance with another aspect of the present invention, there is provided an apparatus for inspecting a lower surface of a cast steel, the sheet casting process comprising: a camera for photographing a cast produced by a casting roll to generate an original image of the surface of the cast; An image processing apparatus for detecting defects from the original image by performing image processing on the original image; And a type discrimination apparatus for determining the type of the detected defect.

Herein, the image processing apparatus is configured to detect a widthwise edge of the slab by extracting a pixel having a change in gray-level of the pixel greater than or equal to a preset range from among pixels constituting the original image. Detection unit; A defect detection unit for extracting a pixel whose color value (RGB value) of the pixel constituting the original image is out of a predetermined range and detecting a defect present on the surface of the cast steel; And a size measuring unit configured to measure the size of the defect using the number of pixels in the width direction of the slab, the number of pixels in the traveling direction of the slab, and the tension applied to the slab. have.

Wherein the size measuring unit,

Lx: the width of the defect, Ax: the number of pixels in the width direction of the defect, S: the number of pixels in the width direction of the slab, W: the length in the width direction of the detected defect, using the width direction length of the cast steel,

Ly: Length of defect, Ax: Number of pixels in width of defect, Ay: Number of pixels in direction of defect, Lx: Width of defect, k: Length of travel direction of detected defect using proportional constant Can be measured.

In this case, the proportional constant k may be set in advance when the magnitude of the tension applied to the slab is maximum and minimum, and the magnitude of the other tension may be determined using linear interpolation.

Here, the type discriminating apparatus may determine the type of the detected defect by using an artificial neural network (ANN).

Here, the type determining device may generate an alarm when the defect is determined to be a periodic defect caused by a surface defect of the casting roll.

According to the cast steel surface inspection system and the cast steel surface inspection method of this invention, the performance of surface defect detection with respect to a cast steel can be improved significantly. In particular, the present invention can easily calculate the longitudinal direction of the advancing direction of a defect by using a proportional constant.

According to the cast steel surface inspection system and the cast steel surface inspection method of the present invention, it is possible to detect the periodic defects caused by the defects occurring on the surface of the casting roll, it is possible to protect additional damage of the casting roll.

1 is a perspective view showing a general twin roll sheet casting device.
Figure 2 is a block diagram showing the lower surface inspection apparatus according to an embodiment of the present invention.
Figure 3 is a flow chart illustrating a method for inspecting the lower surface of the cast steel according to another embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention. In the following detailed description of the preferred embodiments of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. In the drawings, like reference numerals are used throughout the drawings.

In addition, in the entire specification, when a part is referred to as being 'connected' to another part, it may be referred to as 'indirectly connected' not only with 'directly connected' . Also, to "include" an element means that it may include other elements, rather than excluding other elements, unless specifically stated otherwise.

1 is a perspective view showing a general twin roll sheet casting device.

In the twin-roll type sheet casting process, molten steel is supplied to the space formed by a pair of casting rolls and a pair of edge dams to be cooled by using a molten steel supply nozzle to maintain a constant height of the molten steel pool and face the casting rolls. It is a method of continuously producing a thin plate directly from molten steel by rotating the solidified shell (shell) formed on the surface of the casting roll is coalesced at the proximity between the casting rolls to form a cast.

Referring to FIG. 1, in the continuous casting process of a thin plate, a molten steel 5 is supplied between a pair of casting rolls 1 which are cooled by water through a nozzle 2 connected to a tundish, and the casting roll 1 is rotated. As a result, a pair of solidification shells produced on the surface of the casting roll 1 are coalesced at a close point to continuously manufacture the slabs 4 of the thin plate.

Both side portions of the casting roll 1 are provided with edge dams 3 made of refractory material in order to prevent leakage of the molten steel 5 and to limit the slab 4 to a predetermined width, and are supplied from the nozzle 2. The molten steel 5 maintains the hot water surface at a constant height while forming a molten steel pool by the casting roll 1 and the edge dam 3.

The edge dam 3 is in close contact with the side of the casting roll serves to prevent the molten steel penetrates between the side of the casting roll and the edge dam.

In this case, the temperature of the molten steel decreases during the sheet casting process, such that a skull may occur. When the skull is mixed between the side of the casting roll 1 and the edge dam 3, damage may occur to the edge dam 3, and the edge quality of the cast may deteriorate, which may lead to a serious problem of deterioration of merchandise.

Figure 2 is a block diagram showing a cast steel lower surface inspection system according to an embodiment of the present invention.

Referring to FIG. 2, the slab lower surface inspection system according to an exemplary embodiment of the present invention may include a camera 10, an image processing apparatus 20, and a type discrimination apparatus 30.

Hereinafter, with reference to Figure 2 will be described cast slab bottom inspection system according to an embodiment of the present invention.

The camera 10 may photograph the cast steel produced by the casting roll to generate an original image of the surface of the cast steel. The camera 10 may photograph the cast steel in real time, and the camera 10 may be installed to photograph the entire width of the cast steel in consideration of the imaging angle of the camera 10.

The camera 10 further includes a separate illumination device 7 to irradiate light onto the surface of the cast steel, and an image of the surface of the cast steel using light emitted by the lighting device 7. You can shoot.

Since the cast steel is produced directly from the molten metal at high temperature, self-luminous light can be emitted. In particular, the cast steel can emit a lot of infrared rays and ultraviolet rays in addition to visible light. Therefore, when shooting with the general camera 10, light bleeding may occur, the surface defects of the cast steel may not appear well. To prevent this, the camera 10 may include an infrared filter and an ultraviolet filter to remove other light information other than visible light. In addition, in order to more clearly inspect the surface of the cast steel, the cast can be irradiated with a lighting device 7 having a 6,000k color temperature that the camera 10 can best accept.

In addition, the camera 10 may be photographed by the aperture value is very tight (f5 or more) in order to obtain an accurate surface image of the surface of the light-emitting slab, the image blurring due to the high-speed movement of the cast steel In order to prevent the shutter speed, the shutter speed may be taken at a speed of at least 1/2000 second or more.

In addition, a dual cooling system may be provided to protect the camera 10 and the lighting device 7 from the high temperature of the cast steel.

The image processing apparatus 20 may detect a defect from the original image by performing image processing on the original image. The image processing apparatus 20 may include a boundary detector, a defect detector, and a size measurer.

The edge detector may detect a widthwise edge of the slab by extracting a pixel having a change in gray-level of the pixel greater than or equal to a preset range from among pixels constituting the original image.

Regarding the detection of the widthwise boundary of the slab, techniques commonly used for edge detection in an image processing field may be utilized. For example, a Sobel filter, a Laplacian filter, a Roberts filter, a Prewitt filter, or the like may be used.

The gray level is a numerical value representing the degree of contrast of the pixel and may be generally expressed as 8 bits. That is, it may be expressed as a value between 0 and 255 depending on the degree of contrast between white and black, where black is 0 and white is 255, so that the greater the value of the gray degree, the lighter the darker the value of the gray level is. Can be.

In general, since the amount of change in contrast is larger than the surrounding pixels at the edge, the boundary may be found by extracting a portion where the amount of change in gray is greater than that of the surrounding pixel.

In detail, an amount of change of gray level corresponding to the boundary may be set in advance, and then the boundary may be set to a pixel having a change amount of gray degree larger than the amount of change of gray level set in the original image.

By detecting the widthwise boundary of the cast steel, it is possible to determine where the defect exists on the surface of the cast steel, and the type of the defect can be determined using the information about the position of the defect.

The defect detection unit may detect a defect present on the surface of the cast by extracting a pixel in which the color value (RGB value) of the pixel is out of a preset range among pixels constituting the original image.

Defects that occur in the slab may be different in brightness in comparison with the surface of the slab.

Specifically, a defect called scull appearing on the surface of the cast steel is darker than the surroundings, and a defect called white spot caused by the indentation of a foreign material appears brighter than the surroundings. Have

Accordingly, the defect detection unit may detect the defect in such a manner as to detect pixels belonging to a range of color values of the defect such as the skull or the white point out of a color value of the pixel.

More specifically, in order to detect the defect, first, an R value, ie, a red color value, among the color values of the pixel may be used. As described above, since the cast steel is formed from high temperature molten steel, the cast steel mainly contains a lot of red light. Therefore, after determining whether the color value of red corresponds to the defect, first, if it is determined that the defect is determined, the occurrence of the defect may be checked in consideration of the color values of green and blue. .

The defect detection unit may perform defect detection by sampling a portion of pixels without performing defect detection on the entire original image. When a defect is detected in some of the sampled pixels, the defect may be extracted by performing defect detection on a pixel located near the pixel where the defect is detected.

According to the above method, the computation process in the defect detection unit can be reduced, so that defect detection can be performed more quickly.

The size measuring unit may measure the size of the defect by using the number of pixels having the detected defect in the width direction of the cast steel, the number of pixels having the traveling direction of the cast steel, and the tension applied to the cast steel. .

The size of the defect can be measured through the length and width of the defect. Here, the horizontal length means the length of the slab in the width direction, and the vertical length means the length of the slab in the traveling direction. The traveling direction of the cast steel is indicated by the arrow of FIG.

The number of pixels the defect has in the width direction of the slab and the number of pixels in the advancing direction of the cast may be extracted from the original image. The defect detector may detect a defect in the original image, and thus may easily extract the number of pixels from the detected defect.

However, the curvature of the slab may vary depending on the magnitude of the tension applied to the slab. Referring to FIG. 1, the cast steel may be bent much like a solid line, or may be bent little like the dotted line. The vertical length of the defect photographed by the camera 10 may vary depending on the degree of warpage of the cast steel.

Therefore, it is necessary to correct the longitudinal length of the defect in the original image in consideration of the degree of warpage of the slab. A separate proportional constant may be used to correct the vertical length of the defect.

Here, the size measuring unit,

Lx: the width of the defect, Ax: the number of pixels in the width direction of the defect S: the number of pixels in the width direction of the cast, W: the length of the detected width in the width direction of the cast is measured,

Ly: longitudinal length of the defect, Ax: number of pixels in the width direction of the defect Ay: number of pixels in the direction of progression of the defect Lx: length of the width direction of the defect k: measuring the length of the detected direction of the defect using the proportionality constant can do.

Regarding the width Lx of the defect, W / S is a value obtained by dividing the width direction length of the cast steel by the number of pixels in the width direction of the cast steel. That is, the width direction length of one pixel. Therefore, when (W / S) is multiplied by the number Ax of the width pixels of the defect, the width direction length of the defect, that is, the transverse length of the defect can be obtained. Here, the W may be 1340mm.

Regarding the length L y of the advancing direction of the defect, the length Lx / Ax is equal to the width W / S, that is, the width direction length of one pixel. In general, since the pixels have a square shape, multiplying (Lx / Ax) by the number (Ay) of the direction pixels of the defect direction can obtain the length of the defect direction. However, since the slab may be bent in the advancing direction, as shown in FIG. 2, the vertical length of the defect in the original image may be different from the vertical length of the actual defect. In order to correct such a difference, a proportionality constant such as k of the above equation may be used.

In this case, the proportional constant k may be set in advance when the magnitude of the tension applied to the slab is maximum and minimum, and the magnitude of the other tension may be determined using linear interpolation.

Interpolation is a method of estimating a function value at a point where the rest of the function is not known when the entire function form of a continuous function is not known and only a function value at a specific position is known. The linear interpolation method is the simplest of the interpolation methods. It is a method of estimating a function value assuming that a function to be known is a function of a straight line.

Specifically, the value of k is set to 1.2 when the magnitude of the tension is maximum (f_max) and 1 when the minimum is f_min, and k (f_x) = ((1.2-1) / for other magnitudes of tension. The k value can be obtained using the formula (f_max-f_min)) * f_x. Here, the k value can be obtained by inputting the magnitude of the tension instead of f_x.

The type judging device 30 can determine the type of the detected defect.

Here, the type determining device 30 may determine the type of the detected defect by using an artificial neural network (ANN).

Here, the type determining device 30 may generate an alarm when the detected type of the defect is a periodic defect generated by a defect of the casting roll.

The type discrimination apparatus 30 can determine the type of the defect by inputting the distance from the boundary of the defect in the slab, the size of the defect, the color information of the defect, whether the defect is periodic or not. For example, the skull is one of the types of defects. The skull is generated near the edge of the slab, that is, near the boundary, appears darker than the surroundings, and since there is no periodicity, it is possible to determine whether the skull is the skull.

The periodicity of the defects may mean that defects repeatedly appear at regular intervals at a constant position in the width direction of the cast steel. Repeated appearance of defects as described above may mean that there is damage to the surface of the casting roll 1 to produce the cast.

Therefore, when it is determined that the periodic defect has occurred, it is possible to alert the worker that the casting roll 1 is damaged by generating an alarm, and to take measures for damage of the casting roll 1. .

The type discrimination apparatus 30 may be equipped with a fuzzy control method, a look-up table, etc., in addition to the artificial neural network (ANN), which is widely used in artificial intelligence and data mining. It is possible to determine the type of the defect from the information on the defect input by utilizing.

Figure 3 is a flow chart showing a cast slab surface inspection method according to an embodiment of the present invention.

Referring to Figure 3, according to an embodiment of the present invention, the lower surface inspection method, the original photographing step (S10), defect detection step (S20), defect size measurement step (S30) and defect type determination step (S40) It may include.

Hereinafter, a method for inspecting the lower surface of a cast steel according to an embodiment of the present invention will be described with reference to FIG. 3.

Original photographing step (S10), by photographing the cast produced by the casting roll can generate an original image of the surface of the cast.

In the original photographing step S10, after irradiating light to the cast steel, the light reflected from the cast steel may be photographed using a camera. Here, the camera may be photographed in real time, and may be installed to photograph the entire width of the cast in consideration of the imaging angle of the camera.

Regarding the camera, as described above in the lower slab inspection system, a detailed description thereof will be omitted.

Defect detection step (S20), through the image processing for the original image can detect a defect present on the surface of the slab in the original image. Here, the defect detection step S20 may further include a boundary detection process S21 and a color value comparison process S22.

The boundary detection process S21 may detect a widthwise boundary of the slab by extracting a pixel having a change in gray level of the pixel greater than or equal to a preset range among pixels constituting the original image.

In the boundary detection process S21, a technique generally used for edge detection in an image processing field may be used, for example, a Sobel filter, a Laplacian filter, a Roberts filter, a sweet filter, and the like.

In general, since the change amount of contrast is larger than the surrounding pixels at the boundary, the boundary can be detected by extracting a portion where the change in gray level is larger than the surrounding pixels.

Here, only the boundary may be extracted by binarization by setting the value to 1 for the part extracted as the boundary and 0 for the remaining part.

In the color value comparison process (S22), among the pixels constituting the original image, a pixel in which the color value of the pixel is out of a preset range may be extracted to detect a defect present on the surface of the cast.

Defects occurring in the cast steel, there may be a difference in terms of color when compared to the surface of the cast steel, the difference in color may vary depending on the type of the defect.

Specifically, the defect called the skull (skull) is characterized by appearing darker than the surrounding, the defect called the white spot (white spot) is characterized by appearing brighter than the surrounding. Accordingly, in the color value comparison process (S22), the color value of the pixel is out of a preset color value, and the pixels are detected by detecting pixels belonging to a range of color values of the defect such as the skull or the white point. Can be.

More specifically, in order to detect the defect, the color value comparison process S22 may first detect a defect using an R value, that is, a red color value, among the color values of the pixel. The cast steel is formed from high temperature molten steel and mainly contains a lot of red-based light. Therefore, it is first determined whether the defect corresponds to the defect using the red color value, and if it is determined as a defect, the green and blue color values may be determined again to finally determine whether the defect is a defect. have. As described above, if a process of first determining whether a defect is performed using a red color value is used, the green and blue may not be performed on all pixels. Therefore, the image processing for determining the defect can be performed more quickly and simply.

In addition, the color value comparison process (S22) may perform defect detection by sampling a portion of pixels without first performing defect detection on the entire original image. If a defect is detected in some of the sampled pixels, the defect may be extracted by performing defect detection on the neighboring pixels of the pixel from which the defect is detected. Likewise, the above method is to perform the image processing process more quickly and simply.

In the defect size measuring step S30, the size of the detected defect may be measured.

Here, the defect size measuring step S30 may further include a horizontal length measuring process S31 and a vertical length measuring process S32.

The horizontal length means the length of the slab in the width direction, and the vertical length means the length of the slab in the advancing direction.

In the horizontal length measuring process (S31), the defect may measure the number of pixels that the defect has in the width direction of the slab to generate the width of the defect.

The number of pixels the defect has in the width direction of the slab, that is, the number of pixels the defect has in the horizontal direction, can be easily extracted in the color value comparison process (S22). The number of pixels in the transverse direction of the defect may be measured from the defect extracted in the color value comparison process S22.

Wherein the horizontal length measurement process (S31),

Lx: the width of the defect, Ax: the number of pixels in the width direction of the defect S: the number of pixels in the width direction of the cast, W: the length in the width direction of the cast.

With respect to the above-described equation and the measurement of the width of the defect, as described above, a detailed description thereof will be omitted.

In the vertical length measuring process S32, the length of the defect may be generated by measuring the number of pixels that the defect has in the advancing direction of the slab and the magnitude of the tension applied to the slab.

Wherein the vertical length measurement process (S32),

Ly: length of the defect, Ax: number of pixels in the width direction of the defect Ay: number of pixels in the direction of progression of the defect Lx: width of the defect k: proportional constant can be used.

With respect to the above-mentioned formula and the measurement of the longitudinal length of the defect, as described above, detailed description thereof will be omitted.

In the defect type determination step S40, the type of the detected defect may be determined.

Here, in the defect type discrimination step S40, rectification of the detected defect may be determined using an artificial neural network (ANN).

Here, the defect type discrimination step (S40) may generate an alarm when the defect is a periodic defect generated by a surface defect of the casting roll.

In the defect type discrimination step (S40), the type of the defect may be determined by inputting a horizontal position of the defect in the cast steel, color information of the defect, whether the defect is periodic or not. For example, the skull may be generated near the edge of the slab, that is, near the boundary, appear darker than the periphery, and have no periodicity. Therefore, the skull may be used to determine whether the detected defect is a skull. .

The defect type determination step (S40), in addition to the artificial neural network that is widely used in the field of artificial intelligence, data mining, to a defect input by using a fuzzy control method, a look-up table, etc. It is possible to determine the type of the defect from the information on.

The present invention is not limited by the above-described embodiment and the accompanying drawings. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims, .

1: casting roll 2: molten steel supply nozzle
3: edge dam 4: sheet (cast)
5: molten steel 6: pinch roll
7: lighting equipment
10: camera 20: image processing device
30: Type discrimination device
S10: Original shooting step
S20: Defect detection step S21: Boundary detection process
S22: Color value comparison process
S30: Defect size measuring step S31: Width measuring process
S32: length measurement process
S40: Defect type determination step

Claims (14)

An original photographing step of photographing the cast produced by the casting roll to generate an original image of the surface of the cast;
A defect detection step of detecting a defect existing on the surface of the slab in the original image by performing image processing on the original image;
A defect size measuring step of measuring a size of the detected defect; And
Cast iron surface inspection method in a thin plate casting process comprising a defect type discrimination step of determining the type of the detected defect.
The method of claim 1, wherein the defect detection step is
A boundary detection process of detecting a widthwise edge of the slab by extracting a pixel having a change in gray-level of the pixel more than a preset range from pixels constituting the original image; And
Among the pixels constituting the original image, the main surface inspection method comprising a color value comparison process for detecting a defect present on the surface of the cast by extracting a pixel whose color value (RGB) outside the predetermined range of the pixel; .
The method of claim 1, wherein the defect size measuring step
A horizontal length measuring process of measuring the number of pixels the defect has in the width direction of the slab to generate a horizontal length of the coupling; And
And a longitudinal length measuring process of measuring the number of pixels in the advancing direction of the slab and the magnitude of the tension applied to the slab to generate the longitudinal length of the defect.
The method of claim 3, wherein the horizontal length measuring process

Lx: Width of defect, Ax: Number of pixels in width of defect
S: Number of pixels in the width direction of the cast steel, W: Width in the width direction of the cast steel
Cast steel surface measurement method using the.
The method of claim 3, wherein the length measurement process is performed.

Ly: Length of the defect, Ax: Number of pixels in the width of the defect
Ay: Number of pixels in the direction of the defect Lx: Width of the defect
k: proportional constant
Cast steel surface measurement method using the.
The method of claim 5, wherein the length measurement process
The k value when the magnitude of tension applied to the slab is maximum and minimum is set in advance, and the proportional constant k is determined using linear interpolation for the magnitude of the other tension.
According to claim 1, wherein the defect type determination step
A flat surface inspection method for determining the type of the detected defect using an artificial neural network (ANN).
The method of claim 7, wherein the defect type determination step
And an alarm is generated when the defect is a periodic defect caused by a surface defect of the casting roll.
A camera for photographing the cast produced by the casting roll to generate an original image of the surface of the cast;
An image processing apparatus for detecting defects from the original image by performing image processing on the original image; And
A cast steel surface inspection system in a thin plate casting process comprising a type discrimination apparatus for discriminating the type of the detected defect.
The image processing apparatus of claim 9, wherein the image processing apparatus is
A boundary detector for extracting a pixel whose change in gray-level of the pixel is greater than or equal to a preset range among pixels constituting the original image, to detect a width direction edge of the slab;
A defect detection unit for extracting a pixel whose color value (RGB value) of the pixel constituting the original image is out of a predetermined range and detecting a defect present on the surface of the cast steel; And
A slab including a size measuring unit for measuring the size of the defect using the number of pixels having the detected defect in the width direction of the cast steel, the number of pixels having the traveling direction of the cast steel, and the tension applied to the cast steel; Back inspection system.
The method of claim 10, wherein the size measuring unit

Lx: Width of defect, Ax: Number of pixels in width of defect
S: Number of pixels in the width direction of the cast steel, W: Width in the width direction of the cast steel
Measure the length in the width direction of the detected defect using,

Ly: Length of the defect, Ax: Number of pixels in the width of the defect
Ay: Number of pixels in the direction of the defect Lx: Width of the defect
k: proportional constant
Slab surface inspection system for measuring the length of the direction of travel of the detected defect using.
The method of claim 11, wherein the proportional constant k is
And a corresponding value is set in advance when the magnitude of the tension applied to the slab is maximum and minimum, and the magnitude of the other tension is determined by linear interpolation.
13. The apparatus of claim 12, wherein the type discriminating device
A main surface inspection system for determining the type of the detected defect using an artificial neural network (ANN).
The method of claim 13, wherein the type discrimination apparatus
A cast iron surface inspection system for generating an alarm when the defect is determined to be a periodic defect caused by a surface defect of the casting roll.

Images