0001]The present invention relates to a slag detection method in the molten steel flow.
　Priority is claimed on Japanese Patent Application No. 2017-025441 filed in Japan on February 14, 2017, which is incorporated herein by reference.
BACKGROUND
[0002]When the converter of steel out of the ladle, it is common to flow out molten steel flow toward the ladle from the converter by tilting the converter. At this time, the slag was remaining converter furnace, it is ideal to flow out into the ladle the molten steel only from the converter. However, the molten steel stream flowing toward the ladle from a converter furnace, although only substantially molten steel is present in the tapping early, are there the molten steel and slag in the tapping end of tapping metaphase There is common. Therefore, the attempt to suppress the outflow of slag, the residual amount of the molten steel in the converter in increases, the yield tends to be low.
[0003]
　On the other hand, an attempt to reduce the residual amount of the molten steel in the converter in order to slag with the molten steel flows out toward the ladle, so that the slag is often present in the ladle. As a result, there is a possibility that boiling over may occur slag from ladle, such as the components out of the molten steel is generated in the secondary refining process is post-process problems.
[0004]
　Therefore, it quantifies the outflow of slag and detects slag of the molten steel stream flowing toward the ladle from a converter furnace, controlling the slag outflow in the range required in the tapping operation of the converter it has been desired.
[0005]
　Higher than emissivity emissivity of the molten steel slag, when imaging the flow of molten steel, in the site where the slag is present, is brightly imaged than the site of the only molten steel slag is not present. In other words, the concentration of the pixel area corresponding to the slag in the captured image obtained by imaging the flow of molten steel (gray level) is greater than the density of the pixel area corresponding to the molten steel. As a technique for detecting slag using this principle, for example, a method described in Patent Documents 1 and 2.
[0006]
　Patent Document 1 captures molten steel stream (tap stream), depending on whether is in the range of concentrations that pixel area of the captured image is preset manually (gray level), slag or a pixel corresponding to the molten steel determines whether the pixel corresponding to the number of pixels corresponding to the slug, teeth value obtained by dividing the sum of a predetermined the number of pixels corresponding to the number and the molten steel of a pixel corresponding to slag It discloses a method to stop tapping when exceeded threshold.
　That is, in the above-described method of Patent Document 1, in the captured image, by pixel regions within a concentration range previously manual setting is determined to be a pixel area corresponding to the slag, detects the slag in the molten steel flow .
[0007]
　However, the temperature of the molten steel flow, depending on the conditions of the steel grade or tapping operations, also changes, for example 100 ° C. or higher. In this case, in the captured image, and also it varies greatly density of a pixel area corresponding to the molten steel and slag. Thus, in the above method of Patent Document 1, since using a fixed value as the predetermined threshold value, if the conditions of the tapping operation of the converter is changed, it is difficult to accurately detect the slag.
[0008]
　Patent Document 2, for the captured image obtained by imaging the flow of molten steel, the concentration (intensity) and the horizontal axis, to create a concentration (luminance) histogram on the vertical axis the number of pixels, using the density histogram slag It discloses a method of detecting. Specifically, in the method of Patent Document 2, in the density histogram, it regarded as the number of pixels is the maximum peak point is the maximum (the maximum peak position) corresponds to the molten steel, considering the direction of the axis of abscissas of the maximum peak point variation σ with determining the density value (luminance value) N1 or more pixels with the molten steel, and the density value obtained by adding the bias value B to the density value N1 (luminance value) N2 or more pixels determined to slag.
[0009]
　According to the method of Patent Document 2, in the captured image, if the changes greatly depending on the density of the pixel area corresponding to the molten steel and slag for example steel type, also changes the maximum peak point, corresponding to the slag in accordance with the change concentration values ​​N2 for detecting a pixel region is considered will be automatically changed. Therefore, according to the method of Patent Document 2, it believed that the above-described problem that can occur when using a fixed threshold value (concentration range) as above method of Patent Document 1 can be solved to some extent It is.
[0010]
　However, the present inventors have studied, not necessarily the maximum peak point is necessarily correspond to the molten steel in the density histogram, it was found that may correspond to the slag. Therefore, is regarded as the maximum peak point is always correspond to the molten steel, in Patent Document 2 above method of determining the concentration value N2, it is difficult to accurately detect the slag.
CITATION
Patent Document
[0011]
Patent Document 1: Japanese Patent 2001-107127 JP
Patent Document 2: Japanese Patent 2006-213965 JP
Summary of the Invention
Problems that the Invention is to Solve
[0012]
　The present invention has been made in view of the above circumstances, even if the conditions of tapping operations in the converter or the like is changed, slag accurately detectable molten steel flow in the slag in the molten steel flow and an object thereof is to provide a detection method.
Means for Solving the Problems
[0013]
　To solve the above problems, the present inventors have conducted extensive studies.
　First, the present inventors, using a thermal imaging camera (thermography) having a main sensitivity to the infrared region as an imaging device, tapping initial images the tapping metaphase, and various molten steel flow across the tapping end to obtain a large number of captured images. Then, for each of these captured images, the temperature on the horizontal axis, was to create a histogram with the vertical axis the number of pixels, for example in a temperature range of the horizontal axis of 1000 ~ 2000 ° C., the number of pixels on the vertical axis maximum value may the maximum peak point exists is, the maximum peak point in some cases located on the low temperature side was found that the maximum peak point is also located on the high temperature side.
[0014]
　Then, the present inventors have found that when there is only the case or very small amounts there is no slag in the molten steel flow (e.g., tapping initial-tapped metaphase) to, in the histogram, the maximum peak point corresponds to the molten steel and with finding it is positioned on the low temperature side, if the slag is present in large amounts (for example, if the slag is present in the molten steel flow to the extent that easy visually observed in the tapping end), the maximum peak point and in response to the slag it was found to be located on the high temperature side. However, as mentioned above, the temperature of the molten steel flow varies depending on the conditions of the steel grade or tapping operations. Therefore, using a fixed threshold, by determining whether the maximum peak point is located on either the low temperature side and high temperature side threshold value of the fixed, the maximum peak point corresponds to either of the molten steel and slag when determining whether to be, it is difficult to accurately detect the slag.
[0015]
　Accordingly, the present inventors have a result of further intensive studied, when the slag in case or the molten steel flow no slag in the molten steel flow is not only an extremely small amount exists, the temperature variation of the maximum peak point is small (for example, temperature change per 0.2 seconds has been found to be a) that less than 100 ° C.. Specifically, molten steel is refined in a converter is generally because they are stirred by the top-blown or upper bottom blown blowing in a converter furnace during smelting, the temperature variation is small. Therefore, molten steel flow flowing out of the converter has a small temperature variation, for example, a temperature change per 0.2 seconds is less than 100 ° C..
[0016]
　Then, the present inventors have found that compared to the molten steel slag white degree (brightness) is high (compared to the emissivity of the molten steel slag emissivity is approximately 1.5 times) in order, the emissivity of the molten steel in case of measuring the temperature, the temperature of the slag apparent than real temperature is at least 100 ° C. or higher, focused on the phenomenon to be detected in many hundreds ° C. about the case (for example, 300 ~ 600 ° C.) higher temperatures. By utilizing the phenomenon, the actual temperature is comparable molten steel and slag (i.e., flowing out from the converter, molten steel stream comprising molten steel and slag) will be remarkable the relative temperature difference between the apparent molten steel and slag it can, by using the relative temperature difference between the apparent found that the slag in the molten steel flow can be accurately detected.
[0017]
　Based on the above findings, the present invention employs the following in order to solve the above problems.
　(1) slag detection method of the molten steel flow in accordance with one embodiment of the present invention, flows toward the ladle from a converter furnace, imaging by sequentially imaging the molten steel stream comprising molten steel and slag to obtain a plurality of captured images for the concentration parameter corresponding to the density of the pixels constituting the captured images on the horizontal axis, the total number at which the histogram each captured image to the longitudinal axis the number of pixels of the pixels having the concentration parameter; step and for each histogram, the maximum peak point detection step and said number of pixels to detect the maximum peak point is the maximum value; histogram generation process and to create the maximum peak point of each histogram, in any of the slag or the molten steel maximum of the histogram of have, the maximum peak point n th at the type judgment step (n ≧ 2) captured image; corresponding whether the maximum peak point type determination step of determining Over click point P n when determining: maximum peak point P in the histogram of the n-1 th captured image n-1 concentration parameter T n-1 with respect to the maximum peak point P n concentration parameter T of n changes in If the amount ΔT is equal to or greater than a predetermined value, the maximum peak point P n is determined corresponding to the slag; while the variation ΔT is of less than the predetermined value, the maximum peak point is determined to the molten steel j maximum peak point P in the histogram of the captured image of th (j ≦ n-1) j concentration parameter T of j the density parameter T for n variation [Delta] T of 'if said predetermined value or more, the maximum peak point P n is determined corresponding to the slug, the variation [Delta] T' if it is less than the predetermined value, the maximum peak point P n is determined that corresponds to the molten steel.
　(2) In the aspect described in (1) above, wherein the maximum peak point type determination step, the maximum peak point P j as being acquired before the n-th captured image and the n th imaging maximum peak point may be used a maximum peak point in the histogram of the determination captured image and the molten steel most acquisition sequence with close to an image.
　In the aspect described in (3) above (1) or (2), when the maximum peak point in the maximum peak point type determination step determines that corresponding to the slag was determined based on the maximum peak point pixels having a density parameter of less than the first threshold value corresponds to the molten steel, a pixel having a density parameter of more than the first threshold value and a first determination step of determining a corresponding to the slug, the maximum peak point when the maximum peak point in the type determination step determines that corresponding to the molten steel, a pixel having a density parameter of the following second threshold determined based on the maximum peak point corresponds to the molten steel, the second it may further have a; a pixel having a larger concentration parameter than second threshold and a second determination step of determining a corresponding to the slug.
　(4) In the aspect described in (3) above, it may be configured as follows: the first threshold value, in the histogram, first straight line having the maximum peak point and a positive slope as the in is represented, the second threshold value, in the histogram, the expressed maximum peak point in a second straight line having a street and negative slope, the absolute value of the slope of the second straight line, said first larger than the absolute value of a straight line of slope.
　(5) In the aspect described in the above (4), configuration may be as follows: the first straight line, said and said maximum peak point has the number of pixels smaller than the number of pixels threshold among the points having a small concentration parameter greater than a predetermined value relative to the concentration parameter, and the maximum peak point density parameter is a straight line passing through the said maximum peak point, the absolute value of the slope of the second straight line, said first 1 is 1.5 to 2.5 times the absolute value of the slope of the line.　
The invention's effect
[0018]
　According to the above aspect of the present invention, even when the conditions of the tapping operation in the converter or the like is changed, it is possible to accurately detect slag in the molten steel flow.
BRIEF DESCRIPTION OF THE DRAWINGS
[0019]
It is a schematic diagram showing a schematic configuration of a slag detection device used in slag detection method according to an embodiment of FIG. 1 the present invention.
FIG. 2 is a flowchart showing a schematic procedure of the slag detection process.
3 is a flowchart showing a schematic procedure of the maximum peak point type determination step ST4 shown in FIG.
It is a diagram illustrating an example of a captured image acquired in the imaging step ST1 shown in FIG. 4A] FIG.
The histogram generation step ST2 shown in FIG. 4B] FIG. 2 is a diagram showing a histogram created based on the captured image of FIG. 4A.
An example of a captured image acquired in the imaging step ST1 shown in FIG. 5A] FIG. 2 is a diagram illustrating different examples and Figure 4A.
The histogram generation step ST2 shown in FIG. 5B] FIG. 2 is a diagram showing a histogram created based on the captured image of FIG. 5A.
6 is a diagram for explaining an example of a method for determining the predetermined value Th1.
7 is a diagram for explaining a maximum peak point type determination step ST4 shown in FIG.
Is a diagram for explaining a first threshold value which is determined in the first determination step ST5 shown in FIG. 8 Fig.
An example of a captured image acquired in the imaging step ST1 shown in FIG. 9A] FIG. 2 is a diagram illustrating different examples and FIGS. 4A and 5A.
The histogram generation step ST2 shown in FIG. 9B] FIG. 2, a diagram showing a histogram created based on the captured image of FIG. 9A, illustrating a second threshold value which is determined in the second determination step ST6 it is a diagram for.
[FIG. 10A] is a diagram for explaining a slag detection method described in Patent Document 2.
It is a diagram showing a captured image used to create the histogram of FIG. 10B] FIG 10A.
DESCRIPTION OF THE INVENTION
[0020]
　Hereinafter, with reference to the accompanying drawings, slag detection method of the molten steel flow in accordance with an embodiment of the present invention (hereinafter, simply referred to as "slag detection method") will be described. Incidentally, omitted in the present specification and drawings, are denoted by the same reference numerals to constituent elements having substantially the same functional configurations, and their redundant description.
　First, the configuration of the slag detection device 100 used in slag detection method according to the present embodiment.
[0021]
　
　Fig. 1 is a schematic diagram showing a schematic configuration of the slag detecting device 100. In FIG. 1, a converter 3 which accommodates the molten steel M and slag S is shown in cross-section.
　As shown in FIG. 1, the slag detection device 100, when the steel exits from the converter 3 into ladle 4, molten steel flow F flowing out toward the ladle 4 from tapping nozzle 31 of the converter 3 is tilted used to detect slag S in.
[0022]
　Here, the state before tapped from the converter 3 into ladle 4 (i.e., the state before tilting the converter 3), the in converter furnace 3, the molten steel M and slag S are separated to some extent ( slag S is positioned on the surface of the molten steel M of the converter 3). Therefore, tilting of the converter 3, the molten steel M from the tapping port 31 provided on a side surface of the converter 3 is first outlet, after which the slag S from the point in time starts to flow out. That is, the molten steel flow F in tapping early consists primarily of molten steel M, for example molten steel flow F in tapping Middle to tapping end comprises slag S and the molten steel M.
[0023]
　Slag detection device 100, connected to the imaging unit 1 substantially taken from the horizontal direction of molten steel flow F flowing substantially vertically towards the ladle 4 from tapping nozzle 31 of the converter 3, in the imaging device 1 images and a processing unit 2.
[0024]
　The imaging unit 1, for example, can be used such as a CCD camera having a main sensitivity to thermal imaging camera (thermography), or visible light region having a main sensitivity to the infrared region. These thermal imaging camera (thermography) and the CCD camera, for example, can be used commercially.
　In the present embodiment, as the imaging device 1 uses a thermal imaging camera having a main sensitivity to the infrared region. In the case of using a thermal image camera (thermography) as in the present embodiment, it is possible to calculate the value of the temperature or concentration of the pixel region in the captured image (density before converting to temperature). On the other hand, when using a CCD camera, it is possible to calculate the value of the density of the pixel area.
[0025]
　Image processing means 2, for example, a general-purpose personal computer in which a predetermined program for executing such histogram creation step ST2 to be described later is installed. The image processing unit 2 includes a monitor for displaying a captured image obtained by the imaging unit 1.
[0026]
　Slag detection method according to the present embodiment is performed using the slag detection device 100. The following describes slag detection method according to the present embodiment.
[0027]
　
　FIG. 2 is a flowchart showing a schematic procedure of the slag detecting method according to the present embodiment. 3 is a flowchart showing a schematic procedure of the maximum peak point type determination step ST4 shown in FIG.
　Slag detection method according to the present embodiment, flows toward the ladle 4 from the converter 3 on the basis of the flow of molten steel F containing molten steel M and slag S in the plurality of captured images obtained by sequentially captured by the imaging unit 1 , a method of detecting slag S in the molten steel flow F, as shown in FIG. 2, an imaging step ST1, the histogram generation step ST2, the maximum peak point detection step ST3, the maximum peak point type determination step ST4 , has a first determination step ST5, and a second determination step ST6.
　Hereinafter, the contents of each process will be sequentially described.
[0028]
　(Imaging step ST1)
　In the imaging step ST1, the imaging unit 1, to obtain a plurality of captured images by sequentially capturing the molten steel flow F flowing out toward the ladle 4 from the converter 3 (see FIG. 1).
　In this embodiment, the use of the thermal imaging camera as the imaging device 1, the captured image acquired by the imaging step ST1 is to that in terms of temperature the density of each pixel constituting the captured image with a predetermined conversion formula Become. In other words, captured image obtained by the imaging step ST1 has a value of temperatures detected for each pixel.
[0029]
　Field of the imaging means 1, so as not to be affected by the outflow position and extent variations of molten steel flow F, is set to a wide field of view to include a background not molten steel flow F only. Even if the field of view of the imaging means 1 is configured to be in the background, because the temperature of the background is lower than the temperature of the molten steel flow F, at the maximum peak point detection step ST3 will be described later, the pixel region corresponding to the molten steel flow F it is possible to identify a pixel region corresponding to the background with. Note that the field of view of the imaging unit 1, only the molten steel flow F may be previously narrowed adjusted to be imaged. However, the outflow position and extent of the molten steel flow F, in accordance with the tilt angle of the converter 3 (depending on the position of the tapping nozzle 31), it is common to vary to some extent. Therefore, tapping initial, tapping metaphase and that only the molten steel flow F in any of tapping the end to adjust the field of view of the imaging unit 1 to be imaged requires labor work. Accordingly, the field of view of the imaging means 1 is preferably set to a wide field of view to include a background.
[0030]
　The imaging timing of the imaging device 1, at least one of the plurality of captured images acquired, it is necessary is obtained by imaging the molten steel flow F in which the molten steel M mainly. That is, at least one needs to be captured in tapped initial. The reason for this will be described later. Other imaging timing is not particularly limited, in enhancing the time resolution to detect the slag S is continuously captured for each scan period is set to the imaging unit 1 (the inverse of the frame rate) it is preferable.
　Incidentally, relates to an imaging timing of the above, the slag amount and the amount of molten steel in the converter 3 is be estimated, further, whether it caused to what extent tilt the converter 3 molten steel flow F in which the molten steel M mainly flows out geometrically possible estimate. Further, it is possible to confirm whether the molten steel flow F visually are those consisting mainly of molten steel M. Based on these, is discharged from the tapping nozzle 31, it is possible to image the molten steel flow F in which the molten steel M mainly.
　A plurality of captured images obtained by the imaging means 1 is stored in the image processing unit 2.
[0031]
　(Histogram creation step ST2)
　concentration in the histogram generating step ST2, the image processing unit 2, an image processing on each of the plurality of captured images acquired by the imaging step ST1, corresponding to the density of the pixels constituting the captured images the parameter on the horizontal axis, to create a histogram of the number of pixels is a total number of pixels having the concentration parameter and the vertical axis. It histogram may be created for each of captured images may be generated for average image obtained by averaging a plurality of consecutive captured images. Note that the average case of using the image pickup means 1 of the molten steel flow length L of the molten steel flow F velocity V in a division-obtained time of the pixel region corresponding to F in the field of view (= L / V) in in it is desirable to average a plurality of consecutive captured images.
　Examples of the concentration parameter, other concentrations themselves, can be exemplified temperature. When the imaging unit 1 as in the present embodiment is a thermal imaging camera, the horizontal axis can be created histograms is the temperature or concentration (concentration prior to conversion to the temperature). On the other hand, if the imaging means 1 is a CCD camera, the horizontal axis can be created histograms is the concentration.
[0032]
　Here, the "density" herein, for example, refers to the 256 gradations of the image brightness (i.e., luminance of the image). Then, a the concentration, the relationship between the thermal radiance of the molten steel flow is a linear relationship.
[0033]
　Due to the use of thermal imaging camera as the imaging unit 1 in the present embodiment as described above, using a temperature as above concentration parameter (i.e., in this embodiment, the horizontal axis of the histogram is temperature).
　In the present embodiment, as described above, the field of view of the imaging means 1 is set to also include a background not molten steel flow F only. Therefore, when creating the histogram, the image processing means 2, in the captured image, a predetermined threshold (e.g., 1000 ° C.) pixel area is determined as a pixel area corresponding to the flow of molten steel F having a temperature above , a histogram as an object the pixel region (i.e., the temperature is horizontal axis pixel area of less than the predetermined threshold value is not used for creating the histogram). Thus, it is possible to avoid the influence of the background on the histogram (the number of pixels corresponding to the background is not a maximum).
　The image processing unit 2 creates a histogram for the entire captured image including also the pixel area corresponding to the background, from the detection range of the maximum peak point at the maximum peak point detection step ST3 will be described later, a predetermined threshold (e.g. , by excluding the temperature below 1000 ° C.), it may avoid the influence of the background.
[0034]
　FIGS. 4A ~ FIG. 5B is a diagram illustrating an example of a captured image acquired in the imaging step ST1, and an example of a histogram that is created in the histogram generation step ST2.
　Specifically, FIG. 4A, for example, tapped initial acquired in the imaging step ST1 to a diagram showing an example of a captured image of the molten steel flow F in which the molten steel M mainly. Incidentally, FIG. 4A is an average image of the five captured images continuously acquired for each scan period of the image pickup means 1 and averaged (resolution of the captured image is about 3 cm / pixel).
　Further, FIG. 5A, for example, is acquired in the imaging step ST1 in tapping metaphase, is a diagram illustrating an example of a captured image of the molten steel flow F consisting mainly of slag S. Incidentally, FIG. 5A, following the continuously acquired five captured images of five captured images was the source of the five captured images (Figure 4A was continuously acquired for each scan cycle of the imager 1 ) is an average image obtained by averaging (resolution of the captured image is about 3 cm / pixel). The pixel area surrounded by a thick broken line in FIG. 5A is a region of pixels determined to correspond to the slag S in the first determination step ST5 described later, a pixel region thought slag S is present.
[0035]
　4A and FIG. 5A, is displayed by cutting only partially about the pixel region corresponding to the background is located in the vicinity and the pixel area corresponding to the molten steel flow F in the captured image (average image). In other words, captured image actually acquired, than captured images shown in FIGS. 4A and 5A, the pixel region of the left-right direction is wider.
　Furthermore, captured image convenience of illustration shown in FIGS. 4A and 5A, although a monochrome display, in fact, the monitor image processing means 2 is provided, with the different colors depending on the temperature of each pixel It is displayed in a. That is, the temperature of the pixel area corresponding to the molten steel flow F is higher than the temperature of the pixel region corresponding to the background, in the actual captured image obtained by the imaging step ST1, is colored a color corresponding to the high temperatures there.
　Also, among the pixel regions corresponding to the molten steel flow F, the temperature (temperature apparent) of the pixel regions including the slag S shown in FIG. 5A (region surrounded by a thick broken line in FIG. 5A), substantially as shown in Figure 4A manner has become higher than the temperature of the pixel region where only the molten steel M is present (temperature apparent), a color corresponding to the higher temperature is colored. Incidentally, in the molten steel flow F discharged from the converter 3, and the actual temperature of the portion corresponding to the pixel region slag S is present (actual temperature), a portion corresponding to a pixel region where substantially only the molten steel M is present the actual temperature (actual temperature) of the considered equivalent value. This is because, as described above, the temperature variation by being stirred in the top-blown or upper bottom blown blowing during refining is suppressed, the converter molten steel M and slag S of the converter 3 as the molten steel flow F 3 is to be discharged to the outside.
　On the other hand, as described above, for the emissivity of the slag S is higher than the emissivity of the molten steel M, if also set to the same value for any pixel configuration of emissivity in the imaging device 1, in the obtained captured image the temperature of the pixel region where the slag S is present is determined to be higher than temperatures substantially pixel regions only molten steel M is present. The same applies to Figure 9A will be described later.
[0036]
　Figure 4B is a diagram showing a histogram created captured image (average image) shown in FIG. 4A. Upon creation of the histogram of FIG. 4B, in order to avoid the influence of the background, the temperature range of the horizontal axis with a predetermined threshold value (1000 ° C.) or higher (although, 1400 ° C. did not show a characteristic distribution of the number of pixels not shown) for less than, the horizontal axis is divided at 10 ° C. pitch, the vertical axis is the number of pixels having a temperature of each section.
　Figure 5B, for the captured image shown in FIG. 5A (average picture), is a diagram illustrating a histogram generated as in FIG. 4B.
[0037]
　(Maximum peak point detection step ST3)
　in the maximum peak point detection step ST3, the image processing means 2, for each histogram created by the histogram creation step ST2, the number of pixels for detecting the maximum peak point is the maximum value. In the histogram shown in FIG. 4B and 5B, a point indicated by the symbol P is the maximum peak point.
　Hereinafter, the captured image shown in FIG. 5A, n-th of the plurality of captured images acquired by sequentially imaging the molten steel flow F in the imaging step ST1 (n ≧ 2: n is a natural number of 2 or more) captured image and then, the maximum peak point in the histogram shown in FIG. 5B was developed for the captured image shown in FIG. 5A P n and, the P n the temperature T n and. Also, the captured image (average image) to n-1 th captured image shown in FIG. 4A (i.e., one before the captured image of FIG. 5A (immediately preceding) captured image), and the captured image shown in FIG. 4A the maximum peak point in the histogram shown in FIG. 4B created P n-1 and, the P n-1 a temperature of T n-1 and.
[0038]
　(Maximum peak point type determining step ST4)
　the maximum peak point type determining step ST4, the image processing unit 2 has detected the maximum peak point detection step ST3, the maximum peak point P in each histogram, it is discharged as molten steel flow F successively it determines whether or not to correspond to either of the slag S and molten steel M.
　Examples maximum peak point P in the histogram shown in FIG. 5B (P a n ) will be described with reference to FIG. 3 method for determining whether to correspond to any of the slag S and the molten steel M (i.e., determined for the captured image of FIG. 5A It will be described the case of). As shown in FIG. 3, the maximum peak point P of the histogram of FIG. 5B (P n the temperature T of) n from the maximum peak point P (P n-1 temperature T of) n-1 by subtracting the was collected using the variation [Delta] T (temperature T in the histogram of Figure 4A n-1 for the temperature T in the histogram of Figure 5B n variation of T n -T n-1 ) is a predetermined value Th1 (for example, 100 ° C.) or higher determines whether or not (step ST41 in FIG. 3). Then, it is determined that the change amount ΔT may be equal to or greater than the predetermined value Th1, corresponding to the slag S to the maximum peak point P in the histogram of Figure 5B is present in the molten steel flow F (step ST42 in FIG. 3).
　Incidentally, as shown in FIG. 5B, the temperature T in the histogram of FIG. 4B n-1 for the temperature T in the histogram of Figure 5B n amount of change [Delta] T (T n -T n-1 ) for is equal to or greater than the predetermined value Th1 It would be determined that the maximum peak point P in the histogram (the histogram of FIG. 5B) of the captured image shown in FIG. 5A corresponds to slag S present in the molten steel flow F.
[0039]
　On the other hand, if when the amount of change ΔT is less than the predetermined value Th1 (instead of the captured image of FIG. 5A, for example, the molten steel flow captured image F obtained by the molten steel M mainly as a determination target, the variation ΔT is a predetermined value assumed is less than Th1), without immediately determined whether or not the maximum peak point P corresponds to the slag S present in the molten steel flow F, the captured image acquired before the captured image of FIG. 5A among them, the maximum peak point P to identify the determined captured image to correspond to the molten steel M which is present in the molten steel flow F (step ST43 in FIG. 3). That is, such a case, n-th captured image acquired before the captured image of the (j-th (j ≦ n-1: j captured image following a natural number) n-1) a maximum peak the point P to identify the one of the determination captured image to correspond to the molten steel M as a reference image (i.e., j-th acquired before the n-th captured image (j ≦ n-1: j is an image captured by the natural number) of the n-1 or less, the maximum peak point P j to identify the reference image determined by the captured image and corresponds to the molten steel M).
　Then, the maximum peak point P (P in the histogram of the reference image j temperature T of) j the maximum peak point P (P n the temperature T of) n amount of change [Delta] T '(T n -T j predetermined value) Th1 It determines a whether more (step ST44 in FIG. 3). Then, the amount of change [Delta] T '(T n -T j ) is equal to or greater than a predetermined value Th1, the maximum peak point P (P n determines that) corresponds to the slag S present in the molten steel flow F (step ST42 in FIG. 3). On the other hand, the amount of change [Delta] T '(T n -T j is less than) a predetermined value Th1, the maximum peak point P (P n is determined to correspond to the molten steel M which) is present in the molten steel flow F (FIG. 3 step ST45 ).
[0040]
　More detail about the step ST43 in FIG. 3, as shown in FIG. 4A, when the molten steel M in the captured image of the molten steel flow F containing mainly corresponds to the molten steel M is the maximum peak point P in the histogram that are present in the molten steel flow F Then it is determined. Therefore, in step ST43, to identify the captured image in FIG. 4A as a reference image for example, it executes the subsequent step ST44. Incidentally, as a reference image, there is no need to identify the captured image of FIG. 4A, a captured image acquired before the FIG 4A, the maximum peak point P corresponds to the molten steel M which is present in the molten steel flow F and determination been if there is a captured image is to identify the captured image as a reference image, may be performed subsequent step ST44.
　However, the reference image used for calculating the amount of change ΔT is preferable in the captured image to the determination target as the imaging sequence is closer. More imaging sequence is short, the temperature variation of the molten steel flow F decreases, because that can detect more accurately slag. Therefore, in step ST43, when determining the captured image of FIG. 5A, an imaging acquired immediately before FIG. 5A, the captured image of FIG. 4A maximum peak point is determined to correspond to the molten steel M as a reference image it is preferable to identify.
[0041]
　Here, as described above, the maximum peak point type determining step ST4, in step ST41, the maximum peak point P in the histogram of the captured image determination target n the temperature T of n and, taken just before the captured image maximum peak point P in the histogram of the captured image n-1 temperature T of n-1 using a difference between the. Therefore, for the first captured image (first acquired captured image), it is impossible to determine the maximum peak point in a manner as described above.
　As described above, the amount of slag and the molten steel amount of the converter 3, the tilt angle of the converter 3 at the time of tapping, and by visual check or the like, in the tapping early, which molten steel flow F is mainly composed of the molten steel M it is possible to confirm whether or not it is. Therefore, it is determined that based on these, and at least one imaging molten steel flow F when being discharged only predominantly molten steel M is a tapping nozzle 31, the maximum peak point of the captured image corresponding to the molten steel M. Incidentally, if it can be confirmed whether the molten steel flow F are those consisting mainly of molten steel M, to obtain a plurality of captured images in tapping initial, maximum peak point for these captured images may be determined to correspond to the molten steel M .
[0042]
　As described above, in the maximum peak point type determining step ST4, without determination using only the temperature of the maximum peak point of the captured image determination target, the maximum peak point of the captured image to determine the target temperature, the It makes the determination using the change amount with respect to the temperature of the maximum peak point of captured image acquired prior to the captured image. Therefore, even if the conditions of tapping operations has changed, it is possible to accurately detect the slag.
[0043]
　Here, an example of a method of determining the predetermined value Th1. Emissivity of the molten steel M and slag S is caused are some differences with the components. Therefore, even if the components of the molten steel and slag in the converter in has changed, it is preferable to detect more accurately the slag. From such a viewpoint, for example, imaging a variety of molten steel flow components of the molten steel and slag are different, a histogram for each molten steel flow, after calculating the amount of change ΔT of molten steel peak and slag peak for each histogram , to create a graph as shown in FIG. 6 taken frequency N [Delta] T, the vertical axis on the horizontal axis. Then, in the graph shown in FIG. 6, to determine for example the value of the minimum value near the variation ΔT as the predetermined value Th1.
[0044]
　Subsequently, with reference to FIG. 7, the maximum peak point type determining step ST4 more detail. Figure 7 is a view for explaining the maximum peak point type determining step ST4, a schematic diagram showing an example of a plurality of captured images obtained by the imaging step ST1.
　As shown in FIG. 7, it is assumed that a captured image of the molten steel flow F to get 10 sheets in imaging step ST1. Of these captured images, 1-4 th captured image is assumed to have captured the molten steel flow F consisting mainly of molten steel M, 5 ~ 7 th captured image captured molten steel flow F mainly containing slag S shall the 8-10 th captured image is assumed to have captured the molten steel flow F consisting mainly of molten steel M. Then, described below how to determine the maximum peak point of the captured image.
[0045]
　As discussed above, it determines that the first captured image, the maximum peak point corresponds to the molten steel M.
　Subsequently, for the second frame of the captured image, the step of the maximum peak point type determining step ST4 ST41 by (see FIG. 3), the maximum peak point P of the second sheet of the captured image to be currently determined target 2 temperature T of 2 from the maximum peak point P of the first captured image acquired immediately before 1 temperature T of 1 to be subtracted to calculate the amount of change [Delta] T (step ST41 in FIG. 3). The variation ΔT, since the captured image of the second sheet is obtained by imaging the molten steel flow F consisting mainly of molten steel M, becomes less than the predetermined value Th1, 1 sheet of the maximum peak point is determined to the molten steel M identifying the captured image as a reference image (step ST43 in FIG. 3). Then, in step ST44 in FIG. 3, the variation [Delta] T '(T n -T j determines for) is less than the predetermined value Th1, the maximum peak point of the second sheet of the captured image to correspond to the molten steel M.
[0046]
　About 3 th captured image, as in the case of the second sheet of the captured image of the maximum peak point P of the third piece of the captured image and the current determination target 3 temperature T of 3 from being acquired immediately before 2nd peak point P of the captured image of the 2 temperature T of 2 to subtract, calculates the amount of change [Delta] T. The variation ΔT is one captured image of the third sheet is from those of the captured molten steel flow F consisting mainly of molten steel M, becomes less than the predetermined value Th1, the maximum peak point is determined to correspond to the molten steel M a captured image of the captured image or the second sheet of the eye identifying as the reference image. Then, in step ST44 in FIG. 3, the variation [Delta] T '(T n -T j determines for) is less than the predetermined value Th1, the maximum peak point of the third piece of the captured image to correspond to the molten steel M. As the reference image, it may be using the captured image of the first sheet, but it is preferable to use most imaging sequence (acquisition order) of the second sheet captured image close to the captured image of the third piece. More imaging sequence is short, the temperature variation is reduced, it is because it can detect more accurately slag.
　About 4 th captured image, it determines as in the case of the third piece of the captured image.
[0047]
　About 5 th captured image, the maximum peak point P of 5 th captured image as the current determination target 5 temperature T of 5 from 4 th captured image acquired just before the maximum peak point P 4 of the temperature T 4 and is subtracted to calculate the amount of change [Delta] T. What amount of change ΔT is 5 th captured image and is intended to image the molten steel flow F consisting mainly of slag S, the captured image of the 4 th has captured the molten steel flow F mainly of molten steel M since it becomes a predetermined value Th1 or more, the maximum peak point of the 5 th captured image is determined to correspond to the slag S.
[0048]
　The 6 th captured image, the maximum peak point P of this determination target to 6 th captured image 6 the temperature T of 6 from the 5 th captured image acquired just before the maximum peak point P 5 of temperature T 5 a is subtracted to calculate the amount of change [Delta] T. The variation ΔT is and is intended captured image 6 th has captured the molten steel flow F consisting mainly of slag S, which is 5 th captured image obtained by capturing the molten steel flow F mainly containing slag S since it becomes less than the predetermined value Th1. Therefore, the subsequent step ST43 (see FIG. 3) is specified as a reference image one of the first sheet-4th captured image maximum peak point is determined to correspond to the molten steel M. Then, in S step T44 of FIG. 3, the variation [Delta] T '(T n -T j determines that since) becomes a predetermined value Th1 or more, the maximum peak point of the 6 th captured image corresponding to the slag S.
　About 7 th captured image, it determines as in the case of 6 th captured image.
[0049]
　The 8 th captured image, the maximum peak point P of 8 th captured image as the current determination target 8 temperature T of 8 from the maximum peak point of the 7 th captured image acquired immediately before P 7 of temperature T 7 and is subtracted to calculate the amount of change [Delta] T. What amount of change ΔT is and is intended is 8 th captured image obtained by capturing the molten steel flow F consisting mainly of molten steel M, the captured image of the 7th has captured the molten steel flow F mainly containing slag S since it becomes less than the predetermined value Th1. Therefore, the subsequent step ST43 (see FIG. 3), the maximum peak point to identify either the first sheet-4th captured image determined as the molten steel M as the reference image. Then, in step ST44 in FIG. 3, the variation [Delta] T '(T n -T j determines that since) is less than the predetermined value Th1, the maximum peak point of the 8 th captured image corresponds to the molten steel M.
[0050]
　The 9 th captured image, the maximum peak point P of this determination target to 9 th captured image 9 the temperature T of 9 from the maximum peak point of the 8 th captured image acquired immediately before P 8 of temperature T 8 a is subtracted to calculate the amount of change [Delta] T. The variation ΔT is 9 th captured images are those obtained by imaging the molten steel flow F of the molten steel M and, from the captured image of 8 th is obtained by imaging the molten steel flow F of the molten steel M, a predetermined value less than Th1. Therefore, the subsequent step ST43 (see FIG. 3), identifies the first sheet to fourth sheet maximum peak point is determined to molten steel, and any of 8 th captured image as a reference image. Then, in step ST44 in FIG. 3, the variation [Delta] T '(T n -T j determines that since) is less than the predetermined value Th1, the maximum peak point of the 9 th captured image corresponding to the molten steel M.
　About 10 th captured image, it determines as in the case of 9-th captured image.
[0051]
　As described above, the maximum peak point type determining step ST4, the plurality of captured images acquired by the imaging step ST1, a determination described above is executed sequentially.
　Incidentally, that the example shown in FIG. 7, the second frame of the captured image to perform the step ST41 in FIG. 3, the molten steel flow F visually or the like, for example, in tapping early as described above is due to the molten steel M if it can confirm, for example, without executing the step ST41 in FIG. 3 also the second piece of captured image similar to the first captured image, be determined that the maximum peak point P corresponds to the molten steel M good.
[0052]
　(First determination step ST5)
　first determination step ST5 is executed based on the determination result of the maximum peak point type determination step ST4. Specifically, if the maximum peak point P of the captured image determination target is determined to correspond to the slag S present in the molten steel flow F, the image processing means 2 first determination step ST5 is executed. When FIG. 5B (the captured image of the molten steel flow F consisting mainly of slag S) will be described as an example, the first determination step ST5, the image processing means 2, among the pixels constituting the captured image, the maximum peak point slag pixels having a temperature lower than the first threshold determined based on P corresponds to the molten steel M which is present in the molten steel flow F, the pixel having a first threshold value or more temperature present in the molten steel flow F It is determined to correspond to the S. Hereinafter, with reference to FIG. 8 as appropriate, it will be described more specifically.
[0053]
　Figure 8 is a diagram for explaining a first threshold value which is determined in the first determination step ST5. Incidentally, the histogram shown in FIG. 8 is the same as the histogram shown in Figure 5B.
　As shown in FIG. 8, the first threshold value, the histogram created by the histogram creation step ST2, represented the maximum peak point P in the first straight line L1 having the streets and positive slope. More specifically, the first straight line L1 is a straight line passing through the Q and the maximum peak point P point shown in FIG. Point Q, a predetermined number of pixels threshold Th2 (e.g., the maximum peak point of 50% the number of pixels of P) less than a predetermined value than the temperature of and the maximum peak point P has a number of pixels TD (e.g., 50 ° C. ) among the points having the above low temperature, a peak point with the highest temperature (i.e., the point Q is than and the temperature of the maximum peak point P has a number of pixels less than a predetermined number of pixels threshold Th2 of the point at which the maximum value be a point having a lower temperature than a predetermined value TD, the point with the highest temperature).
　Here, among the points having a temperature lower than a predetermined value TD, when determining whether the peak point as a determination target point with the highest temperature, and a point adjacent to the cold side of the point and the point connecting attention to the slope of the line, the line is considered a positive slope (the line is upward-sloping line) a long if the point to point of the determination target Q.
　Incidentally, slag detection method according to the present embodiment is particularly suitably applied to the molten steel flow peaks less than a predetermined pixel number threshold value Th2 in the histogram. Further, regardless of the setting whether a predetermined number of pixels threshold Th2, are particularly preferably applied to the molten steel flow 50% or more peaks in the number of pixels of the maximum peak point P is for example equal to or less than 2 points. Feature of such peaks is determined by the mixed condition of molten steel and slag in the refining.
　The above first threshold value (first straight line L1), the temperature of the horizontal axis is X, and the number of pixels in the vertical axis and Y, will be expressed by the following equation (1).
　Y = aX + b ··· (1 )
　where, a is a positive constant, b is a constant. These constants are determined from the first straight line L1 passes through the point Q and the maximum peak point P.
[0054]
　Predetermined value TD is not particularly limited, for example, 50 ° C.. Experience on a range of within ± 50 ℃ of the maximum peak temperature is often not the foot. Therefore, for example, by setting the predetermined value TD to 50 ° C., preferably it is possible to determine a first threshold value at a peak excluding the foot.
　The predetermined number of pixels threshold Th2 is not particularly limited, 1200 ° C. ~ not to capture the peak temperature regions appear to background such 1300 ° C., for example at the maximum peak point P of the number of pixels preferably to 50% and Th2.
[0055]
　As described above, the image processing unit 2 determines that a pixel having a temperature lower than the first threshold value corresponds to the molten steel M which is present in the molten steel flow F. That is, the pixel which satisfies Y> aX + b will be determined to correspond to the molten steel M which is present in the molten steel flow F.
　On the other hand, the image processing unit 2 determines that a pixel having a first threshold value or more temperature corresponds to the slag S present in the molten steel flow F. That is, in this embodiment, it is determined that Y ≦ aX + b pixels satisfying the (pixel in the region hatched in FIG. 8) corresponds to the slag S present in the molten steel flow F.
[0056]
　(Second determination step ST6)
　second determination step ST6 is executed based on the determination result of the maximum peak point type determination step ST4. Specifically, at the maximum peak point type determining step ST4, if the maximum peak point P of the captured image determination target is determined to correspond to the molten steel M which is present in the molten steel flow F, the second determination step by the image processing means 2 ST6 is executed. In the second determination step ST6, the image processing means 2, among the pixels constituting the captured image, pixels having a second threshold temperature below which was determined based on the maximum peak point P is present in the molten steel flow F determines that corresponds to the molten steel M, pixels having a temperature higher than the second threshold value corresponds to the slag S present in the molten steel flow F to. Hereinafter, referring as appropriate to FIGS. 9A and 9B, is described more specifically.
[0057]
　Figure 9A is a captured image acquired in the imaging step ST1, a diagram showing other different examples and FIGS. 4A and 5A. Incidentally, FIG. 9A shows the mean image of the five captured images continuously acquired for each scan period of the image pickup means 1 and averaged. Imaging conditions are the same as FIGS. 4A and 5A.
　Further, FIG. 9B is a diagram showing a histogram created based on the captured image of FIG. 9A, a diagram for explaining a second threshold which is determined in the second determination step ST6.
[0058]
　In the example shown in FIG. 9B, the maximum peak point type determination step ST4 described above, the maximum peak point P in the histogram is determined to correspond to the molten steel M which is present in the molten steel flow F. Therefore, the image processing unit 2 performs a second determination step ST6.
[0059]
　As shown in FIG. 9B, second threshold is expressed maximum peak point P on the second straight line L2 with the street and negative slope. If the absolute value is the larger are (preferably towards the slope of the second straight line L2 than the absolute value of the inclination of the first straight line L1, the absolute value of the slope of the second straight line L2 is the inclination of the first straight line L1 of 1.5 to 2.5 times the absolute value). The first straight line L1 is a straight line passing through the Q and the maximum peak point P point shown in FIG. 9B. Since the point Q is the slope of a line between a point adjacent to the low temperature side positive, as in FIG. 8, the temperature of and the maximum peak point P has a number of pixels less than the number of pixels threshold Th2 predetermined value TD (e.g., 50 ° C.) than among the points having a temperature lower than a peak point with the highest temperature.
[0060]
　As described above, the first straight line L1, the temperature of the horizontal axis is X, and the number of pixels in the vertical axis and Y, will be expressed by the following equation (1).
　Y = aX + b ··· (1 )
　where, a is a positive constant, b is a constant. These constants are determined from the first straight line L1 passes through the point Q and the maximum peak point P.
[0061]
　On the other hand, for example, when the absolute value of the slope of L2 of the second straight line is to be set to twice the absolute value of the slope a of the first straight line L1, the second straight line L2, it is represented by the following formula (2) become.
　Y = -2aX + c ··· (2 )
　where, a is a positive constant, c is a constant. Then, a is determined from the first straight line L1, c is determined from the second straight line passing through the maximum peak point P.
[0062]
　As described above, the image processing unit 2 determines that a pixel having a second threshold temperature below corresponds to the molten steel M which is present in the molten steel flow F. That is, in the histogram shown in FIG. 9B, pixels satisfying Y ≦ -2aX + c will be determined to correspond to the molten steel M which is present in the molten steel flow F.
　On the other hand, the image processing unit 2 determines that a pixel having a temperature higher than the second threshold value corresponds to the slag S present in the molten steel flow F. That is, the histogram shown in FIG. 9B, Y> -2aX + c satisfies the pixels (pixels in the region hatched in FIG. 9B) will be determined as corresponding to slag S present in the molten steel flow F.
[0063]
　According to slag detection method according to the present embodiment described above, the temperature at the maximum peak point of the captured image determination target, the temperature at the maximum peak point of the captured image obtained before the captured image based on the difference, the maximum peak point of the captured image determination target is to determine whether to correspond to or slag S corresponding to the molten steel M, the temperature of the molten steel flow F by changing the conditions of the tapping operation has changed even if it is possible to perform the determination with high accuracy. Therefore, it accurately detects the slag S in the molten steel flow F.
[0064]
　Further, according to the slag detection method according to the present embodiment, in the first determination step ST5 or the second determination step ST6, the number of pixels corresponding to the slag S present in the molten steel flow F (area), the molten steel flow F the number of pixels corresponding to the molten steel M which present the (area) can be calculated. Thus, for example, it can be determined the area ratio of the slag S in the molten steel flow F, and the volume fraction of the slag S in the molten steel flow F. By using further the specific gravity of the molten steel M and slag S, it is possible to calculate the mass ratio of the slag S in the molten steel flow F, flow rate of the molten steel flow F can be estimated from the inclination angle of the converter 3 when the tapping . Therefore, by using the flow rate of the mass ratio and the molten steel flow F of the slag S, it is possible to estimate runoff slag S a (flow), tapping operation of the converter 3 the outflow amount of slag S it becomes possible to control the range that is required in.
　Specifically, according to the slag detection method according to the present embodiment, or to exit the tapped operation if the outflow amount of the slag S (outflow, the number of pixels, the area, etc. volume) is greater than zero termination or the tapped operation if the outflow amount of the slag S is greater than a predetermined value, if the ratio of the outflow amount of the slag S against outflow amount of the molten steel M is greater than a predetermined value the it is possible to perform the control such as to terminate the tapping operation.
Example
[0065]
　Next, a description will be given of an embodiment carried out for confirming the effect of the present invention.
[0066]
　Using the captured image shown in FIG. 5A as the evaluation target, and slag detection method according to the present embodiment was compared with the slag detection method described in Patent Document 2.
　Specifically, the slag detecting method according to the present embodiment, as described above, the histogram shown in FIG. 5B was developed for the captured image shown in FIG. 5A, the slag S to the maximum peak point P is present in the molten steel flow F It is determined corresponding with. Then, as shown in FIG. 8, the first straight line L1 represented by the above formula (1), a pixel in the region hatched is determined to correspond to the slag S.
　In the example shown in FIG. 8, 309 pixels is determined to correspond to the slag S.
[0067]
　On the other hand, the use of slag detection method described in Patent Document 2, the maximum peak point P in the histogram shown in FIG. 5B is considered to correspond to the molten steel M which is present in the molten steel flow F. As described above, the slag detecting method described in Patent Document 2, the horizontal axis direction of the dispersion density value N1 or more pixels consideration σ of the maximum peak point P is determined as the molten steel M, the bias value to the density value N1 B have determined that the slag S concentration value N2 or more pixels by adding the. Replacing the density value of the temperature, the slag detecting method described in Patent Document 2, the horizontal axis direction of the variation temperature N1 or more pixels even considering σ of the maximum peak point P is determined as the molten steel M, the bias value to the temperature N1 the temperature N2 or more pixels obtained by adding B will determine that a slug S. Here, the slag detection method disclosed in Patent Document 2, since the maximum peak point P is considered to correspond to the molten steel M, the bias value B for distinguishing between pixels corresponding to the pixels and the slag S corresponding to the molten steel M set above 2 [sigma] (i.e., the temperature N2 set to at least the temperature + sigma of the maximum peak point P) of it is reasonable. In this evaluation, using the minimum value 2σ most detection error of slag S is decreased as the bias value B. The pixel of the histogram shown in FIG. 5B, assuming that the temperature or the number of pixel distribution of the maximum peak point P is a normal distribution, the temperature of the maximum peak point P to a temperature N2 (temperature of the maximum peak point P + sigma) the sum of the number, divided by the sum of the number of or more pixels temperature of the maximum peak point P is set to σ to approximately 68%.
[0068]
　10A and 10B are views for explaining a slag detection method described in Patent Document 2. Figure 10A is a histogram, FIG. 10B shows a captured image (average image). Histogram shown in FIG. 10A is the same as the histogram shown in FIG. 5B or FIG. Captured image shown in FIG. 10B is the same as the captured image shown in Figure 5A. According to the slag detection method described in Patent Document 2, a pixel in the region hatched in FIG. 10A is determined to correspond to the slag S. According to this evaluation, 50 pixels in the pixel area surrounded by a thick broken line in FIG. 10B is determined to correspond to the slag S.
[0069]
　The number of pixels corresponding to the molten steel flow F in the captured image shown in FIG. 5A was 465 pieces. Since the entire pixel area corresponding to the molten steel flow F is a temperature exceeding the operating temperature, the entire pixel area corresponding to the molten steel flow F is assumed to correspond to the slag S, a true value of the number of pixels corresponding to the slag S 465 it can be regarded as a number. Thus, if the true value of the number of pixels corresponding above 465 pieces of slag S, slag detection method according to the present embodiment, -33.5% of the true value ((309-465) / 465 × 100 = - to which the the error of 33.5), the slag detecting method described in Patent Document 2, an error of -89.2% of the true value ((50-465) /465×100=-89.2) there were. Therefore, according to the slag detection method according to the present embodiment, as compared with the slag detection method described in Patent Document 2, it can be said that the slag S in the molten steel flow F is accurately detectable. This is more of a thick broken line shown in FIG. 5A, as compared with the thick broken line shown in FIG. 10B, a point close to the contour of the pixel area of ​​the molten steel flow F is considered slag S is present, also apparent in visual manner is there.
[0070]
　Incidentally, when converting the true value of the number of pixels corresponding to the slug S to the area (actual size), the area of one pixel is about 9cm 2 for a, 9 × 465 = 4185Cm 2 becomes. Simply Convert volume this (converted visual axis dimension of the imaging unit 1 of the slag S is assumed to be the same as the dimension in the viewing plane of the image pickup means 1) Then, (4185) 3/2 = 270734Cm 3 270 734 × 10 = -6 m 3 a. Therefore, the specific gravity of the slag S 2 × 10 -3 m 3 When / kg, the mass of slag S is, (two hundred seventy thousand seven hundred and thirty-four × 10 -6 ) / (2 × 10 -3 a) = 135 kg.
　Similarly, when converting the slag S detected by the slag detection method according to the present embodiment mass 73kg (54.1% of the true value), and the slag S detected by the slag detection method described in Patent Document 2 mass a 5 kg (3.7% of the true value) when converted into. That is, according to the slag detection method according to the present embodiment, the mass in becomes -45.9% error, as compared to the method described in Patent Document 2 an error of -96.3% occurs in the molten steel flow F it can be said that the slag S is accurately detectable.
[0071]
　Having described the embodiments of the present invention, the above embodiments have been presented by way of example, the scope of the present invention is not limited to the above embodiment. The embodiments described herein may be embodied in other various forms, without departing from the spirit of the invention, various omissions, substitutions, and changes can be made. The above embodiments and their modifications as would fall within the scope and spirit of the invention, and are included in the invention and the scope of their equivalents are claimed.
[0072]
　For example, in the above embodiment, showing a case where the first threshold value is represented by a first straight line L1 and having a positive slope as the maximum peak point P. Relates method of determining the threshold value, from the viewpoint of detecting a slag with higher accuracy, it is preferable to determine by performing fitting, etc. each Gaussian distribution for the peak of the molten steel and slag. However, in such a method, the calculation time becomes long, industrially not preferred. Therefore, by representing the first threshold by a straight line, it determines a threshold value more easily.
[0073]
　The first and second threshold values ​​is not limited to the case represented by the first straight line L1 and the second straight line L2. For example, the first threshold value and second threshold value, considering the horizontal axis direction of the variation of the maximum peak point, may be represented by a straight line perpendicular to the horizontal axis (a straight line of slope is infinite).
DESCRIPTION OF SYMBOLS
[0074]
1: imaging means
2: image processing means
3: the converter
4: ladle
100: slag detection device
ST1: imaging step
ST2: histogram generation step
ST3: maximum peak point detection step
ST4: Peak point type determining step
ST5: first determination step
ST6: second determination step
F: the molten steel flow
M: the molten steel
S: slag

WE CLAIM

Flows toward the ladle from a converter furnace, an imaging process and which sequentially images the molten steel stream comprising molten steel and slag to obtain a plurality of captured images
　density parameters corresponding to the density of each pixel constituting the respective pickup images was a horizontal axis, the histogram generation step and creating for the total number at which the histogram each captured image to the longitudinal axis the number of pixels of the pixels having the concentration parameter;
　for each histogram, the number of pixels is the maximum value maximum peak point detection step of detecting a certain maximum peak point;
　wherein the maximum peak point type determination step of determining whether the maximum peak point of each histogram corresponds to one of the slug or the molten steel;
have,
　the maximum maximum peak point P of the histogram of the captured image of the n th peak point type determining step (n ≧ 2) n when determining:
　Histo of n-1 th captured image Maximum peak point P in ram n-1 concentration parameter T of n-1 with respect to the maximum peak point P n concentration parameter T of n when the change amount ΔT of a predetermined value or more, the maximum peak point P n is the slug determining a corresponding;
　On the other hand, the case variation ΔT is smaller than the predetermined value, the maximum peak point P in the histogram of the captured image Like j the maximum peak point is determined to the molten steel th (j ≦ n-1) j concentration parameter T of j the concentration parameter T for n of variation [Delta] T 'if said predetermined value or more, the maximum peak point P n is determined corresponding to the slug, the variation [Delta] T' if it is less than the predetermined value, the maximum peak point P n is determined that corresponds to the molten steel;
slag detection method of the molten steel flow, characterized in that.
[Requested item 2]
　In the maximum peak point type determination step,
　the maximum peak point P j as the n-th acquired before the captured image and the n th largest peak point most acquisition order in the captured image with the near of the using the maximum peak point in the histogram in the molten steel judged captured image
slag detection method in the molten steel flow according to claim 1, characterized in that.
[Requested item 3]
　When the maximum peak point in the maximum peak point type determination step determines that corresponding to the slug, the pixels having a density parameter of less than the first threshold determined based on the maximum peak point corresponding to the molten steel and a pixel having a density parameter of more than the first threshold value first determination process and determines that corresponding to said slug;
　if it is determined that the maximum peak point in the maximum peak point type determination process corresponding to the molten steel , the pixels having a density parameter of the following second threshold determined based on the maximum peak point corresponds to the molten steel, a pixel having a larger concentration parameter than the second threshold value corresponding to said slug ; second determination step and determining a result
, further comprising a
slag detection method in the molten steel flow according to claim 1 or 2, characterized in that.
[Requested item 4]
　Wherein the first threshold value, in the histogram, the expressed maximum peak point in a first straight line having a street and a positive slope,
　the second threshold value, in the histogram, as the maximum peak point and is represented by a second straight line having a negative slope,
　the absolute value of the second slope of the line is greater than the absolute value of the slope of the first linear
flow of molten steel according to claim 3, characterized in that slag detection method in.
[Requested item 5]
　The first straight line, among the points having a small concentration parameter greater than a predetermined value with respect to the concentration parameter of and having the number of pixels smaller than the number of pixels threshold the maximum peak point, density parameters maximum peak and the point, a straight line passing through the said maximum peak point,
　the absolute value of the slope of the second straight line is 1.5 to 2.5 times the absolute value of the slope of the first linear
and wherein the slag detection method in the molten steel flow according to claim 4.