KR20230158301A - Method for identify of slag, method for slag skimming and apparatus for skimming slag - Google Patents

Abstract

The method for identifying slag according to an embodiment of the present invention is a process of acquiring a photographed image (I _s ) including chromatic colors by photographing a container charged with molten material with slag floating on the surface, on the photographed image (I _s ) The process of extracting the region of interest (A _i ) containing slag, which is the object to be excluded, and generating the region of interest setting image (I _Ai ) showing the extracted region of interest (A _i ), the region of interest setting image including chromatic colors ( A process of generating an object identification image (I _d ) by converting the color of a pixel with a brightness below the identification standard brightness among a plurality of pixels included in the area of interest (A _i ) of I Ai ) into an achromatic first _color , and the object It may include a process of identifying a pixel displayed in the first color on the identification image (I _d ) as slag, which is an exclusion target.
Therefore, according to embodiments of the present invention, the identification accuracy for slag can be improved during the operation of removing slag from inside the vessel. Accordingly, the location of the slag inside the container can be determined more accurately, and thus the slag inside the container can be effectively excluded. In addition, the amount of slag remaining inside the container can be quantitatively determined, making it possible to more accurately predict the completion of slag removal.

Description

Method for identifying slag, method for excluding slag, and apparatus for skimming slag {Method for identify of slag, method for slag skimming and apparatus for skimming slag}

The present invention relates to a slag identification method, a slag drainage method, and a slag drainage device. It relates to a slag identification method, a slag drainage method, and a slag drainage device that can improve the identification accuracy of slag.

When the converter refining is completed, the molten steel inside the converter is tapped using a ladle. At this time, the slag floating on the upper part of the molten steel is tapped into the ladle together with the molten steel. Slag deteriorates the quality of steel. Accordingly, when molten steel is tapped into a ladle, slag is excluded before moving the molten steel to the casting device.

Meanwhile, in order to effectively drain slag inside the ladle, it is important to adjust the position of the drainer. In other words, it is necessary to adjust the horizontal position of the slag so that the slag is in contact with slag rather than molten steel inside the ladle.

In order to adjust the position of the slag, it is necessary to determine the location of the slag inside the ladle. For this purpose, the worker checks the inside of the ladle with the naked eye to determine the location of the slag. However, this method has problems in that the accuracy of controlling the slag position and scavenger position is low, and workers are exposed to danger.

To solve this problem, a technology was applied to photograph the inside of the ladle using a camera and determine the location of the slag inside the ladle using the captured image. Molten steel and slag have different temperatures and therefore different emissivity. Accordingly, images in which at least one of color and saturation differ depending on the difference in emissivity are obtained.

However, although the molten steel and slag have different temperatures, the difference is not large and can be displayed with similar color and saturation in the image. Therefore, it is difficult to identify slag on the image. In particular, the slag floating on top of the molten steel for stainless steel has a high temperature and exists in a liquid state. In addition, slag in the liquid state has a high emissivity due to its high temperature, and its emissivity has a small difference from the emissivity of molten steel. Therefore, in the case of images taken of a ladle filled with molten steel for stainless steel, it is more difficult to distinguish molten steel and liquid slag.

Additionally, the position of the slag on the surface of the molten steel may vary depending on the tilt angle of the ladle, the amount of molten steel charged inside the ladle, the internal volume of the ladle according to the number of times the ladle is used, and the amount of residual slag. Therefore, the position of the slag on the surface of the molten steel cannot be identified by fixing it to a specific position.

Korean registered patent 10-1412550

The present invention provides a slag identification method, a slag removal method, and a slag removal device that can improve the identification accuracy of slag during the operation of removing slag from inside a container.

The present invention provides a slag identification method, a slag drainage method, and a slag drainage device that can easily identify slag even if the position of the slag inside the container changes.

The present invention provides a slag identification method, a slag drainage method, and a slag drainage device that can improve identification accuracy for high-temperature slag existing in a liquid phase.

A method for identifying slag according to an embodiment of the present invention includes the steps of acquiring a photographed image (I _s ) including chromatic colors by photographing a container charged with molten material with slag floating on the surface; A process of extracting a region of interest (A _i ) containing slag, which is an object to be excluded, from the captured image (I _s ), and generating a region of interest setting image (I _Ai ) displaying the extracted region of interest (A _i ). ; Among the plurality of pixels included in the region of interest (A _i ) of the region of interest setting image (I _Ai ) including chromatic colors, the color of the pixel having a brightness below the identification standard brightness is converted to an achromatic first color, and the object A process of generating an identification image (I _d ); and a process of identifying a pixel displayed in the first color on the object identification image I _d as slag, which is an exclusion object.

The process of extracting the region of interest (A _i ) includes distinguishing an inner area of the inner wall surface of the container and other areas on the captured image (I _s ); and a process of setting the inner area of the divided inner wall surface of the container as the area of interest (A _i ).

The process of distinguishing the inner area of the inner wall of the container from other areas on the captured image (I _s ) and the process of setting the inner area of the inner wall of the container as the area of interest (A _i ) are segmentation using artificial intelligence ( This can be done using segmentation techniques.

The process of generating the object identification image (I _d ) includes assigning a first identification value to each of a plurality of pixels included in the captured image (I _s ) according to the brightness of each pixel; And a first converted image (I) that is a gray scale image by converting the color of each of the plurality of pixels of the captured image (I _s ) into one of the achromatic colors according to the first identification value assigned to each. It may include a process of generating _c1 ).

The process of generating the object identification image (I _d ) includes setting a region of interest (A i ) in the first converted image (I _c1 ) using the region of interest setting image (I _{Ai )} ; Calculating an identification reference value (V _ds ) using a first identification value for all pixels included in the first converted image (I _c1 ); And using the identification reference value (V _ds ), the first converted image (I _c1 ) is converted into an image that appears in two types of achromatic colors, including the first color and a second color that is an achromatic color distinct from the first color. A process of converting to generate a second converted image (I _c2 ) may be included, and the identification reference value (V _ds ) may be used as the identification reference brightness.

In calculating the identification reference value (V _ds ), the average of the first identification values for all pixels included in the first converted image (I _c1 ) (PV _t-mean ); The maximum value (PV _Ai-max ) among the first identification values of pixels in the region of interest (A _i ) of the first converted image (I _c1 ); and the average (PV _Ai-mean ) of the first identification values of pixels in the region of interest (A _i ) of the first converted image (I _c1 ).

In calculating the identification reference value (V _ds ), it can be calculated using the formula below.

[formula]

(C: constant)

The process of generating the second converted image (I _c2 ) includes, in each of the plurality of pixels included in the first converted image (I _c1 ), a first identification value being applied to a pixel whose first identification value is less than or equal to the identification reference value (V _ds ). A process of assigning a second identification value of 1 and assigning a second value, which is a second identification value different from the first value, to a pixel whose first identification value exceeds the identification reference value (V _ds ); In each of the plurality of pixels included in the first converted image (I _c1 ), a pixel whose assigned second identification value is a first value is displayed with the first color, and a pixel whose assigned second identification value is a second value is displayed with the first color. It may include a process of converting the first converted image (I _c1 ) by displaying it in the second color.

The process of generating the object identification image (I _d ) is the same as the pixel marked with the first color on the second converted image (I _c2 ) among the plurality of pixels included in the region of interest setting image (I _Ai ). It may include a process of converting the color of the pixel at the location to the first color.

The first color is black, the second color is white, the second converted image (I _c2 ) is a black-and-white image in which a plurality of pixels are displayed in black or white, and the object identification image ( Among the pixels in the area of interest (A _{i )} of the second converted image (I _c2 ), pixels at the same position as the pixels displayed in black in the area of interest (A _i ) of the second converted image (I c2 ) are displayed in black, The process of identifying slag, which is an exclusion object, may include identifying a pixel displayed in black in the object identification image I _d as slag, which is an exclusion object.

A method for discharging slag according to an embodiment of the present invention includes the process of identifying slag on the object identification image (I _d ); A process of adjusting the position of the slag so that the slag is positioned on the identified object identification image (I _d ); and a process of discharging slag out of the container by moving the exhauster toward the outside of the container.

A process of determining whether slag drainage is complete using slag on the identified object identification image (I _d ), wherein the process of determining whether slag drainage is complete includes adjusting the position of the slag exhauster. It can be done before.

The process of determining whether the slag exclusion is complete includes adding up the area (B _c ) of pixels identified as slag on the object identification image (I _d ); And a process of determining that slag exclusion is complete when the added area (B _c ) is less than or equal to a preset reference area (B _s ). If it is determined that slag exclusion is complete in the determination process, the tilted container is It may include an erection process.

The melt may include molten steel, and the container may include a ladle.

The molten steel may be molten steel for manufacturing stainless steel.

The slag drainage device according to an embodiment of the present invention,

a drainage unit capable of moving into a container charged with molten material with slag floating on the surface; a tilting unit on which the container can be seated and which is installed to tilt the container; a photographing unit installed on the outside of the container to obtain images according to emissivity; A region of interest (A _i ) containing slag, which is an exclusion object, is extracted from the photographed image (I _s ) acquired by the photographing unit, and a brightness of a plurality of pixels included in the region of interest (A _i ) is equal to or less than the brightness standard for identification. An identification unit that generates an object identification image (I _d ) by converting the color of the pixel having to any one color of the achromatic color series; and a control unit that controls the operation of at least one of the distribution unit and the tilt mover using the object identification image (I _d ).

The photographed image (I _s ) acquired by the photographing unit includes a chromatic color, and the identification unit positions the slag to be excluded from the inner area of the inner wall surface of the container using the exhaust portion on the photographed image (I _s ). An extraction unit for extracting a region of interest (A _i ) and displaying the extracted region of interest (A _i ) on the captured image (I _s ) to generate a region of interest setting image (I _Ai ) including chromatic colors; And a first converter that converts each of the plurality of pixels included in the captured image (I _s ) into one color among achromatic colors according to the brightness of each pixel, thereby generating a first converted image that is a gray scale image. and a conversion unit, wherein the first conversion unit may set a region of interest ₍ A i ) in the first converted image (I _c1 ) using the region of interest setting image (I _Ai ).

The identification unit determines the average brightness of all pixels included in the first converted image (I _c1 ) and the maximum brightness among the brightness of pixels included in the region of interest (A _i ) of the first converted image (I _c1 ). , a calculation unit that calculates the identification standard brightness using the brightness average of pixels included in the region of interest (A _i ) of the first converted image (I _c1 ); Each of the plurality of pixels included in the first converted image is converted to one of black and white depending on whether the brightness of each pixel is less than or equal to the identification standard brightness, thereby producing a black-and-white image. 2 A second conversion unit that generates a converted image; And among a plurality of pixels included in the region of interest setting image (I _Ai ), the color of a pixel at the same position as a pixel marked black on the second converted image (I _c2 ) is converted to black to create the object identification image. It may include an identification image generator that generates (I _d ).

The control unit may include a determiner that determines whether slag exclusion is complete using a pixel identified as slag, which is an exclusion object, on the object identification image (I _d ); a first controller that controls the operation of the slag slag so that its position is adjusted when the judge determines that slag slag has not been completely drained; and a second controller that controls the operation of the tilting unit so that the container is upright when the judge determines that the slag has been completely drained.

According to embodiments of the present invention, the identification accuracy for slag can be improved during the operation of removing slag from inside a container. In other words, slag can be easily identified even if the position of the slag on the surface of the melt changes depending on the amount of molten material charged into the container, the volume inside the container, the tilt angle of the container, and the amount of slag remaining inside the container, and the identification accuracy can be improved. It can be improved. In addition, even if the slag has a high emissivity or is in a liquid state, the slag can be easily identified and the identification accuracy can be improved.

Therefore, the location of the slag inside the container can be determined more accurately. Additionally, it is easy to adjust the position of the slag so that it can contact the slag inside the container. Because of this, slag inside the container can be effectively excluded.

In addition, the amount of slag remaining inside the container can be determined, and the timing of completion of slag removal can be more accurately predicted.

Figure 1 is a diagram showing a state in which slag inside a first container is being drained into a second container using a slag draining device according to an embodiment of the present invention.
Figure 2 is a view showing the main portion of the slag drainage device according to an embodiment of the present invention and the interior of the tilted first container.
Figure 3(a) and Figure 3(b) are top views seen from the upper side of the tilted ladle.
Figure 4 (a) and Figure 5 (a) are examples of captured images (I _s ) obtained from the photographing unit of the slag distribution device according to an embodiment of the present invention.
Figures 4(b) and 5(b) show the region of interest setting image (I _Ai ) shown by extracting the region of interest (A _i ) from the captured image (I _s ) using the extraction unit according to an embodiment of the present invention. ) is an example.
Figures 4(c) and 5(c) show a first converted image (I) converted from a captured image (I _s ) to a gray scale image using a first conversion unit according to an embodiment of the present invention. _c1 ) is an example.
4(d) and 5(d) show a second converted image ( _{I c2} ) is an example.
Figures 4(e) and 5(e) are examples of object identification images (I _d ) generated using an identification image generator according to an embodiment of the present invention.
Figure 6 is a flowchart showing a method for discharging slag according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the attached drawings. However, the present invention is not limited to the embodiments disclosed below and will be implemented in various different forms. These embodiments only serve to ensure that the disclosure of the present invention is complete and to those skilled in the art to fully convey the scope of the invention. This is provided to inform you. The drawings may be exaggerated to explain embodiments of the present invention, and like symbols in the drawings refer to like elements.

The present invention relates to a slag identification method that can identify slag that is suspended at the top of a melt and is a target for slag. Additionally, the present invention relates to a slag slag method and a slag slag device for slag slag floating on top of a melt.

The melt may be molten steel for manufacturing a slab or steel. As a more specific example, the melt may be molten steel for manufacturing stainless steel, or may be molten steel that has undergone converter refining.

An example of how to prepare molten steel for manufacturing stainless steel is described below. First, the molten iron tapped from the blast furnace is transferred to HMPS (Hot Metal pre-treatment Station), where it is desiliconized, dephosphorized, and deoiled. Next, the molten iron that has been desiliconized, dephosphorized, and defluorinated is charged into a converter and refined. In other words, converter refining is performed by blowing oxygen into the converter using a lance to lower the carbon (C) content contained in the molten iron in the converter. Accordingly, molten iron whose decarburization has been completed or molten iron whose content of carbon (C) has been adjusted, that is, molten steel, is manufactured. In addition, slag containing reaction by-products generated during converter refining may float on the upper part of the molten steel.

When the converter refining is completed, the molten steel is tapped into another container, such as a ladle, located outside the converter. That is, the converter is tilted to tap or discharge the molten steel into the ladle. At this time, the slag floating on the upper part of the molten steel is tapped or discharged into the ladle together with the molten steel.

Meanwhile, slag deteriorates the quality of steel. Accordingly, when molten steel is tapped into a ladle, slag is removed from the molten steel, that is, excluded, before moving the molten steel to the casting device.

The temperature of the ladle tapped molten steel for manufacturing stainless steel may be 1600°C to 1700°C. Additionally, the slag floating on top of the tapped molten steel may be exposed to the atmosphere. Accordingly, the temperature of slag may be lower than that of molten steel. More specifically, the temperature of slag may be 25°C to 50°C lower than that of molten steel. At this time, the temperature of molten steel for manufacturing stainless steel is as high as 1600°C to 1700°C as described above. Accordingly, the slag at the top of the molten steel has a high temperature and may exist in a liquid state. That is, the slag floating on top of molten steel for manufacturing stainless steel has a higher temperature than the slag floating on top of molten steel for manufacturing other steel materials. For example, it is molten steel for manufacturing carbon steel, and the slag floating on top of molten steel that has been ladle tapped after converter refining has a temperature of 1300°C to 1350°C, and may be in the solid phase or solid particle state due to its low temperature. there is. However, the temperature of molten steel for stainless steel production and ladle taping after completion of converter refining is high at 1600°C to 1700°C, and the temperature of slag floating on top of the molten steel is also high. Accordingly, the slag may be in a liquid state.

The object to be drained using the slag draining device according to the embodiment may be slag floating on the upper part of molten steel for stainless steel production in which converter refining has been completed. That is, the object to be excluded may be molten steel for stainless steel production and slag floating on top of the molten steel that has undergone converter refining.

Additionally, the object to be drained using the slag draining device may be high-temperature slag floating on top of molten steel with a temperature of 1600°C to 1700°C, or slag in a liquid state.

Of course, the object to be excluded is not limited to the slag floating on top of the molten steel for manufacturing stainless steel as in the above-mentioned example, and may be slag floating on the top of the molten steel for manufacturing various steel materials. For example, it may be slag floating on top of molten steel for manufacturing carbon steel.

In addition, the object to be excluded is not limited to the slag floating on the top of the molten steel after converter refining as in the above-mentioned example, and may be slag generated at various refining operation stages. For example, it may be slag generated from refining other than converter refining, such as KR (Kanvara Reactor) refining, RH (Rheinstahl Huttenwerke) refining, etc.

Additionally, the object to be excluded is not limited to slag in a liquid state, and may also be slag in a solid or solid particle state.

Hereinafter, molten steel will be described as an example of molten material with floating slag. In addition, the molten steel is molten steel for manufacturing stainless steel, and the molten steel that has undergone converter refining will be explained as an example. Accordingly, the slag may be molten steel for manufacturing stainless steel and may be slag floating on top of molten steel that has undergone converter refining.

In addition, hereinafter, for convenience of explanation, melt and molten steel will be described by referring to the same reference numeral 'M'.

Figure 1 is a diagram showing a state in which slag inside a first container is being drained into a second container using a slag draining device according to an embodiment of the present invention. Figure 2 is a view showing the main portion of the slag drainage device according to an embodiment of the present invention and the interior of the tilted first container. Here, in Figure 2, in order to show the state of the molten steel and slag charged inside the first tilted container, the tilted portion on which the first container is seated is shown with a dotted line.

Before explaining the slag drainage device according to the embodiment, the first and second containers will first be described with reference to FIGS. 1 and 2.

The first container 1000 may be a container that can accommodate the molten steel (M) tapped from the converter. This first container 1000 has an internal space in which molten steel M can be accommodated, a body 1100 with an opening 1100a on the upper side, and a body 1100 so that it can be seated on the tilting portion 4600. It may include a lifting lug 1200 installed and a tilt axis 1300 connecting the body 1100 and the lifting lug 1200.

The body 1100 may be made of a material containing a refractory material to have heat resistance against high-temperature molten steel (M) and slag (S). More specifically, the body 1100 may include a recess formed by building a steel shell made of metal or alloy and a brick made of a material containing a refractory material on the inner surface of the steel shell.

Referring to FIG. 2, the body 1100 may include a body 1111 having an internal space and a discharge member 1112 extending laterally from the body 1111 to communicate with the internal space of the body 1111. .

The discharge member 1112 may be a means for providing a passage so that the charge inside the body 1111, such as slag (S) or molten steel (M), can be discharged to the outside. This discharge member 1112 may be formed to extend from the outer surface of the upper part of the body 1111 in the width (or diameter) direction of the body 1111. Additionally, the discharge member 1112 may extend not only in the width direction of the body 1111 but also in the upper direction of the body 1111.

Since the body 1100 includes the body 1111 and the discharge member 1112 extending laterally from the body 1111 as described above, the opening 1100a provided on the upper side of the body 1100 is connected to the body 1111. It can be described as including an upper opening (hereinafter, main opening 1111a) and an upper opening of the discharge member 1112 (hereinafter, discharge port 1112a). Here, the main opening 1111a of the body 1111 may be mainly used as a passage through which molten steel (M) and slag (S) are charged into the interior of the body 1100. In addition, the discharge port 1112a of the discharge member 1112 can be mainly used as a passage for discharging the charge inside the body 1100.

The lifting lug 1200 may be provided to protrude outward from the side of the body 1100. Additionally, the lifting lugs 1200 may be provided as a pair, and the pair of lifting lugs 1200 may be installed on both sides of the body 1100 to face each other. And the tilt of the first container 1000 may be centered on the tilt axis 1300 connecting the body 1100 and the pair of lifting lugs 1200.

The first container 1000 as described above may be a ladle in which at least one of molten iron, molten steel, and molten metal is accommodated in the steelmaking technology field. Accordingly, it can be explained that the first container 1000 containing the slag to be drained using the slag draining device 4000 is a ladle. Hereinafter, the first container 1000 will be described as a 'ladle' as an example. In addition, for convenience of explanation, the first container and the ladle will be described by referring to the same reference numeral '1000'.

In addition, the port as the second container 2000, which is located on the lower side of the ladle 1000 and receives and accommodates the slag S, will be described as an example. At this time, for convenience of explanation, the second container and the port will be described by referring to the same reference numeral '2000'.

Hereinafter, a slag drainage device according to an example will be described.

Referring to FIGS. 1 and 2, the slag exhaust device 4000 according to the embodiment includes an exhaust portion 4100 including a slag exhaust device 4100a that is inserted into the ladle 1000 and discharges the slag (S) to the outside. Images acquired by the tilting unit 4600 for tilting the ladle 1000, the photographing unit 4200 for photographing the interior of the ladle 1000, and the photographing unit 4200 to identify the slag (S), which is the object of exclusion. It may include an identification unit 4300 that generates an image (hereinafter referred to as an object identification image (I _d )) in which the exclusion object is distinctively displayed.

In addition, the slag drainage device 4000 includes a control unit 4400 that controls the operation of at least one of the drainage unit 4100 and the tilting unit 4600 using the object identification image (I _d ) generated by the identification unit 4300. may include. In addition, the slag drainage device 4000 may further include a display unit 4500 that displays the generated object identification image I _d on the screen.

The exhaust unit 4100 may include an exhaust unit 4100a that can be inserted into the ladle 1000 so as to contact the slag S, and a mover 4100b that moves the exhaust unit 4100.

The exhaust device 4100a may have a plate shape with a predetermined area. Additionally, the exhaust device 4100a may have a plate shape with a lower area opposite to the mover 4100b larger than the upper part connected to the mover 4100b.

Referring to FIG. 1, the mover 4100b has a main body 4110b connected to an exhaust device 4100a at one end, which is the end of the moving machine 4100b, in the direction in which the ladle 1000 is located, or between the ladle 1000 and the main body 4110b. It may include a traveling member 4120b that moves horizontally in the opposite direction.

The main body 4110b has an arm 4111b connected to one end of the exhaust device 4100a, a support 4112b installed to support the lower part of the arm 4111b, and a support 4112b installed to connect the arm 4111b and the support 4112b. It may include an elevating body 4113b that adjusts the height of the arm 4111b, and a rotating body 4114b installed at the lower part of the support 4112b so as to be rotatable.

The arm 4111b may have a shape that extends in one direction, and an exhaust device 4100a may be connected to one end facing the ladle 1000, and an elevating body 4113b may be connected to the other end, which is the opposite end of the arm 4111b.

The elevating body 4113b is a means capable of moving in the up and down direction and may be installed to connect the arm 4111b and the support 4112b. This elevating body 4113b may be a means including, for example, a hydraulic cylinder or a pneumatic cylinder. For example, the elevating body 4113b is installed to connect between the drive source 41, which provides a lifting and lowering driving force by hydraulic or pneumatic pressure, and the other end of the arm 4111b and the drive source 41, and is raised by the operation of the drive source 41. It may include a driving body 42 that extends (or expands) or descends (or contracts).

The height of the exhaust device (4100a) can be adjusted by the operation of the raising and lowering body (4113b). That is, the arm 4111b is raised and lowered around the connection portion connected to the support 4112b by the operation of the raising and lowering body 4113b, and the exhaust device 4100a is raised and lowered accordingly. In other words, the arm 4111b is raised and lowered by the operation of the raising and lowering body 4113b so that the heights of one end and the other end change. Accordingly, the exhaust device 4100a connected to one end of the arm 4111b rises or falls. For example, when the driving body 42 is raised by the operation of the driving source 41, the arm 4111b operates so that the height of the other end rises and the height of one end decreases, as shown in FIG. 1. Accordingly, the exhaust device 4100a connected to one end of the arm 4111b descends. Conversely, when the driving body 42 is lowered by the operation of the driving source 41, the arm 4111b operates so that the height of the other end decreases and the height of one end increases. Accordingly, the exhaust device 4100a connected to one end of the arm 4111b rises. In this way, the height of the exhaust device 4100a can be adjusted.

The rotating body 4114b may be installed at the lower part of the support 4112b so that it can rotate around its width direction center. And when the rotating body 4114b rotates, the support 4112b, the elevating and lowering body 4113b, the arm 4111b, and the exhaust device 4100a connected thereto may rotate. At this time, the arm 4111b of the main body 4110b can rotate around its other end. That is, the arm 4111b can rotate around the width direction center of the rotating body 4114b. By this rotation, the horizontal position of one end of the arm 4111b can be changed or adjusted. And this rotation of the arm 4111b can be explained as changing the position of one end of the arm 4111b in the width direction of the ladle 1000. Additionally, since the exhaust 4100a is connected to one end of the arm 4111b, the exhaust 4100a can move in the horizontal direction by rotating the arm 4111b. That is, the rotation of the arm 4111b moves the drainer 4100a horizontally, and thus the horizontal position of the drainer 4100a in the width direction of the ladle 1000 can be adjusted.

The traveling member 4120b horizontally moves the main body 4110b to the front, where the ladle 1000 is located, or to the rear, opposite to the ladle 1000. For example, the traveling member 4120b may be a means including a rail on which the rotating body 4114b of the main body 4110b can be mounted and slide. When the traveling member 4120b operates, the main body 4110b mounted on the upper part of the traveling member 4120b and the exhaust device 4100a connected to the arm 4111b of the main body 4110b move forward so as to approach the ladle 1000. It can move horizontally or move horizontally backwards away from the ladle 1000.

The tilting unit 4600 is a means for tilting the ladle 1000. This tilting unit 4600 is, for example, a seating table 4610 on which the lifting lug 1200 of the ladle 1000 can be seated, and a tilting machine that tilts the ladle 1000 seated on the seating table 4610. It may include (4620).

The seating base 4610 may have a shape extending in the vertical direction and may be provided as a pair. The pair of mounting bases 4610 may be arranged in the direction in which the pair of lifting lugs 1200 provided on the ladle 1000 are aligned. In addition, when the ladle 1000 is seated or mounted on the mounting base 4610, a pair of lifting lugs 1200 of the ladle 1000 are seated so as to be mounted on each upper part of the pair of seating bases 4610. You can.

The tilting machine 4620 connects the ladle 1000 and the seating table 4610 and provides rotational force to the tilting arm 4621 and the tilting arm 4621 installed to rotate around the tilting axis 1300. It may include a driving source 4622.

The tilting arm 4621 is installed so that one end is connected to the mounting base 4610, and the tilting arm 4621 can be rotated around its one end connected to the seating base 4610. That is, the tilt arm 4621 is connected to the mounting base 4610, and can be installed on the seating base 4610 so that it can rotate with one end of it as a rotation axis.

The tilt drive source 4622 may include, for example, a rod 4622-1 and a cylinder member 4622-2 that moves the rod 4622-1 forward and backward by hydraulic or pneumatic pressure. By moving the rod 4622-1 of the tilt drive source 4622 forward or backward, the tilt arm 4621 connected to the rod 4622-1 can be rotated. And, as the tilt arm 4621 rotates according to the forward or backward motion of the rod 4622-1, the ladle 1000 rotates, that is, tilts, around the tilt axis 1300. At this time, the rotation direction of the tilt arm and can be adjusted depending on the forward or backward movement of the rod 4622-1. Accordingly, the tilt direction of the ladle 1000 can be adjusted. Additionally, the rotation amount of the tilt arm 2210 can be adjusted depending on the forward or backward movement amount of the rod 4622-1, and the tilt angle of the ladle 1000 can be adjusted accordingly.

The tilting unit 4600 is not limited to the above-described example, and various means for tilting the ladle 1000 may be applied.

Figure 3(a) and Figure 3(b) are top views seen from the upper side of the tilted ladle. Here, Figure 3(a) and Figure 3(b) are diagrams to explain that the positions of slag on the molten steel surface are different.

Meanwhile, the position of the slag (S) floating on the upper surface of the molten steel (M), that is, the molten steel surface, is determined by the tilt angle of the ladle (1000), the amount of molten steel charged inside the ladle (1000), and the number of uses of the ladle (1000). It may vary depending on the internal volume and amount of residual slag.

In addition, when the slag (S) floats on the surface of the molten steel (M), the slag (S) may cover the entire upper surface of the molten steel, but as shown in Figures 3 (a) and 3 (b), the molten steel (S) M) It may be distributed in the form of an island so that a portion of the hot water surface is exposed.

And at this time, the location of the slag on the molten steel (M) surface may be different, as shown in Figure 3 (a) and Figure 3 (b). That is, as described above, slag is formed on the molten metal surface according to the tilt angle of the ladle 1000, the amount of molten steel (M) charged inside the ladle 1000, the internal volume of the ladle 1000, and the amount of residual slag (S). The location of (S) may be different.

Additionally, slag drainage is performed while changing the tilt angle of the ladle 1000 until the slag drainage operation is completed. That is, the exhaust can be performed while increasing the tilt angle over time after the exhaust is started. Therefore, while performing slag removal by tilting the ladle 1000, the slag (S) on the molten metal surface varies depending on the change in tilt angle and the remaining amount of slag, for example, as shown in Figures 3(a) and 3(b). can change.

In order to effectively slag (S) inside the ladle (1000), the location of the slag (4100a) is important. That is, it is necessary to adjust the horizontal position of the slag filter (4100a) so that it contacts the slag (S) rather than the molten steel (M) inside the ladle (1000). The horizontal position of the ladle 1000 can be adjusted by controlling the operation of at least one of the traveling member 4120b and the rotating body 4114b of the mobile device 4100b. Then, when the horizontal position of the exhaust device 4100a is adjusted so that it contacts the slag S, the main body 4110b is moved backward toward the discharge member 1112 using the traveling member 4120b. Accordingly, the slag (S) in contact with the exhaust device (4100a) moves toward the discharge member (1112), that is, is pushed out, and is then discharged out of the discharge member (1112).

In this way, in order to adjust the horizontal position of the slag slag (4100a) so that the slag slag (4100a) can be in contact with the slag (S), the position of the slag (S) inside the ladle 1000 must be determined. In addition, since the position of the slag (S) changes during the slag slag operation, it is desirable to determine the position of the slag (S) in real time and adjust the horizontal position of the slag slag (4100a).

In the embodiment, the inside of the ladle 1000 is photographed using the photographing unit 4200 in order to determine the location of the slag (S) inside the ladle 1000. Then, the identification unit 4300 generates an object identification image (I _d ) using the image (I _s ) captured by the photographing unit 4200 . That is, the identification unit 4300 converts or processes the image (I _s ) captured by the photographing unit 4200 to generate an image (I _d ) in which the slag to be excluded can be more clearly identified.

Additionally, the control unit 4400 uses the object identification image (I _d ) to determine the location of the slag (S) inside the ladle (1000). When the location of the slag (S) inside the ladle 1000 is identified, the control unit 4400 controls the operation of the mover 4100b to adjust the horizontal position of the slag slag (4100a).

Additionally, the control unit 4400 may determine whether slag removal is complete using the object identification image (I _d ). Additionally, the control unit 4400 may control the operation of the tilt unit 4600 according to the determination result.

The reason for using the image captured by the photographing unit 4200 as is to determine the location of the slag (S) or using the object identification image (I _d ) processed or converted without determining whether slag exclusion is complete will be discussed later. Explained again.

Figure 4 (a) and Figure 5 (a) are examples of captured images (I _s ) obtained from the photographing unit of the slag distribution device according to an embodiment of the present invention. Figures 4(b) and 5(b) show the region of interest setting image (I _Ai ) shown by extracting the region of interest (A _i ) from the captured image (I _s ) using the extraction unit according to an embodiment of the present invention. ) is an example. Figures 4(c) and 5(c) show a first converted image (I) converted from a captured image (I _s ) to a gray scale image using a first conversion unit according to an embodiment of the present invention. _c1 ) is an example. 4(d) and 5(d) show a second converted image ( _{I c2} ) is an example. Figures 4(e) and 5(e) are examples of object identification images (I _d ) generated using an identification image generator according to an embodiment of the present invention.

Here, (a) to (e) in Figures 5 are shown by showing pixels on each image (a) to (e) in Figures 4.

The photographing unit 4200 may be installed on the upper side of the ladle 1000 as shown in FIGS. 1 and 2 to photograph the inside of the ladle 1000 through the opening 1100a of the ladle 1000. At this time, the photographing unit 4200 may be installed at a position where it can photograph the ladle 1000 in an upright state before tilting, the tilted ladle 1000, and the inside of the ladle 1000 being tilted to change the angle. Additionally, the photographing unit 4200 may be installed in a position where it can photograph both the inside of the ladle 1000 and the outside of the ladle 1000.

Also, the imaging unit 4200 can be moved. That is, a separate mobile device can be connected to the photographing unit 4200, and the position of the photographing unit 4200 can be adjusted by operating the mobile device. By moving the photographing unit 4200 using a mover in this way, it is possible to photograph the ladle 1000 in an upright state before tilting, the tilted ladle 1000, and the inside of the ladle 1000 being tilted to change the angle.

The photographing unit 4200 may acquire a video or image through the photographing operation. And the photographing unit 4200 transmits the image extracted from the captured image or the photographed image to the identification unit 4300. Hereinafter, for convenience of explanation, the photographed material acquired by the photographing unit 4200 will be collectively referred to as an 'image.'

An image captured by the photographing unit 4200 (hereinafter referred to as a captured image I _s ) may include a horizontal image of the inside of the ladle 1000 . In other words, the acquired image (I _s ) captured by the photographing unit 4200 may be an image of the inside of the ladle 1000 in the lateral direction. And since the inside of the ladle 1000 is photographed while the ladle 1000 is tilted or tilted for the operation of distributing the slag (S), the photographed image (I _s ) obtained from the photographing unit 4200 is tilted. It may be a horizontal image of the inside of the ladle 1000 in its current state.

The molten steel (M) and slag (S) tapped by the ladle 1000 have different temperatures. That is, the slag (S) floating on top of the molten steel (M) is exposed to the outside air and has a lower temperature than the molten steel (M). Accordingly, the emissivity of molten steel (M) and slag (S) may vary depending on the temperature difference. In addition, the ladle 1000 in which the molten steel (M) and slag (S) are accommodated, the exhaust device (4100a) brought into the ladle (1000), and the mobile device (4100b) connected to the exhaust device (4100a) adjust the emissivity according to the temperature. You can have it.

Accordingly, the image I _s captured by the photographing unit 4200 may be a color image reflecting the emissivity of the temperature at each location of the photographed area. That is, the captured image (I _s ) acquired by the photographing unit 4200 may be a color image with color, brightness, and saturation, and an image with different colors, brightness, and saturation depending on the emissivity of each location of the photographing area. It can be. In other words, the captured image I _s acquired by the photographing unit 4200 may be an image with full natural colors. For example, when the exhaust device 4100a is brought into the tilted ladle 1000, the photographed image I _s acquired by the photographing unit 4200 is a color image such as (a) in FIG. 4 or It may be a chromatic image or a full-color image.

Meanwhile, the temperature of the slag (S) is lower than that of the molten steel, but since the slag (S) is floating on the surface of the molten steel (M), the temperature difference between the slag (S) and the molten steel (M) is not large. For example, the temperature of slag (S) may be 25°C to 50°C lower than that of molten steel (M), but the difference is not large. Accordingly, the difference in emissivity between the molten steel (M) and the slag (S) may not be large.

In addition, the slag (S) floating on top of the molten steel (M) for stainless steel has a high temperature and may exist in a liquid state. In addition, the slag in the liquid state has a high emissivity due to its high temperature, and the difference between the emissivity and the emissivity of the molten steel (M) may be small.

Accordingly, it may be difficult to distinguish between molten steel (M) and slag (S) in the captured image (I _s ) obtained by photographing the inside of the ladle 1000 by the photographing unit 4200. That is, due to the high emissivity of the slag (S), it may be difficult to identify the slag (S) using differences in color or saturation on the captured image (I _s ).

And, when the slag (S) is in a liquid state rather than a solid state, it is more difficult to distinguish it from the liquid molten steel (M). In addition, even if the slag (S) is solid, it is difficult to identify the slag using shape characteristics because the particles of solid slag do not have a constant or regular shape and have various shapes.

In addition, as explained in Figures 3 (a) and (b), the position of the slag (S) on the surface of the molten steel (M) is determined by the tilt angle of the ladle (1000), the amount of molten steel charged into the ladle (1000), It may vary depending on the internal volume of the ladle 1000 and the amount of residual slag. Therefore, it is difficult to identify the position of the slag (S) by fixing it to a specific position on the captured image (I _s ).

In this way, it is difficult to identify the slag (S) only with the captured image (Is) acquired by the photographing unit 4200. Also, if it is not easy or difficult to identify slag on the image, it is difficult to adjust the position of the slag slag (4100a) using the image.

Therefore, in the embodiment, the image (I _s ) captured by the photographing unit 4200 is processed or converted using the identification unit 4300 to display the slag S so that it is clearly distinguished from the surrounding molten steel or other components. An object identification image (I _d ) is generated.

Hereinafter, the identification unit 4300 will be described.

The identification unit 4300 uses the captured image (I _s ) acquired by the photographing unit 4200 to generate an object identification image (I _d ) through which the slag (S), which is an excretion object, can be identified. Referring to FIGS. 2 and 4 (a) to (e), the identification unit 4300 extracts the region of interest (A _i ) where the slag (S), which is the exclusion object, is located on the captured image (I _s ), and An extraction unit 4310 that generates an image displaying an area (A _i ) (hereinafter referred to as an area-of-interest setting image (I _Ai )) and an image (converted image) that produces or displays the captured image (I _s ) in two types of colors. It includes an image conversion unit 4320 that converts to (I _c )).

Additionally, the identification unit 4300 may include an identification image generator 4330 that generates an object identification image (I _d ) using the region of interest setting image (I _Ai ) and the converted image (I _c ).

The extraction unit 4310 extracts and displays the region of interest (A _i ) on the captured image (I _s ) acquired by the photographing unit 4200 .

Here, the area of interest (A _i ) may be an area where slag (S), which is an exclusion object, is located on the captured image (I _s ). For a more detailed explanation, FIGS. 4(a) and 4(b) will be described as examples. The photographed image I _s acquired by the photographing unit 4200 while the exhaust device 4100a is brought into the tilted ladle 1000 may be, for example, as shown in (a) of FIG. 4 . Referring to (a) of FIG. 4, the captured image (I _s ) includes a ladle 1000, molten steel (M) charged inside the ladle 1000, slag (S) floating on top of the molten steel (M), It may include a slag 4100a introduced into the ladle 1000, a mobile device 4100b connected to the slag 4100a, and a port 2000 located on the lower side of the ladle 1000 to receive the slag S. there is.

Here, the ladle 1000, the molten steel M, the exhaust device 4100a, the moving device 4100b, the port 2000, and the surrounding area outside the ladle 1000 included in the captured image I _s are not the exhaust target.

Meanwhile, the drainer 4100a is a means of pulling or scraping the slag S inside the body 1111 of the ladle 1000 and moving it to the discharge member 1112 connected to the body 1111. Accordingly, the object to be drained using the drainer 4100a is the slag (S) inside the body 1111 of the ladle 1000. That is, the slag (S) that has already moved to the discharge member 1112 or is discharged out of the discharge member 1112 and falls into the port 2000 and the slag (S) charged into the port 2000 are discharged through the discharger 4100a. It is not an object that you want to exclude by using it.

In summary, among the slag included in the captured image I _s , the slag S inside the body 1111 of the ladle 1000 is the target to be drained using the slag extractor 4100a. And the slag (S) that has already been moved to the discharge member 1112 of the ladle 1000, the slag (S) that has already fallen out of the discharge member 1112, and the slag (S) charged into the pot 2000 are discharged from the discharger ( This is not the target you want to exclude using 4100a).

The extraction unit 4310 extracts the area where the excluded object is located on the captured image I _s and displays or inputs it as the area of interest A _i . That is, the extraction unit 4310 distinguishes the exclusion object from the exclusion object and other components (or objects) other than the exclusion object on the captured image (I _s ), and extracts and displays the area where the exclusion object is located as the area of interest (A _i ). . To explain this in other words, the extraction unit 4310 extracts the inner area of the inner wall surface of the body 1111 of the ladle 1000 as the region of interest A _i on the captured image I _s . Here, the inner area of the inner wall surface of the body 1111 is the inner side of the body 1111 on the captured image I _s and may mean an area surrounded by the inner wall surface of the body 1111.

Hereinafter, the 'inner area of the inner wall surface of the ladle 1000 body 1111' will be abbreviated as 'the inner area of the ladle 1000 body 1111'.

In this way, extracting the inner region of the body 1111 of the ladle 1000 as the region of interest (A _i ) means that the slag (S) inside the body 1111 of the ladle 1000 is drained using the slag extractor 4100a. This is because it is the object that you want to do something to.

Extracting the region of interest (A _i ) from the captured image (I _s ) in the extraction unit 4310 can be performed using a segmentation technique using artificial intelligence or deep learning.

The segmentation technique is a method of extracting objects of interest on a pixel basis from an image. In other words, when you want to know the location of an object in an image, the shape of that object, which pixel belongs to which object, etc., the method is to segment the image and assign a label to each pixel in the image.

The extraction unit 4310 uses a segmentation technique to distinguish between an inner area of the body 1111 of the ladle 1000, which is an object of interest, and other areas on the captured image I _s . And the extractor 4310 displays or inputs the inner area of the body 1111 of the ladle 1000 as the area of interest (A _i ) on the captured image (I _s ), as shown in (b) of FIG. 4 . In other words, the extractor 4310 generates a region-of-interest setting image (I _Ai ) in which the region of interest (A _i ) is displayed on the captured image (I _s ).

To explain this again with reference to (a) and (b) of FIG. 5, the extraction unit 4310 extracts the body 1111 of the ladle 1000 in units of pixels (P) included in the captured image (I _s ). The inner area and other areas can be distinguished. Then, the extraction unit 4310 extracts pixels (P) in the inner area of the body 1111 of the ladle 1000 on the captured image (I _s ) as objects of interest, and the pixels (P) extracted as objects of interest are located at Mark or enter the area as the area of interest (A _i ). At this time, since the captured image (I _s ) is a color image implemented or displayed in chromatic colors, the region of interest (A _i ) displayed or the input region of interest setting image (I _Ai ) is shown in (b) of FIG. 5 It may be a color image including chromatic colors, such as .

In this way, extracting the region of interest (A _i ) using the segmentation technique as described above in the extraction unit 4310 can be performed by deep learning. In other words, the extraction unit 4310 learns data obtained by performing the process of extracting the region of interest (A _i ) from the captured image (I _s ) and generating the region of interest setting image (I _Ai ) dozens or hundreds of times by the user or worker. You can use the deep learning model built by doing this.

In addition, the region of interest (A _i ) extracted by the extraction unit 4310 or the region of interest setting image (I _Ai ) generated by the extraction unit 4310 may be transmitted to and utilized by the image conversion unit 4320, which will be described later.

The image conversion unit 4320 includes a first conversion unit 4321 that converts the captured image (I _s ) acquired by the photographing unit 4200 into a first conversion image (I _c1 ), which is a gray scale image, and a first conversion image ( A calculation unit 4322 that calculates an identification reference value (V _ds ) using I _c1 ) and a first converted image (I _c1 ) using the identification reference value (V _ds ) to produce a first color and a color that is distinguished from the first color. It may include a second conversion unit 4323 that converts the second color into two colors and generates a second conversion image (I _c2 ).

The first conversion unit 4321 converts the captured image (I _s ) into a gray scale image and generates the first converted image (I _c1 ). Here, the gray scale image may mean an achromatic image without color, as shown in (c) of FIG. 4. In other words, a gray scale image or an achromatic image is an image in which pixels included in the image appear in white, black, or a gray range between white and black, as shown in (c) of Figure 5, and chromatic color It can mean an image without this.

At this time, the first conversion unit 4321 converts each pixel P on the captured image I _s to appear in different series of achromatic colors depending on the brightness, that is, brightness. That is, the first converter 4321 converts any one of white, black, and gray colors between white and black according to the brightness of each pixel (P) on the captured image (I _s ). Convert it to appear as Accordingly, the first converted image (I _c1 ), which is a gray scale image such as (c) of FIG. 4 and (c) of FIG. 5, may be generated.

To be more specific, the first conversion unit 4321 assigns a first identification value to each pixel (P) included in the captured image (I _s ). Here, the first identification value may be a number within a preset range. Then, the first conversion unit 4321 selects a first identification value as a number within a preset range according to the brightness of each pixel (P) and assigns the selected first identification value. Here, the first identification value may be a rational number, and the preset range may be 0 to 255. Accordingly, the first identification value may be a rational number of 0 to 255, and may be selected and assigned as any number from 0 to 255 depending on the brightness of each pixel (P). That is, pixels P having different brightnesses may be given different first identification values in the range of 0 to 255. At this time, the lower the brightness, the closer to 0 the first identification value is assigned, and the higher the brightness, the closer to 255 the first identification value is assigned. Accordingly, the captured image I _s may be given a first identification value of any one of 0 to 255 for each pixel.

Next, the first conversion unit 4321 converts the captured image I _s into a gray scale image, that is, the first converted image I _c1 , using the first identification value assigned to each pixel P. At this time, the pixel (P) with a first identification value of 0 is black, the pixel (P) with a first identification value of 255 is white, and the pixel (P) with a first identification value between 0 and 255 is white. (P) can be converted to appear as a shade of gray between black and white. And for the pixel (P) with a first identification value between 0 and 255, the closer it is to 0, the closer it is to black or a gray color with lower brightness, and the closer it is to 255, the closer it is to white (or a gray color with low brightness). is converted to high gray.

Accordingly, for example, the captured image I _s shown in (a) of FIG. 4 may be converted into a gray scale image such as (c) of FIG. 4 . If this is explained by displaying a pixel P on the image, the captured image I _s as shown in (a) of FIG. 5 can be converted into a gray scale image as shown in (c) of FIG. 5, for example.

Additionally, the first conversion unit 4321 may display the region of interest (A _i ) on the first converted image (I _c1 ). At this time, the region of interest (A _i ) on the first converted image (I _c1 ) is displayed at the same position or pixel as the region of interest (A _i ) on the region of interest setting image (I _Ai ).

The calculation unit 4322 calculates the identification reference value (V _ds ) using the first identification value assigned to each pixel (P) of the first converted image (I _c1 ). As described above, the first conversion unit 4321 generates the first converted image (I c1 ) _by assigning a first identification value of 0 to 255 according to the brightness of each pixel (P) of the captured image (I _s ). . Accordingly, each of the plurality of pixels P included in the first converted image I _c1 is assigned a first identification value of any one of 0 to 255. In other words, each pixel (P) of the first converted image (I _c1 ) has a first identification value of any one of 0 to 255.

The calculation unit 4322 calculates the identification reference value (V _ds ) using the first identification value of each pixel (P) of the first converted image (I _c1 ). That is, the calculation unit 4322 uses the first identification value of all pixels (P) included in the first converted image (I _c1 ) and the first identification value of the pixels (P) in the region of interest (A _i ). This calculates the identification reference value (V _ds ). To be more specific, the calculation unit 4322 calculates the average (PV _t-mean ) of the first identification values of all pixels (P) included in the first converted image (I _c1 ), the first converted image (I _c1 ) The maximum value (PV _Ai-max ) of the first identification values of the pixels (P) in the area of interest (A _i ) in the image, the pixel (P) in the area of interest (A _i ) in the first converted image (I _c1 ) The identification reference value (V _ds ) is calculated using the average of the first identification values (PV _Ai-mean ) (see Equation 1).

[Formula 1]

'C' included in Equation 1 is a constant and may be, for example, '0.65'. Here, the constant 'C' is not limited to '0.65' and can be changed to an optimized value depending on the waste management environment or conditions.

The identification reference value (V _ds ) calculated in this way is used to generate an object identification image (I _d ) in the second conversion unit 4323.

Meanwhile, the overall brightness or brightness of the photographed image (I _s ) acquired by the photographing unit 4200 may vary depending on the brightness or illuminance of the surrounding environment in which the drainage operation is conducted. That is, the brightness of each pixel P of the captured image I _s acquired by the photographing unit 4200 may vary depending on the brightness or illuminance around the photographing unit 4200 or the ladle 1000.

Therefore, it is calculated using the first identification values of all pixels (P) included in the first converted image (I _c1 ). That is, all pixels (P) included in the first converted image (I _c1 ) are calculated using the average (PV _t-mean ) of the first identification values. Here, since the first identification value is given according to the brightness in the captured image (I _s ), the identification reference value (V) is obtained by using or reflecting the brightness of all pixels (P) included in the first converted image (I _c1 ). It can be explained as calculating _ds ). In other words, it can be explained that the identification reference value (V _ds ) is calculated by reflecting the brightness or illuminance of the surrounding environment where the waste management is conducted.

Accordingly, the object identification image (I _d ) generated using the identification reference value (V _ds ) is an image generated by reflecting the brightness or illuminance of the surroundings of the photographing unit 4200 or the ladle 1000 and where the distribution operation is performed. You can.

The second conversion unit 4323 uses the calculated identification reference value (V _ds ) to convert the first conversion image (I _c1 ) to appear in a first color and a second color distinct from the first color. That is, the second conversion unit 4323 converts the first conversion image (I _c1 ), which is a gray scale image containing three or more achromatic colors, to appear in two types or two colors to produce a second conversion image. Create (I _c2 ). In other words, the second conversion unit 4323 converts the first conversion image (I _c1 ), which is a gray scale image containing three or more achromatic colors, into only two types of colors or only two colors. It is converted to appear and a second converted image (I _c2 ) is generated. That is, the second converted image I _c2 generated by the second conversion unit 4323 may be an image in which each of the plurality of pixels is displayed in only one of two types of colors or two colors.

Here, the first color and the second color are preferably set to colors that are not included in the captured image I _s . For example, the first color may be an achromatic color, and the second color may be an achromatic color different from the first color. As a more specific example, the first color may be black, and the second color may be white. Accordingly, the second converted image I _c2 can be described as a black-and-white image implemented or represented in two colors, black and white. And, the second conversion unit 4323 can be described as a means for converting the first conversion image (I _c1 ) into a black and white image.

The second conversion unit 4323 selects one of the first and second values for each pixel (P) of the first converted image (I _c1 ) depending on whether the first identification value is less than or equal to the identification reference value (V _ds ). One second identification value is assigned. That is, the second conversion unit 4323 assigns a first value to a pixel (P) whose first identification value is less than or equal to the identification reference value (V _ds ) in each pixel (P) of the first converted image (I _c1 ). , a second value is assigned to the pixel (P) whose first identification value exceeds the identification reference value (V _ds ). For example, the first value may be '0' and the second value may be '1'. Thereafter, the second conversion unit 4323 determines that in each pixel (P) of the first converted image (I _c1 ), the pixel (P) to which the first value (0) is assigned is black and the second value (1) is changed to black. A black-and-white image is created so that the pixel (P) assigned with ) appears white. That is, the first converted image (I _c1 ), which is a gray scale image, is converted into the second converted image (I _c2 ) which is a black-and-white image.

For example, the second converted image (I c2 ) _obtained by converting the first converted image (I _c1 ) of FIG. 4 (c) in the second converter 4323 may be the same as (d) of FIG. 4 . To explain this by displaying pixels (P) on the image, in the first converted image (I _c1 ) of FIG. 5 (c), each pixel (P) is black and black as shown in (d) of FIG. 5. It can be converted to appear white.

Meanwhile, in the second converted image (I _c2 ), as shown in FIGS. 4 (d) and 5 (d), the pixels (P) for other components except the molten steel (M) are in the first color, that is, black. ) can be displayed. That is, except for the pixel P identified as molten steel M in the second converted image I _c2 , the rest may be displayed in black. More specifically, the slag (S), the ladle (1000), the outside of the ladle (1000), the drainer (4100a), the mover (4100b), the port (2000), and the port (2000) in the area of interest (A _i ). The slag (S) present may be displayed in black.

In this way, the second conversion unit 4323 converts the first conversion image (I _c1 ) into the second conversion image (I _c2 ) using the identification reference value (V _ds ). Here, the identification reference value (V _ds ) is a value calculated using the brightness of the first converted image (I _c1 ), and is a value calculated to identify the slag (S), which is the exclusion object, so the identification reference value (V _ds ) can be named ‘identification standard brightness’.

Accordingly, the identification standard brightness, which is the identification reference value (V _ds ), is the average brightness of all pixels (P) included in the first converted image (I _c1 ), and the region of interest of the first converted image (I _c1 ). The maximum brightness among the brightnesses of the pixels (P) included in (A _i ), calculated using the brightness average of the pixels (P) included in the region of interest (A _i ) of the first converted image (I _c1 ). It can be explained as:

In addition, the average brightness of all pixels (P) included in the first converted image (I _c1 ) is the average (PV _t-mean ) of the first identification value, and the region of interest of the first converted image (I _c1 ) The maximum brightness of the pixels (P) included in (A _i ) is the maximum value (PV _Ai-max ) among the first identification values, and is included in the region of interest (A _i ) of the first converted image (I _c1 ). The brightness average for the pixels (P) can be explained as the average (PV _Ai-mean ) of the first identification values of the pixels in the region of interest (A _i ) of the first converted image (I _c1 ).

And, the second conversion unit 4323 converts the first conversion image (I _{c1 )} into the second conversion image (I _c2 ) using the identification reference value (V _ds ). It can be described as converting the first converted image (I _c1 ) into the second converted image (I _c2 ) using the brightness of each pixel (P) and the brightness of the identification standard. In other words, the second converted image (I _c2 ) converts the pixels (P) with a brightness below the identification standard brightness into the first color among the pixels (P) of the first converted image (I _c1 ). , It can be explained as converting the pixel P having a brightness exceeding the identification standard brightness into a second color.

The first and second conversion units 4321 and 4323 as described above may be means including a device including the cv2 library program.

The identification image generator 4330 uses the excluder 4100a to generate an object identification image I _d in which the object to be excluded appears distinct from other components. That is, the identification image generator 4330 generates the object identification image I _d so that the slag S in the area of interest A _i is displayed in the first color on the area of interest setting image I _Ai .

More specifically _, the _{identification} image generator 4330 generates an area of interest setting image (I The pixel (P) on _Ai ) is converted to the first color and displayed.

Here, the first color of the second converted image I _c2 is an achromatic color, such as black, and the region of interest setting image I _Ai is an image including chromatic colors, that is, a color image. Accordingly, some areas on the region of interest setting image (I _Ai ), which is a color image, are converted to black. In other words, an object identification image (I _d ) in which at least some of the plurality of pixels included in the region of interest ( _Ai ) of the region of interest setting image (I Ai), which is a chromatic image, are converted to black, which is an achromatic color. is created. That is, an object identification image (I _d ) is generated in which the area or pixel (P) corresponding to the slag (S), which is the exclusion object, in the area of interest (Ai) is displayed in black. Accordingly, the object identification image I _d may be an image in which the pixels P are displayed in black and the remaining pixels P other than the black pixels P are displayed in chromatic colors.

For example, the object identification image (I _d ) generated by the identification image generator 4330 using the region of interest setting image (I _Ai ) of FIG. 4 (b) may be the same as (e) of FIG. 4 there is. If this is explained by displaying a pixel (P) on the image, the object identification image (I _d ) generated using the region of interest setting image (I _Ai ) of FIG. 5 (b) will be the same as (e) of FIG. 5 You can.

In this way, the object identification image (I _d ) is a pixel (P) corresponding to the slag (S) in the area of interest (A _i ), as shown in Figure 4 (e) or Figure 5 (e), and other pixels around it. It is created to appear in a completely different color from (P), such as black. That is, the object identification image I _d is displayed in black and a chromatic color different from black. To be more specific, the region of interest (A _i ) of the object identification image (I _d ) is displayed in black and a chromatic color different from black. Additionally, in the object identification image (I _d ), areas other than the area of interest (A _i ) are displayed in chromatic colors different from black. To explain in another way, the object identification image I _d may be an image including chromatic color and black, which is an achromatic color.

In this way, the object identification image I _d is generated as an image including black and a chromatic color different from black. Accordingly, the position or pixel P displayed in black in the object identification image I _d can be identified as slag. Therefore, the accuracy of identifying slag (S) on the image can be improved. That is, the identification visibility of the slag, which is the exclusion object, on the object identification image (I _d ) is improved.

As described above, the identification image generator ₄₃₃₀ generates an area of interest setting image ₍ I _Ai ) is converted to the first color to generate an object identification image (I _d ). Accordingly, the identification image generator 4330 mixes the region-of-interest setting image (I _Ai ) and the second transformed image (I _c2 ), or adds the second transformed image (I _c2 ) to the region-of-interest setting image (I _Ai ). It can be described as generating an object identification image (I _d ) by combining them. That is, the identification image generator 4330 generates an object identification image (I _d ) by mixing or combining the region of interest setting image (I _Ai ), which is a chromatic or color image, and the second converted image (I _c2 ), which is a black-and-white image. It can be explained as: Accordingly, the object identification image (I _d ) can be described as an image created by mixing or combining the region-of-interest setting image (I _Ai ), which is a chromatic or color image, and the second converted image (I _c2 ), which is a black-and-white image.

The control unit 4400 controls the operation of at least one of the exhaust unit 4100 and the tilt unit 4600 using the object identification image I _d generated by the identification image generator 4330. Referring to FIG. 2, the control unit 4400 includes a determiner 4410 that determines whether slag has been completely eliminated using the area of pixels P displayed in the first color, for example, black, in the object identification image I _d . It may include a first controller 4420a that controls the operation of the exhaust portion 4100 and a second controller 4420b that controls the operation of the tilting portion 4600 according to the judgment result in the device 4410.

The judge 4410 quantitatively determines the amount of slag (S) inside the ladle 1000 and determines whether slag removal is complete. To this end, the determiner 4410 sums (B _c ) the areas of pixels (P) displayed in black on the object identification image (I _d ) as shown in (e) of FIG. 4 or (e) of FIG. 5 . At this time, the pixels P included in the object identification image I _d may have the same area. Accordingly, by multiplying the number of pixels P displayed in black in the object identification image I _d by the area of one pixel, the total area B _c of the pixels displayed in black is calculated.

Here, the pixels (P) displayed in black in the object identification image (I _d ) are the objects to be excluded, so the summed area (B _c ) of the pixels (P) displayed in black is 'the summed area of the excluded objects (B _c )'.

Next, the judge 4410 compares the calculated total area (B _c ) of the exclusion object with a preset reference area (B _s ). At this time, if the total area (B _c ) of the slag objects is less than or equal to the reference area (B _s ), it can be determined that slag (S) slag removal is complete. In other words, it is determined that the slag remaining inside the ladle 1000 is below the target amount, and there is no need to carry out the slag removal operation anymore. In other words, when the total area (B _c ) of the drainage object becomes less than the reference area (B _s ), it is determined that it is time to erect the ladle 1000 and end drainage. Conversely, when the total area (B _c ) of the slag objects exceeds the reference area (B _s ), it is determined that more slag (S) is needed. In other words, it is determined that the slag S remaining inside the ladle 1000 exceeds the target amount and that further slag removal operation must be performed.

In addition, by comparing the total area (B _c ) of the drainage object with the reference area (B _s ) in real time, it is possible to predict when drainage is completed or when the ladle must be erected.

The first and second controllers 4420a and 4420b control the operation of the tilt unit 4600 and the exhaust unit 4100 according to the determination result in the judge 4410.

For example, if the total area (B _c ) of the slag object exceeds the reference area (B _s ) and the judge 4410 determines that more slag (S) needs to be excluded, the first controller 4420a operates the slag ( The mover 4100b is controlled to maintain the state in which the ladle 1000 (4100a) is brought into the ladle 1000. And the second controller 4420b controls the tilt unit 4600 to further tilt the ladle 1000 or maintain the tilt state.

In addition, as in the above-mentioned example, if the judge 4410 determines that more slag S is needed to be excluded, the first controller 4420a uses the object identification image I _d to determine the horizontal direction of the slag 4100a. Adjust the position. That is, the first controller 4420a operates the mover 4100b to adjust the position of the slag 4100a so that the slag 4100a can be positioned at the position of the slag on the object identification image I _d .

At this time, the method by which the first controller 4420a adjusts the position of the slag 4100a so that the slag 4100a can be positioned at the position of the slag on the object identification image I _d is not particularly limited.

When the total area (B _c ) of the exclusion object becomes less than or equal to the reference area (B _s ), the determiner 4410 determines that exclusion is complete. Accordingly, the first controller 4420a operates the moving unit 4100b of the exhaust unit 4100 to move the exhaust unit 4100a out of the ladle 1000, and the second controller 4420b operates the tilting unit 4600. The ladle 1000 is erected.

The first and second controllers 4420a and 4420b as described above may include a Programmable Logic Controller (PLC).

In the above, it has been described as an example that the first controller 4420a controls the operation of the mover 4100b so that the slag 4100a can be positioned at the position of the slag on the object identification image I _d . However, moving the slag 4100a to the position of the slag on the object identification image I _d may be performed by an operator.

For example, when the object identification image (I _d ) is displayed or displayed on the display unit 4500, the operator checks the display unit 4500 to determine the location of the pixel (P) displayed in black on the object identification image (I _d ). to determine the location of the slag. In addition, the operator can control the operation of the mover 4100b so that the exhauster 4100a can be positioned at the location of the slag identified on the object identification image I _d .

In addition, in the above, the determiner 4410 calculates the area (B _c ) of the area or pixel (P) displayed in black on the object identification image (I _d ), and compares it with the reference area (B _s ). It was explained as determining whether drainage was completed. However, it is not limited to this, and the operator may determine whether exclusion is complete by calculating the area (B _c ) of the area or pixel displayed in black on the object identification image (I _d ). In other words, when the object identification image (I _d ) is displayed or displayed on the display unit 4500, the operator can determine the area of the area or pixel (P) displayed in black on the object identification image (I _d ) displayed on the display unit 4500. can be calculated. In addition, it is possible to determine whether drainage is complete by comparing the calculated area, that is, the total area (B _c ) of the drainage object, with the reference area (B _s ).

Figure 6 is a flowchart showing a method for discharging slag according to an embodiment of the present invention.

Hereinafter, a method for excluding slag according to an embodiment of the present invention will be described with reference to FIG. 6.

First, the ladle 1000 containing the tapped molten steel M is seated on the seating table 4610 of the tilt section 4600. Then, the ladle 1000 is tilted at a predetermined angle (S110), and the exhauster 4100a is brought into the ladle 1000 (S120). Then, the position of the slag extractor 4100a is adjusted so that it can contact the slag S inside the ladle 1000 or face the slag S (S130). Thereafter, the slag (S) is pushed out of the ladle (1000) using the exhauster (4100a) to exhaust the slag (S) (S140). At this time, the tilt angle of the ladle 1000 is increased using the tilt portion 4600 over time.

While the slag S is excluded in this way, the inside of the ladle 1000 is photographed using the photographing unit 4200, and the identification unit 4300 identifies the slag on the acquired image (S200).

To be more specific, the photographing unit 4200 acquires a photographed image (I _s ) by photographing the inside of the ladle 1000 (S210). Then, the extraction unit 4310 extracts a region of interest (A i ), _which is an area containing slag (S), which is an exclusion object, on the captured image (I _s ). Then, the region of interest (A _i ) is displayed on the captured image (I _s ) to generate the region of interest setting image (I _Ai ), for example, as shown in (b) of FIG. 4 (S220).

Additionally, the first conversion unit 4321 converts the captured image I _s into gray scale and generates a first converted image I _c1 (S231). That is, the first conversion unit 4321 assigns a first identification value to the plurality of pixels P included in the captured image I _s according to the brightness of each pixel P. And the first conversion unit 4321 converts each pixel (P) into black, white, and a gray color between black and white according to the given first identification value and displays it. Accordingly, for example, the first converted image I _c1 as shown in (c) of FIG. 4 is generated.

Next, the calculation unit 4310 calculates the identification reference value (V _ds ) using the first converted image (I _c1 ) (S232). At this time, _the calculation unit 4310 calculates the average (PV _t-mean ) of the first identification values of all pixels (P) included in the first converted image (I _c1 ), the region of interest ( The maximum value (PV _Ai-max ) among the first identification values of the pixels (P) in the A _i ), the first identification of the pixels (P) in the region of interest (A _i ) on the first transformed image (I _c1 ) The identification reference value (V _ds ) is calculated using the value average (PV _Ai-mean ) (see Equation 1).

Next, the second conversion unit 4323 converts the first conversion image (I _c1 ) into the second conversion image (I _c2 ) using the identification reference value (V _ds ) (S233). That is, the second converter 4323 sets the first value '1' to each pixel (P) of the first converted image (I _c1 ) whose first identification value is less than or equal to the identification reference value (V _ds ). is assigned, and a second value, '2', is assigned to the pixel P whose first identification value exceeds the identification reference value (V _ds ). Thereafter, the second conversion unit 4323 determines that in each pixel (P) of the first converted image (I _c1 ), the pixel (P) to which the first value (0) is assigned is black and the second value (1) is changed to black. A black-and-white image is created so that the pixel (P) assigned with ) appears white. That is, the first converted image (I _c1 ), which is a gray scale image, is converted into the second converted image (I _c2 ) which is a black-and-white image.

Next, the identification image generator 4330 generates an object identification image (I _d ), which is an image in which the excluded object is distinctly displayed. That is, the identification image generator 4330 generates a pixel on the region of interest setting image (I _Ai ) at the same position as the pixel (P) displayed in the first color in the region of interest (A _i ) of the second converted image (I _c2 ). (P) is converted to the first color and displayed (S240). Accordingly, some areas on the region of interest setting image (I _Ai ), which is a chromatic color image, are converted to the first color, that is, black. That is, in the region of interest (Ai) of the region of interest setting image (I _Ai ), which is a chromatic image such as (b) in FIG. 4, the area or pixel (P) corresponding to the slag (S), which is the exclusion object, is black. An object identification image (I _d ) indicated by is generated. The generated object identification image (I _d ) may be, for example, as shown in (e) of FIG. 4 .

When the object identification image (I _d ) is generated in this way, it is transmitted to the determination unit 4410 of the control unit 4400. The judge 4410 quantitatively determines the amount of slag (S) inside the ladle 1000 and determines whether slag removal is complete. For this purpose, the determiner 4410 sums (B _c ) the areas of pixels (P) displayed in black on the object identification image (I _d ), for example, as shown in (e) of FIG. 4 . Then, the judge 4410 compares the calculated total area (B _c ) of the exclusion object with a preset reference area (B _s ) (S300).

For example, when the total area (B _c ) of the slag object is less than or equal to the reference area (B _s ) (example), it may be determined that slag (S) slag removal has been completed. In this case, the second controller 4420b operates the mover 4100b of the exhaust unit 4100 to move the exhaust unit 4100a out of the ladle 1000 (S420), and the second controller 4420b operates the tilt unit 4600. ) is operated to erect the ladle 1000 (S410).

On the contrary, when the total area (B _c ) of the slag object exceeds the reference area (B _s ) (No), the judge 4410 determines that more slag (S) needs to be sagged. Accordingly, the first controller 4420a controls the mover 4100b to maintain the drainer 4100a brought inside the ladle 1000, and the second controller 4420b further tilts the ladle 1000. Alternatively, the tilt unit 4600 is controlled to maintain the tilt state.

Additionally, the first controller 4420a adjusts the horizontal position of the spreader 4100a using the object identification image I _d (S130). That is, when the first controller 4420a controls the operation of the mover 4100b so that the slag 4100a can be positioned at the position of the slag on the object identification image I _d , the position of the slag 4100a is adjusted ( S130). That is, the first controller 4420a operates the mover 4100b to adjust the position of the slag sieve 4100a so that the slag 4100a can contact the slag S inside the body 1111 of the ladle 1000. Adjust (S130).

According to these embodiments, a region of interest (A _i ) containing slag, which is an exclusion object, is extracted from the captured image (I _s ) using a segmentation technique. Accordingly, even if the position of the slag (S) on the surface of the molten steel (M) in the ladle (1000) is not fixed and changes, the slag (S), which is an object of slag, can be easily identified.

_In addition _, the object identification image ₍ I _d ) is created. Accordingly, the object identification image I _d is displayed in black and a chromatic color different from black. Additionally, the region of interest (A _i ) of the object identification image (I _d ) is displayed in black and a chromatic color different from black. Accordingly, the position or pixel P displayed in black in the object identification image I _d can be identified as slag. Therefore, the visibility and identification accuracy of the slag (S) on the image can be improved.

And, when generating the object identification image (I _d ), the entire brightness of the captured image (I _s ) is reflected and generated. That is, the object identification image (I _d ) is generated by reflecting the overall brightness or brightness of the captured image (I _s ), which changes depending on the brightness or illuminance of the environment around the ladle (1000) or where distribution operations are performed. Therefore, even if the brightness or illuminance of the environment surrounding the ladle 1000 or where the drainage operation is performed changes, the object identification image (I _d ) in which the slag, which is the object of drainage, appears in a clearly different color from other objects can be generated. . Therefore, the visibility and identification accuracy of the slag (S) on the image can be improved.

Accordingly, the location of the slag (S) within the ladle 1000 can be determined more accurately. In addition, this makes it easier to adjust the position of the slag slag (4100a) so that it can contact the slag (S) inside the ladle (1000). Therefore, the slag (S) inside the ladle 1000 can be effectively excluded.

In addition, the amount of slag remaining inside the ladle 1000 can be determined, and the timing of completion of slag removal can be more accurately predicted.

1000: First container (ladle) 4100: Baejae part
4100a: Bae Jae-gi 4200: Filming Department

Claims (19)

A process of acquiring a photographed image (I _s ) including chromatic colors by photographing a container charged with molten material with slag floating on the surface;
A process of extracting a region of interest (A _i ) containing slag, which is an object to be excluded, from the captured image (I _s ), and generating a region of interest setting image (I _Ai ) displaying the extracted region of interest (A _i ). ;
Among the plurality of pixels included in the region of interest (A _i ) of the region of interest setting image (I _Ai ) including chromatic colors, the color of the pixel having a brightness below the identification standard brightness is converted to an achromatic first color, and the object Process of generating an identification image (I _d ); and
Identifying a pixel displayed in the first color on the object identification image I _d as slag, which is an exclusion object. In claim 1,
The process of extracting the region of interest (A _i ) is,
A process of distinguishing between an inner area of the inner wall surface of the container and other areas on the captured image (I _s );
A method of identifying slag comprising a process of setting the inner area of the divided inner wall surface of the container as a region of interest (A _i ). In claim 2,
The process of distinguishing the inner area of the inner wall of the container from other areas on the captured image (I _s ) and the process of setting the inner area of the inner wall of the container as the area of interest (A _i ) are segmentation using artificial intelligence ( A method of identifying slag using segmentation technique. In claim 1,
The process of generating the object identification image (I _d ) is,
A process of assigning a first identification value to each of a plurality of pixels included in the captured image (I _s ) according to the brightness of each pixel; and
The color of each of the plurality of pixels of the captured image (I _s ) is converted to one of the achromatic colors according to the first identification value assigned to each, and a first converted image (I _c1 ) is a gray scale image. ) process of producing slag; identification method of slag including. In claim 4,
The process of generating the object identification image (I _d ) is,
Setting a region of interest (A _i ) in the first converted image (I _c1 ) using the region of interest setting image (I _Ai );
Calculating an identification reference value (V _ds ) using a first identification value for all pixels included in the first converted image (I _c1 ); and
Using the identification reference value (V _ds ), the first converted image (I _c1 ) is converted into an image that appears in two types of achromatic colors, including the first color and a second color that is an achromatic color distinct from the first color. A process of generating a second converted image (I _c2 ) by
A slag identification method using the identification reference value (V _ds ) as the identification reference brightness. In claim 5,
In calculating the identification reference value (V _ds ),
An average (PV _t-mean ) of the first identification value for all pixels included in the first converted image (I _c1 );
The maximum value (PV _Ai-max ) among the first identification values of pixels in the region of interest (A _i ) of the first converted image (I _c1 ); and
A slag identification method calculated using the average (PV _Ai-mean ) of the first identification values of pixels in the region of interest (A _i ) of the first converted image (I _c1 ). In claim 6,
In calculating the identification reference value (V _ds ), a slag identification method calculated using the formula below.
[formula]

(C: constant) In claim 5,
The process of generating the second converted image (I _c2 ) is,
In each of the plurality of pixels included in the first converted image (I _c1 ), a second identification value of the first value is assigned to a pixel whose first identification value is less than or equal to the identification reference value (V _ds ), and the first identification value is assigned to the pixel. A process of assigning a second value, which is a second identification value different from the first value, to the pixel exceeding the identification reference value (V _ds );
In each of the plurality of pixels included in the first converted image (I _c1 ), a pixel whose assigned second identification value is a first value is displayed with the first color, and a pixel whose assigned second identification value is a second value is displayed with the first color. Converting the first converted image (I _c1 ) by displaying in the second color. In claim 8,
The process of generating the object identification image (I _d ) is,
A process of converting the color of a pixel at the same location as a pixel marked with the first color on the second converted image (I _c2 ) among a plurality of pixels included in the region of interest setting image (I _Ai ) to the first color. Identification method of slag containing ; In claim 8 or claim 9,
The first color is black, the second color is white,
The second converted image (I _c2 ) is a black-and-white image in which a plurality of pixels are displayed in black or white,
Among the pixels in the area of interest (A _i ) of the object identification image (I _d ), the pixel at the same position as the pixel displayed in black in the area of interest (A _i ) of the second converted image (I _c2 ) is black. It is displayed as
The process of identifying slag as an exclusion object includes identifying pixels displayed in black in the object identification image (I _d ) as slag as an exclusion object. The slag identification method of any one of claims 1 to 9, comprising: identifying slag on the object identification image (I _d );
A process of adjusting the position of the slag so that the slag is positioned on the identified object identification image (I _d ); and
A process of discharging slag out of the container by moving the slag to the outside of the container. In claim 11,
A process of determining whether slag exclusion is complete using the slag on the identified object identification image (I _d ),
The process of determining whether the slag slag is complete is performed before the process of adjusting the position of the slag slag. In claim 12,
The process of determining whether the slag drainage is complete is,
A process of adding up the area (B _c ) of pixels identified as slag on the object identification image (I _d ); and
A process of determining that slag exclusion is complete when the added area (B _c ) is less than or equal to a preset reference area (B _s ),
When it is determined that slag drainage is completed in the above determination process, a slag drainage method comprising the step of erecting the tilted container. In claim 11,
A method for discharging slag, wherein the melt includes molten steel, and the container includes a ladle. In claim 14,
The molten steel is a method of removing slag, which is molten steel for stainless steel production. a drainage unit capable of moving into a container charged with molten material with slag floating on the surface;
a tilting unit on which the container can be seated and which is installed to tilt the container;
a photographing unit installed on the outside of the container to obtain images according to emissivity;
A region of interest (A _i ) containing slag, which is an exclusion object, is extracted from the photographed image (I _s ) acquired by the photographing unit, and a brightness of a plurality of pixels included in the region of interest (A _i ) is equal to or less than the brightness standard for identification. An identification unit that generates an object identification image (I _d ) by converting the color of the pixel having to any one color of the achromatic color series; and
A control unit that controls the operation of at least one of the exhaust unit and the tiller using the object identification image (I _d ). In claim 16,
The captured image (I _s ) obtained from the photographing unit includes chromatic colors,
The identification unit,
On the captured image (I _s ), the inner area of the inner wall of the container is extracted as the area of interest (A _i ) where the slag to be excluded is located using the exhaust portion, and the extracted area of interest (A _i ) is photographed. An extraction unit that displays an image (I _s ) and generates a region-of-interest setting image (I _Ai ) including chromatic colors; and
A first conversion that converts each of the plurality of pixels included in the captured image (I _s ) into one color among achromatic colors according to the brightness of each pixel, thereby generating a first converted image that is a gray scale image. Part; includes,
The first conversion unit is a slag drainage device that sets a region of interest (A _i ) in the first converted image (I _c1 ) using the region of interest setting image (I _Ai ). In claim 17,
The identification unit,
The average brightness of all pixels included in the first converted image (I _c1 ), the maximum brightness of the pixels included in the region of interest (A _i ) of the first converted image (I _c1 ), the first converted image (I c1 ), a calculation unit that calculates the identification standard brightness using the brightness average of pixels included in the region of interest (A _i ) of the converted image (I _c1 );
Each of the plurality of pixels included in the first converted image is converted to one of black and white depending on whether the brightness of each pixel is less than or equal to the identification standard brightness, thereby producing a black-and-white image. 2 A second conversion unit that generates a converted image; and
Among the plurality of pixels included in the region of interest setting image (I _Ai ), the color of the pixel at the same position as the pixel marked in black on the second converted image (I _c2 ) is converted to black to obtain the object identification image ( I _d ) A slag drainage device comprising an identification image generating unit that generates. The method of any one of claims 16 to 18,
The control unit,
a judge that determines whether slag drainage is complete using a pixel identified as slag, which is an exclusion object, on the object identification image (I _d );
a first controller that controls the operation of the slag slag so that its position is adjusted when the judge determines that slag slag has not been completely drained; and
When the judge determines that slag removal is complete, a second controller controls the operation of the tilting unit so that the container is upright.

Images