KR102161348B1 - Apparatus and Method for Determining Descaling Method Based on Image of Slab - Google Patents

Abstract

An apparatus for determining a descaling type according to an aspect of the present invention capable of automatically determining a descaling type to be applied to a corresponding slab based on an image of the slab comprises: a photographing unit including a plurality of cameras for photographing the surface of the slab; A modeling unit for generating an FSB type determination model by learning the FSB (Finished Scale Breaking) type actually applied to the slabs corresponding to the training images acquired by the photographing unit for a predetermined period of time; And one FSB to be applied to the target slab based on the determined FSB types by inputting target images obtained by photographing the target slab by each camera into the FSB type determination model to determine an FSB type for each target image of each camera. It characterized in that it comprises a FSB type determination unit for determining the type.

Description

Apparatus and Method for Determining Descaling Method Based on Image of Slab}

The present invention relates to the steel industry, and more particularly, to a technique for removing the scale of the slab.

In general, the rolling process is a process of manufacturing a slab in a coil shape. In the heating furnace of the rolling process, the slab can be preheated and heated. When the slab is heated in a furnace, scale may be generated on the surface of the slab. Scale means an oxide film formed on the surface of a slab due to an oxidation reaction. In addition, scale may be generated on the surface of the slab as the slab is exposed to the atmosphere while the slab is transferred from the heating furnace to the rolling mill.

Therefore, in the rolling process, various descaling methods are used to remove the scale of the slab generated on the surface of the slab. The descaling method includes a scale breaker method that removes scale by spraying water after applying pressure to the scale to generate cracks, a method of removing scale by spraying high pressure water, and a method of removing scale by applying pressure using small solids. There are a short ball method to remove the scale by immersing the slabs in an aqueous acid solution, and a pickling method.

However, in the conventional rolling process, since the operator determines the descaling method by visually checking the state of the slab surface, not only is there a limit to detailed inspection on the entire length of the slab, but the scale removal method is different depending on the skill of the operation. There is a problem that it can be damaged.

In addition, when the scale is not properly removed due to the application of an incorrect scale removal method, the scale removal operation must be re-executed, so that the descaling operation may additionally consume a lot of time and cost.

Korean Patent Publication No. 10-2004-0047285 (Name of invention: Scale scatter prevention device of descaler, Publication date: June 5, 2004)

The present invention is to solve the above-described problem, and provides an apparatus and method for determining a descaling type based on a slab image capable of automatically determining a descaling type to be applied to a corresponding slab based on the image of the slab. To do.

In addition, another technical feature of the present invention is to provide an apparatus and method for determining a descaling type based on a slab image that can improve the accuracy of determining a descaling type.

In addition, another technical object of the present invention is to provide an apparatus and method for determining a slab image-based descaling type that can more accurately capture the amount and pattern of the scale of the slab surface.

Descaling type determination apparatus according to an aspect of the present invention for achieving the above object is a photographing unit consisting of a plurality of cameras for photographing the surface of the slab; A modeling unit for generating an FSB type determination model by learning the FSB (Finished Scale Breaking) type actually applied to the slabs corresponding to the training images acquired by the photographing unit for a predetermined period of time; And one FSB to be applied to the target slab based on the determined FSB types by inputting target images obtained by photographing the target slab by each camera into the FSB type determination model to determine an FSB type for each target image of each camera. It characterized in that it comprises a FSB type determination unit for determining the type.

A method for determining a descaling type according to another aspect of the present invention for achieving the above object comprises: photographing the surface of the slab with a plurality of cameras; Generating an FSB type determination model by modeling an FSB type actually applied to a slab corresponding to the training images acquired during a predetermined period and each training image; Inputting test images acquired during a predetermined period into the FSB type determination model to obtain weights for each camera; Determining an FSB type for each camera by inputting target images obtained by photographing a target slab with the plurality of cameras into the FSB type determination model; And determining one FSB type to be applied to the target slab among the FSB types determined for each camera based on the weight.

According to the present invention, the amount and pattern of the scale of the slab surface is determined based on the image obtained by photographing the surface of the slab, and the descaling type to be applied to the slab is automatically determined according to the determined amount and pattern of the scale. Therefore, not only can the deviation of the descaling type determined by the operator be removed, but also the entire length of the slab can be inspected in detail, thereby improving product quality.

In addition, according to the present invention, since the descaling type to be applied to the slab is accurately determined according to the amount and pattern of the scale on the surface of the slab, it is possible to prevent re-performing of the descaling operation due to insufficient removal of the scale. It has the effect of saving time and cost.

Further, according to the present invention, it is possible to improve the accuracy of determining the descaling type by continuously learning the image of the photographed slab surface.

In addition, the present invention has an effect that it is possible to more accurately photograph the amount and pattern of the scale of the slab surface by setting different shooting angles of cameras for photographing the slab surface.

1 is a view showing the flow of a rolling process to which a descaling type determination device according to an embodiment of the present invention is applied
2 is a diagram showing a configuration of an apparatus for determining a descaling type according to an embodiment of the present invention
3 is a perspective view showing the configuration of a photographing unit according to an embodiment of the present invention.
4 is a perspective view showing a configuration of a photographing unit according to another embodiment of the present invention.
5 is a table showing an example of an FSB type applied to a slab according to an embodiment of the present invention.
6 is a table showing an example in which an FSB type determination unit according to an embodiment of the present invention selects a final FSB type applied to a slab
7 is a flowchart showing a method of determining a descaling type according to an embodiment of the present invention.
8 is a flowchart showing a method of generating an FSB type determination model according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The meaning of the terms described in this specification should be understood as follows.

Singular expressions should be understood as including plural expressions unless clearly defined differently in context, and terms such as “first” and “second” are used to distinguish one element from other elements, The scope of rights should not be limited by these terms.

It is to be understood that terms such as "comprise" or "have" do not preclude the presence or addition of one or more other features or numbers, steps, actions, components, parts, or combinations thereof.

The term “at least one” is to be understood as including all possible combinations from one or more related items. For example, the meaning of “at least one of the first item, the second item, and the third item” means 2 among the first item, the second item, and the third item, as well as the first item, the second item, and the third item. It means a combination of all items that can be presented from more than one.

Before describing the apparatus for determining a descaling type according to an embodiment of the present invention, an overall flow of a rolling process to which the apparatus for determining a descaling type according to the present invention is applied will be described.

1 is a view showing the flow of a rolling process to which a descaling type determination apparatus according to an embodiment of the present invention is applied. As shown in Figure 1, the rolling process includes a heating furnace 110, rough rolling mill 120, edge heater 130, FSB 140, Finished Scale Breaking, finishing mill 150, runout table 160, winding It is performed through the odor 170, and SDD (180, Surface Defect Detector), in particular, the descaling type determination device 190 according to the present invention is the edge heater 130 and the FSB during the rolling process as shown in FIG. It is arranged between 140 and can determine the descaling type to be applied to the slab.

First, the slab is heated by the heating furnace 110 and then pressed through the rough rolling mill 120 to an appropriate thickness and width required by the finishing mill 150. Thereafter, the edge portion of the rolled slab is heated by the edge heater 130.

The slab heated by the edge heater 130 passes through the FSB 140 and the scale formed on the surface thereof is removed. The FSB 140 refers to a device for removing scale by spraying water after generating cracks by applying pressure to the scale formed on the surface of the slab. In one embodiment, the FSB 140 removes the scale by applying the FSB type determined by the descaling type determination apparatus 190 according to the present invention among a plurality of predetermined FSB types. In this case, the plurality of FSB types may be classified according to a position where water is to be sprayed on the surface of the slab and a pressure applied to the crack.

The slab from which the scale has been removed by the FSB 140 is manufactured in a final shape having the thickness or width required by the customer or the cold rolling process through the finish rolling machine 150, and then cooled to the target temperature on the runout table 160. . In one embodiment, the runout table may cool the slab of the final shape to a target temperature using laminar flow coolant.

After the slab cooled to the target temperature is wound in a coil shape through the take-up machine 170, the surface defect is inspected through the SDD 180. In FIG. 1, the SDD 180 is shown to be located on the outlet side of the take-up machine 170, but this is only an example, and the SDD 180 may be installed on the inlet side of the take-up machine 170. When the SDD 180 is located at the entrance side of the take-up machine 170, if a defect exists on the surface of the coil and the slab is determined to be abnormal, the slab can be supplied back to the FSB 140, thereby descaling operation. It can be done again.

The descaling type determination device 190 is disposed between the edge heater 130 and the FSB 140 to determine the descaling type to be applied to the slab. Specifically, the descaling type determination device 190 is applied to a target slab among a plurality of FSB types based on an image obtained by photographing the surface of a slab (hereinafter referred to as'target slab') (hereinafter referred to as a'target image'). Decide on the type of FSB to do.

In one embodiment, the descaling type determination apparatus 190 according to the present invention may determine the FSB type to be applied to the target image using the FSB type determination model generated by learning images photographed by slabs in advance. .

Hereinafter, the configuration of the descaling type determination apparatus according to the present invention will be described in more detail with reference to FIG. 2.

2 is a diagram showing a configuration of an apparatus for determining a descaling type according to an embodiment of the present invention. As shown in FIG. 2, the descaling type determining device 190 includes a photographing unit 210, a modeling unit 220, and an FSB type determining unit 270.

The photographing unit 210 photographs the surface of the slab and provides an image obtained by photographing the surface of the slab to the modeling unit 220 and the FSB type determination unit 270. The photographing unit 210 is disposed in the space between the edge heater 130 and the FSB 140, as shown in FIG. 1, so that the surface of the slab supplied from the edge heater 130 to the FSB 140 can be photographed. . In an embodiment, the photographing unit 210 may include a first photographing part for photographing a first surface of the slab and a second photographing part for photographing a second surface of the slab. In this case, each of the first photographing unit and the second photographing unit may include a plurality of cameras for photographing the first and second surfaces of the slab.

Hereinafter, the configuration of the photographing unit 210 according to the present invention will be described in more detail with reference to FIGS. 3 and 4.

3 is a perspective view showing the configuration of the photographing unit 210 according to the first embodiment of the present invention. As shown in FIG. 3, the photographing unit 210 may include a first photographing unit 213 and a second photographing unit 216.

The first photographing unit 213 is disposed on a first surface that is an upper surface of the slab to photograph the upper surface of the slab. In this case, the plurality of cameras 213a to 213d constituting the first photographing unit 213 may be sequentially arranged according to the traveling direction FD of the slab. Accordingly, when the first camera 213a photographs the first area D1 on the first surface of the slab, the second camera 213b will capture the first area D1 at a second view point after the first view point. Photographing is performed, and the third camera 213c photographs the first area D1 at a third viewpoint after the second viewpoint, and the fourth camera 213d photographs the first area D1 at a fourth viewpoint after the third viewpoint. One area D1 is photographed. That is, the first to fourth cameras 213a to 213d photograph the same area on the first surface of the slab at different viewpoints.

In FIG. 3, the first photographing unit 213 is illustrated as being composed of four cameras 213a to 213d, but this is only an example, and the first photographing unit 213 is composed of two or three cameras. It may consist of more than one camera.

The first to fourth cameras 213a to 213d constituting the first photographing unit 213 provide the captured slab image to the modeling unit 220 and the FSB type determining unit 230. In this case, each of the cameras may transmit an image captured by them together with an identification number assigned to them.

The second photographing unit 216 is disposed under the second surface that is the lower surface of the slab to photograph the lower surface of the slab. In this case, a plurality of cameras 216a to 216d constituting the second photographing unit 216 may also be sequentially arranged according to the traveling direction FD of the slab. The fifth to eighth cameras 216a to 216d constituting the second photographing unit 216 have different viewpoints in the same manner as the first to fourth cameras 213a to 213d constituting the first photographing unit 213. Thus, the same area can be photographed on the second surface of the slab, and may be composed of two or three cameras, or five or more cameras.

The fifth to eighth cameras 216a to 216d constituting the second photographing unit 216 provide the captured slab image to the modeling unit 220 and the FSB type determining unit 270. In this case, each of the cameras may transmit an image captured by them together with an identification number assigned to them.

3 shows that the first photographing unit 213 photographs the upper surface of the slab and the second photographing unit 216 photographs the lower surface of the slab. However, the first photographing unit 213 is disposed on the lower surface of the slab and The lower surface of the image may be photographed, and the second photographing unit 216 may be disposed on the upper surface of the slab to photograph the upper surface of the slab.

4 is a perspective view showing the configuration of a photographing unit according to a second embodiment of the present invention. As shown in FIG. 4, the photographing unit 210 according to the second embodiment includes a first photographing unit 213 for photographing a first surface of a slab and a second photographing unit 216 for photographing a second surface of the slab. ).

The first photographing unit 213 includes first and second cameras 213a and 213b sequentially arranged according to the traveling direction of the slab, and third and fourth cameras 213c arranged in an inclined direction in the traveling direction of the slab. 213d). Accordingly, the first and second cameras 213a and 213b photograph the first surface of the slab in a direction perpendicular to the first surface of the slab, and the third and fourth cameras 213c and 213d are 1 The first surface of the slab is photographed in a direction inclined to the surface.

As described above, according to the second embodiment of the present invention, some of the cameras constituting the first photographing unit 213 photograph the first surface of the slab in a direction perpendicular to the first surface of the slab. Accurately photographing the height of the scale existing on the first surface of the slab by making it possible to accurately acquire the pattern of the scale existing on the surface, and photographing the first surface of the slab in a direction inclined to the first surface of the slab This makes it possible to more accurately determine the amount of scale.

4 shows that only two of the four cameras of the first photographing unit are photographed in a vertical direction and only two are photographed in an inclined direction, but the number of cameras photographed in a vertical direction and the number of cameras photographed in an inclined direction are It will be changeable.

The second photographing unit 216 includes fifth and sixth cameras 216a and 213b sequentially arranged according to the traveling direction of the slab, and seventh and eighth cameras 216c arranged in an inclined direction in the traveling direction of the slab. 216d). Accordingly, the fifth and sixth cameras 216a and 216b photograph the second surface of the slab in a direction perpendicular to the second surface of the slab, and the seventh and eighth cameras 216c and 216d are 2 The second surface of the slab is photographed in a direction inclined to the surface.

As described above, according to the second embodiment of the present invention, some of the cameras constituting the second photographing unit 213 allow the second surface of the slab to be photographed in a direction perpendicular to the second surface of the slab. Accurately photographing the height of the scale existing on the second surface of the slab by allowing the pattern of the scale existing on the surface to be accurately acquired, and allowing the second surface of the slab to be photographed in an inclined direction to the second surface of the slab. This makes it possible to more accurately determine the amount of scale.

4 shows that only two of the four cameras of the second photographing unit are photographed in a vertical direction and only two are photographed in an inclined direction, but the number of cameras photographed in a vertical direction and the number of cameras photographed in an inclined direction are It will be changeable.

Referring back to FIG. 2, the modeling unit 220 generates an FSB type determination model by learning images previously captured by the photographing unit 210 using a predetermined learning algorithm. To this end, the modeling unit 220 includes an image collecting unit 230, a model generating unit 240, and a database 245 as shown in FIG. 2. The modeling unit 220 may further include an accuracy calculation unit 250 and a weight calculation unit 260.

The image collecting unit 230 collects an image captured by the photographing unit 210 for a predetermined period. In an embodiment, the images collected by the image collection unit 230 may select images that are determined to be normal as a result of SDD determination among images captured by the photographing unit 210 as training images. At this time, the training image includes the FSB type actually applied to the slab corresponding to the image.

The model generation unit 240 generates the FSB type determination model by learning the training images selected by the image collection unit 230 using a predetermined learning algorithm. As an example, the model generator 240 uses a convolution neural network (CNN) deep learning algorithm suitable for image learning to determine the amount of scale included in the plurality of training images, the pattern of the scale, and the FSB type actually applied for each training image. By learning, it is possible to create an FSB type determination model to determine the FSB type to be applied to a specific image.

The model generation unit 240 stores the generated FSB type determination model in the database 245.

The accuracy calculation unit 250 calculates the accuracy of the FSB type determination model generated by the model generation unit 240 for each camera included in the first and second photographing units 213 and 216. To this end, the image collection unit 230 may select some of the selected training images as a test image, and the accuracy calculation unit 250 may calculate the accuracy of the FSB type determination model for each camera using the test image.

Specifically, the accuracy calculation unit 250 inputs a plurality of test images selected by the image collection unit 230 into the FSB type determination model, and determines the FSB type obtained and the FSB type actually applied to the slab corresponding to the test image. By comparing, the accuracy of each camera is calculated.

The accuracy calculation unit 250 stores the calculated accuracy for each camera in the database 245.

The weight calculation unit 260 calculates a weight for each camera based on the accuracy for each camera calculated by the accuracy calculation unit 250.

In an embodiment, the weight calculator 260 may calculate a weight for each camera using 1 below.

는 i개의 카메라 중 k번째 카메라의 가중치를 나타내고

는 k번째 카메라의 정확도를 나타낸다.In Equation 1

Represents the weight of the k-th camera among i cameras

Represents the accuracy of the k-th camera.

The weight calculator 260 stores the calculated weight for each camera in the database 245.

As described above, according to the present invention, since the modeling unit 220 can generate the FSB type determination model for determining the FSB type to be applied to the slab corresponding to the target image photographed by the photographing unit 210, the target It is possible to accurately determine the FSB type to be applied to the target slab with only the target image taken of the slab.

When the target image is input from the photographing unit 210, the FSB type determination unit 270 inputs the input target image to the FSB type determination model generated by the modeling unit 220 to be applied to the slab corresponding to the target image. Determine the FSB type.

Specifically, the FSB type determination unit 270 inputs each of the first target images in which the cameras photographed the same area of the first surface of the slab from each camera constituting the first photographing unit 213 By inputting the first target images for each camera into the FSB type determination model, the FSB type to be applied to the region of the slab corresponding to the first target images for each camera is obtained. The FSB type determination unit 270 determines any one of the obtained FSB types as the final FSB type to be applied to the area of the slab.

In an embodiment, the FSB type determination unit 270 may determine a final FSB type using the weights for each camera calculated by each modeling unit 220. Specifically, the FSB type determination unit 270 determines the FSB type for each first target image of each camera, and determines the weight of the camera as the weight of the FSB type determined for the first target image of the corresponding camera. The FSB type determination unit 270 adds the weights given to the same FSB type to determine the FSB type having the highest weight as the final FSB type to be applied to the region of the slab corresponding to the first target images for each camera.

Hereinafter, an example of the FSB type determined by the FSB type determination unit 270 is shown in FIG. 5.

Referring to FIG. 5, the first FSB type is applied when the scale of the first reference amount or more exists only on the first surface of the slab, and the second FSB type is applied when the scale of the first reference amount or more exists only on the second surface of the slab, The third FSB type is applied when a scale equal to or greater than the first reference amount exists on both the first surface and the second surface of the slab, and the fourth FSB type is applied to both the first surface and the second surface of the slab. It is applied when a scale greater than or equal to the reference amount exists, and the fifth FSB type is applied when the scale of the first reference amount or more exists in a predetermined pattern only on the first surface of the slab, and the sixth FSB type is applied only to the second surface of the slab. It is applied when the scale of the reference amount or more exists in a predetermined pattern, and the seventh FSB type is applied when the scale of the first reference amount or more exists in a predetermined pattern on both the first surface and the second surface of the slab. In this case, the first surface may be the upper surface of the slab and the second surface may be the lower surface of the slab. Conversely, the first surface may be a lower surface of the slab and the second surface may be an upper surface of the slab.

Hereinafter, an example in which the FSB type determination unit 270 selects the FSB type to be applied to the slab in consideration of the weight and FSB type for each camera will be described with reference to FIG. 6.

6 is a table showing an example in which the FSB type determination unit 270 according to an embodiment of the present invention selects a final FSB type applied to a target slab. 6A, the weight of the first camera calculated by the modeling unit 220 is 0.4, and the FSB type of the first camera determined by the FSB type determination unit 270 is the first FSB type, and the modeling unit The weight of the second camera calculated by 220 is 0.3, and the FSB type of the second camera determined by the FSB type determination unit 270 is the second FSB type, and the third camera calculated by the modeling unit 220 The weight of is 0.2, the FSB type of the third camera determined by the FSB type determination unit 270 is the second FSB type, and the weight of the fourth camera calculated by the modeling unit 220 is 0.1, and the FSB type determination unit ( The FSB type of the fourth camera determined by 270) is the third FSB type.

Therefore, the FSB type determination unit 270 adds the weights assigned to the same FSB type as shown in FIG. 6B, so that the weight of the first FSB type is 0, 4, the weight of the second FSB type is 0.5, and the third FSB The weight of the type is 0.1. The FSB type determination unit 270 determines the FSB type having the highest weight among the summed weights as the final FSB type to be applied to the slab, and thus determines the second FSB type as the final FSB type to be applied to the slab.

Hereinafter, a method of determining a descaling type according to the present invention will be described with reference to FIGS. 7 to 8.

7 is a flowchart showing a method of determining a descaling type according to an embodiment of the present invention. The descaling type determination method illustrated in FIG. 7 may be performed by a descaling type determination apparatus having a configuration as illustrated in FIG. 2.

First, the descaling type determination apparatus acquires a target image photographing the surface of the target slab for each camera by using a plurality of cameras (S700). In this case, the target image may include a first target image obtained by photographing the first surface of the target slab and a second target image obtained by photographing the second surface of the target slab.

At this time, the plurality of cameras that photograph the first surface of the target slab photograph the same area on the first surface of the slab with different viewpoints, and the plurality of cameras photographing the second surface of the target slab are The same area is photographed on the second surface of the slab.

In one embodiment, as shown in FIG. 4, some of the plurality of cameras are sequentially arranged according to the traveling direction of the slab, and the first surface or the second surface of the target slab in a direction perpendicular to the first surface of the slab And the remaining cameras may photograph the first surface or the second surface of the target slab in a direction inclined to the first surface or the second surface of the slab according to the moving direction of the slab.

Referring back to FIG. 7, the apparatus for determining the descaling type inputs the target image acquired for each camera into the FSB type determination model and determines the FSB type for each camera (S710).

In this case, the FSB type determination model may be generated by learning a plurality of previously acquired training images. Hereinafter, a method of generating an FSB type determination model by the descaling type determination apparatus will be described in detail with reference to FIG. 8.

8 is a flowchart illustrating a method of generating an FSB type determination model according to an embodiment of the present invention.

First, as shown in FIG. 8, the apparatus for determining a descaling type collects a learning image for a predetermined period (S800).

In one embodiment, the training image may be selected as images determined as normal as a result of SDD determination. At this time, the training image includes the FSB type actually applied to the slab corresponding to the image.

Thereafter, the apparatus for determining the descaling type generates an FSB type determination model by learning the learning image using a predetermined learning algorithm (S810).

As an example, the descaling type determination device is applied to a specific image by learning the amount of scale included in the plurality of training images and the pattern of the scale and the actual applied FSB type for each training image using a CNN Deep Learning algorithm suitable for image learning. An FSB type determination model can be created to determine the FSB type to be performed.

Thereafter, the descaling type determination apparatus inputs a plurality of predetermined test images into the FSB determination model to calculate accuracy for each camera (S820). In an embodiment, the apparatus for determining the descaling type may select some of the training images as a test image.

Specifically, the descaling type determination device compares the FSB type obtained by inputting a plurality of test images into the FSB type determination model generated in S810 with the FSB type actually applied to the slab corresponding to the test image to determine the accuracy of each camera. Calculate.

Thereafter, the apparatus for determining the descaling type calculates a weight of each camera based on the accuracy of each camera calculated in S820 (S830).

In an embodiment, the apparatus for determining a descaling type may calculate a weight for each camera using Equation 1 described above. Since the method of calculating the weight for each camera by the descaling type determination apparatus using Equation 1 has been described in Equation 1, a detailed description thereof will be omitted.

As described above, the descaling type determination apparatus according to the present invention generates an FSB type determination model based on training images acquired by photographing slabs, and calculates the accuracy and weight for each camera based on the test image. Not only can the type be automatically determined, but also the accuracy of FSB type determination can be improved.

Referring back to FIG. 7, the apparatus for determining the descaling type determines the FSB type to be applied to the target slab corresponding to the target image by using a weight predetermined for each camera (S720). In this case, the weight for each camera may be determined by the process S830 as described above.

Specifically, the apparatus for determining the descaling type determines the FSB type for each target image of each camera, and determines the weight of the camera as the weight of the FSB type determined for the target image of the camera. The apparatus for determining the descaling type determines the FSB type having the highest weight as the final FSB type to be applied to the target slab by summing the weights assigned to the same FSB type.

Thereafter, the descaling type determination apparatus transfers the FSB type determined through S720 to the FSB 140 of FIG. 1 so that the FSB 140 performs descaling to the FSB type determined by the descaling type determination apparatus.

Those skilled in the art to which the present invention pertains will appreciate that the above-described present invention can be implemented in other specific forms without changing the technical spirit or essential features thereof.

Therefore, it should be understood that the embodiments described above are illustrative in all respects and not limiting. The scope of the present invention is indicated by the claims to be described later rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. do.

190: descaling type determination device 210: photographing unit
213: first photographing unit 216: second photographing unit
220: modeling unit 230: image collection unit
240: model generation unit 250: accuracy calculation unit
260: weight calculation unit 270: FSB type determination unit

Claims (13)

A photographing unit comprising a plurality of cameras for photographing the surface of the slab;
Using a predetermined learning algorithm, the amount of scale included in the learning images acquired by the photographing unit for a predetermined period and the pattern of the scale, and the FSB (Finished Scale Breaking) type actually applied to the slabs corresponding to each learning image are determined. A modeling unit that learns and generates an FSB type determination model; And
The FSB type is determined for each target image using the amount of scale and the scale pattern included in the target images acquired by the camera photographing the target slab and the FSB type determination model, An apparatus for determining a descaling type based on a slab image, comprising: an FSB type determining unit that determines one FSB type to be applied to the slab. delete The method of claim 1,
The modeling unit obtains a test FSB type by inputting the test image acquired by each camera into the FSB type determination model, and based on a comparison result of the acquired test FSB type and the FSB type actually applied to the slab corresponding to the test image. And further includes a weight calculator for calculating the weight of each camera,
The FSB type determination unit determines the single FSB type by reflecting a weight of each camera. The method of claim 3,
The weight calculation unit, Equation

Calculate the weight of each camera using
In the above equation

Represents the weight of the k-th camera among i cameras

Is a slab image-based descaling type determination device, characterized in that it indicates the accuracy of the k-th camera among i cameras.

The method of claim 3,
The FSB type determination unit,
The weight calculated for each camera is determined as the weight of the FSB type calculated for each camera, and the FSB type with the highest weight is determined as the one FSB type by summing the weights assigned to the same FSB type. Slab image-based descaling type determination device. The method of claim 1,
The photographing unit,
A first photographing unit comprising a plurality of cameras for photographing the first surface of the slab; And
And a second photographing unit comprising a plurality of cameras for photographing the second surface of the slab. The method of claim 1,
The slab image-based descaling type determination apparatus, wherein the plurality of cameras are sequentially arranged according to the moving direction of the slab. The method of claim 1,
At least one first camera among the plurality of cameras photographs the surface of the slab in a direction perpendicular to the surface of the slab, and the second cameras other than the first camera among the plurality of cameras are on the surface of the slab. Slab image-based descaling type determination apparatus, characterized in that photographing the surface of the slab in an inclined direction. The method of claim 1,
The FSB type is a first FSB type applied when a scale of a first reference amount or more exists only on the first surface of the slab, a second FSB type applied when a scale of a first reference amount or more exists only on the second surface of the slab, A third FSB type applied when a scale equal to or greater than the first reference amount exists on both the first surface and the second surface, and when there is a scale greater than or equal to the second reference amount greater than the first reference amount on both the first and second surfaces A fourth FSB type that is applied, a fifth FSB type that is applied when a scale of the first reference amount or more exists in a predetermined pattern only on the first surface, and when a scale of the first reference amount or more is applied only on the second surface in a predetermined pattern Slavic image-based D, characterized in that it includes at least one of a sixth FSB type that is applied, and a seventh FSB type that is applied when a scale equal to or greater than the first reference amount exists in a predetermined pattern on both the first surface and the second surface. Device for determining the scaling type. The method of claim 1,
The learning images are a slab image-based descaling type determination apparatus, characterized in that consisting of data determined to be normal as a result of a surface defect detector (SDD) determination. Photographing the surface of the slab with a plurality of cameras;
Generating an FSB type determination model by modeling an amount of scale and a pattern of scale included in the training images acquired during a predetermined period of time, and a Finished Scale Breaking (FSB) type actually applied to the slab corresponding to each training image;
Inputting test images acquired during a predetermined period into the FSB type determination model to obtain weights for each camera;
Determining an FSB type for each camera by inputting target images obtained by photographing a target slab with the plurality of cameras into the FSB type determination model; And
And determining one FSB type to be applied to the target slab from among the FSB types determined for each camera based on the weight. The method of claim 11,
In the step of obtaining the weight, the equation

Calculate the weight for each camera by using,
In the above equation

Represents the weight of the k-th camera among i cameras

Is a slab image-based descaling type determination method, characterized in that it represents the accuracy of the k-th camera among i cameras.
The method of claim 11,
The step of determining the one FSB type,
The weight calculated for each camera is determined as the weight of the FSB type calculated for each camera, and the FSB type with the highest weight is determined as the one FSB type by summing the weights assigned to the same FSB type. Descaling type determination method based on slab image.

Images