JP6372272B2 - Apparatus for determining furnace wall surface state of coking chamber of coke oven, method for determining furnace wall surface state of coking chamber of coke oven, and program - Google Patents

Description

  The present invention relates to a furnace wall surface state determination device for a coking chamber of a coke oven, a method for determining a furnace wall surface state of a coking chamber of a coke oven, and a program. It is suitable for use in determining the state.

In the steel industry, a coke oven called a chamber type is used. A coke oven is a huge industrial furnace configured by alternately connecting a number of carbonization chambers and combustion chambers through a furnace wall formed of refractory bricks or the like. The standard size of one carbonization chamber is about 16 m in length, 6.5 m in height, and about 0.4 m in width, and is characterized by a narrow width compared to the length and height.
Coal is charged into the carbonization chamber, and a high temperature of 1000 ° C. or higher is continuously applied from the combustion chamber through the furnace wall to the carbonization chamber charged with coal for about 20 hours to dry-coal the coal to produce a coke cake. (In the following description, “coke cake” is simply referred to as “coke”). When dry distillation is completed, the doors at both ends of the carbonization chamber are opened, and the coke is discharged in a horizontal direction by means of an apparatus called an extrusion ram. Coke is produced at different times in carbonization chambers lined up in about 100 rooms. Because refractory bricks are damaged by thermal shocks that lower the temperature, the interior of the coke oven is always maintained at a high temperature.

  Coke ovens in the domestic steel industry were intensively built during the period of high economic growth, and many furnaces have been operating for more than 30 years. In such a coke oven that has been operating for a long period of time, the refractory bricks that make up the furnace chamber (carbonization chamber and combustion chamber) have deteriorated due to thermal, chemical, or mechanical factors. Concavities and convexities are generated on the wall of the furnace due to the occurrence of meat, corner chipping, and the like. In operating a coke oven, it is desirable that the extrusion load be small when extruding coke after dry distillation. The extrusion load depends on various factors such as coal blending / moisture, dry distillation state, etc. In particular, the unevenness of the furnace wall greatly affects the extrusion load as the frictional resistance between the coke and the furnace wall. If the extrusion load is high, there may occur a situation in which it is impossible to discharge, which is called “clogging”. In this case, coke production is inevitably reduced.

  By the way, normally, carbon is partially attached to the furnace wall surface of the carbonization chamber. Carbon grows when the carbon component of the gas generated during the carbonization of coal adheres to the furnace wall. In some cases, this carbon fills the unevenness generated on the furnace wall surface of the carbonization chamber and brings about the effect of smoothing the furnace wall surface. On the other hand, when the carbon further grows, the surface of the carbon becomes irregularly uneven, increasing the extrusion load. Although the amount of carbon deposited in one dry distillation is small, the carbon gradually becomes thicker in daily operations. Therefore, after the coke is extruded, before the next coal to be carbonized is charged into the carbonization chamber, an operation is performed in which air is blown from the coal charging inlet at the upper portion of the carbonization chamber to burn the carbon. However, if all the carbon is removed, fine irregularities on the furnace wall of the carbonization chamber will be exposed, and the extrusion load will rise rapidly. It is known that when the temperature of the furnace wall is high, the growth rate of carbon is known to be high, and the manner of carbon deposition varies depending on the operation state and the part of the furnace wall. In this way, it is very difficult to manage the carbon adhesion on the furnace wall of the carbonization chamber, and in a coke oven that has deteriorated and the refractory bricks have become thin and rough, the carbon adhesion on stable operation is poor. It is one of the important monitoring items.

  The present inventors have developed and put into practical use an apparatus for imaging a furnace wall surface of a coking chamber by inserting a large heat-resistant measuring device into the coking chamber of a coke oven (see, for example, Patent Document 1). In the following description, this apparatus is referred to as a “furnace wall observation apparatus” as necessary. For example, the furnace wall observation device is equipped with an imaging device called a line CCD camera in which 1000 or more photodetecting elements are arranged in a row. A line CCD camera with a linear field of view along the vertical direction of the furnace wall is mounted on a water-cooled heat insulation device and sent to the coking oven carbonization chamber. Thus, a two-dimensional thermal image is generated. Thus, a high-definition thermal image of the entire furnace wall can be obtained by the furnace wall observation apparatus. If a person sees the thermal image of the furnace wall surface of the carbonization chamber, the state of carbon adhesion on the furnace wall surface of the carbonization chamber can be understood.

Moreover, there are techniques described in Patent Documents 2 and 3 as techniques for automatically determining the adhesion state of carbon on the furnace wall surface of the carbonization chamber.
Patent Document 2 discloses a technique for extracting a plurality of damaged areas from a carbonized interior wall image and deriving damaged site data including the extracted damaged area images, feature amounts, and damage names.
In Patent Document 3, after correcting the brightness of the current furnace wall image to the brightness of the previous furnace wall image, a difference image between the current furnace wall image and the previous furnace wall image is acquired. A technique for extracting a portion having a change as a furnace wall abnormal portion is disclosed.

Japanese Patent No. 3590509 JP-A-11-256166 JP 2009-57491 A

Supervised by Mikio Takagi and Yoshihisa Shimoda, “Image Analysis Handbook”, The University of Tokyo Press, published on January 17, 1991, p503

  However, the thermal image of the furnace wall of the carbonization chamber has a luminance change due to temperature distribution. For this reason, with the techniques described in Patent Documents 2 and 3, it is difficult to extract carbon from the density information of the thermal image of the furnace wall surface of the carbonization chamber. Thus, there has never been a technique for automatically reading the carbon deposition status from the thermal image of the furnace wall of the carbonization chamber. At present, the size of the carbon deposition area is judged by human eyes (sensuously). Yes.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to enable automatic and accurate determination of carbon adhesion from a thermal image of a furnace wall of a carbonization chamber.

  The furnace wall surface state determination device for the coking chamber of the coke oven according to the present invention is composed of each pixel having a density value corresponding to the thermal radiance distribution of the thermal radiance distribution on the furnace wall surface of the coking chamber of the coke oven. Wall surface image acquisition means for acquiring as a thermal image, first area dividing means for dividing the thermal image acquired by the wall surface image acquisition means into a plurality of areas, and a plurality of areas divided by the first area dividing means Based on the density histogram deriving means for deriving a density histogram showing the relationship between the density and the number of pixels for each pixel of the thermal image for each area, and the shape of the density histogram derived by the density histogram deriving means, It is determined for each of the plurality of regions whether the state of the furnace wall surface is a healthy brick surface state, a mottled carbon adhesion state, or a uniform carbon adhesion state. Wall surface determination means, and the sound brick surface state is a state where the furnace wall surface is exposed, and carbon is clogged linearly in joints and cracks of the refractory bricks constituting the furnace wall surface. The uniform carbon adhering state is a state in which carbon is substantially uniformly adhering so as to cover the furnace wall surface, and the mottled carbon adhering state is a refractory that constitutes the furnace wall surface. It is in a state where carbon having a width equal to or greater than the width of the joints and cracks is attached to the area including the joints and cracks of the brick, and the wall surface state determination means has one peak in the density histogram Satisfying the condition and satisfying the condition that the state of the region satisfying the first condition is the sound brick surface state, satisfying the condition that there is one peak in the density histogram, and The state of the region that does not satisfy the condition 1 is determined to be the uniform carbon adhering state, and the state of the region that satisfies the second condition that there are two peaks in the concentration histogram is defined as the patchy carbon. The first condition is that the number of pixels at a density value farther from the peak than on the peak in the density histogram is higher than the peak in the density histogram. It is a condition indicating that there are many.

  The method for determining the surface state of a coke oven coke chamber according to the present invention comprises a distribution of thermal radiance on the wall surface of the coke oven coke chamber comprising each pixel having a density value corresponding to the thermal radiance. A wall surface image acquiring step acquired as a thermal image, a first region dividing step of dividing the thermal image acquired by the wall surface image acquiring step into a plurality of regions, and a plurality of regions divided by the first region dividing step Based on the density histogram derivation step for deriving a density histogram indicating the relationship between the density and the number of pixels for each pixel of the thermal image for each region, and the shape of the density histogram derived by the density histogram derivation step, It is determined for each of the plurality of regions whether the state of the furnace wall surface is a healthy brick surface state, a mottled carbon adhesion state, or a uniform carbon adhesion state. Wall surface state determining step, and the sound brick surface state is a state where the furnace wall surface is exposed, and carbon is clogged linearly in joints and cracks of the refractory brick constituting the furnace wall surface. The uniform carbon adhering state is a state in which carbon is substantially uniformly adhering so as to cover the furnace wall surface, and the mottled carbon adhering state is a refractory that constitutes the furnace wall surface. It is in a state where carbon having a width equal to or greater than the width of the joints and cracks is attached to the area including the joints and cracks of the brick, and the wall state determination step has one peak in the density histogram. Satisfying the condition and satisfying the condition that the state of the region satisfying the first condition is the sound brick surface state, satisfying the condition that there is one peak in the density histogram, and The state of the region that does not satisfy the condition 1 is determined to be the uniform carbon adhering state, and the state of the region that satisfies the second condition that there are two peaks in the concentration histogram is defined as the patchy carbon. The first condition is that the number of pixels at a density value farther from the peak than on the peak in the density histogram is higher than the peak in the density histogram. It is a condition indicating that there are many.

  The program of the present invention causes a computer to function as each means of the furnace wall surface state determination device for the coking chamber of the coke oven.

  According to the present invention, a density histogram indicating the relationship between the density and the number of pixels of the thermal image of the furnace wall surface of the coking chamber of the coke oven is derived. The condition that there is only one peak in the density histogram is satisfied, and the number of pixels exceeds 0 over a wider range in the region on the higher density side than the peak in the density histogram. A region that satisfies the first condition is determined to be a region in a healthy brick surface state. In addition, it is determined that the region having one peak in the density histogram and not satisfying the first condition is a region having a uniform carbon adhesion state. Further, an area that satisfies the second condition that the number of peaks in the density histogram is two is determined to be an area having a mottled carbon adhesion state. Therefore, it is possible to automatically and accurately determine the carbon adhesion state from the thermal image of the furnace wall surface of the carbonization chamber.

It is a figure which shows an example of the thermal image of the furnace wall surface of the carbonization chamber of a coke oven. It is a figure which shows the 1st example of a functional structure of the furnace wall surface state determination apparatus of the carbonization chamber of a coke oven. It is a figure which shows an example of a furnace wall surface thermal image. It is a figure which shows an example of the furnace wall surface thermal image after a shading correction | amendment. It is a figure which shows an example of the threshold value by the image which cut out the block which is a mottled carbon adhesion state from the furnace wall surface thermal image data after shading correction | amendment, the density histogram in the said block, and the method of Otsu. It is a figure which shows an example of the threshold value by the method of the image which cut out the block which is a healthy brick surface state after shading correction | amendment, and the density histogram in the said block, and the method of Otsu. It is a figure which shows an example of the threshold value by the image which cut out the block which is a uniform carbon adhesion state from the furnace wall surface thermal image data after a shading correction | amendment, the density histogram in the said block, and the method of Otsu. It is a figure which shows an example of the display screen of the adhesion state of carbon in the furnace wall surface of the carbonization chamber of a coke oven. It is a flowchart explaining the 1st example of operation | movement of a furnace wall surface state determination apparatus. It is a figure which shows an example of the result of the determination of the adhesion state of the carbon in the furnace wall surface of a carbonization chamber. It is a figure which shows the 2nd example of a functional structure of the furnace wall surface state determination apparatus of the carbonization chamber of a coke oven. It is a figure which illustrates notionally an example of the method of determining whether there exists local maximum in the area | region of the high concentration side from a peak concentration. It is a figure which shows an example of the peak density in a density histogram, the low density side gravity center density, and the high density side gravity density. It is a flowchart explaining the 2nd example of operation | movement of a furnace wall surface state determination apparatus.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
First, the outline | summary of embodiment of this invention is demonstrated.
FIG. 1 is a diagram (photograph) showing an example of a thermal image of a furnace wall surface of a coking chamber of a coke oven, taken by a furnace wall observation apparatus.
In the embodiment of the present invention shown below, the state of the furnace wall surface of the coking chamber of the coke oven is divided into a plurality of small regions obtained by dividing the furnace wall surface, the sound brick surface state, the uniform carbon adhesion state, and the patchy carbon adhesion. The state is classified into one of the three states.

The sound brick surface means that the furnace wall composed of refractory bricks is exposed, and carbon is clogged linearly in joints and cracks (mainly vertical cracks (cracks that occur in the vertical direction)). The state that is. In the example illustrated in FIG. 1, for example, the region 101 is a region in a healthy brick surface state.
The uniform carbon adhering state refers to a state in which carbon is adhering substantially uniformly so as to cover the furnace wall surface constituted by the refractory bricks. In the example illustrated in FIG. 1, for example, the region 102 is a region having a uniform carbon adhesion state.

  The mottled carbon adhering state is a state in which carbon is adhering in a mottled manner, in areas including joints and cracks (mainly vertical cracks (cracks formed in the vertical direction)). It means a state where carbon having a width equal to or larger than the width is attached. More specifically, the mottled carbon adhering state refers to a state in which carbon having a width about three times larger than the width of joints or cracks is adhering. In the spotted carbon adhesion state, there is a state where carbon fills minute irregularities generated on the surface of the refractory brick, and there is no defect on the surface of the refractory brick, but there is an ongoing state in the uniform carbon adhesion state. It is done. In FIG. 1, for example, the region 103 is a region in a mottled carbon adhesion state.

  As shown in FIG. 1, the area where the carbon is attached has a feature that the concentration is slightly higher (that is, brighter) than the area where the surrounding refractory brick is exposed. Here, the density refers to light and dark (that is, luminance on the image) of an image having 256 gradations. The relationship between the image density and the thermal radiance on the furnace wall of the carbonization chamber is a linear relationship. This density value becomes the pixel value of each pixel of the thermal image.

As described above, although there is a difference between the carbon concentration and the refractory brick concentration, it is not easy to accurately identify the carbon adhesion state on the entire furnace wall of the carbonization chamber only by performing a simple binarization process. . The thermal image of the coke oven coking chamber wall is a thermal image of thermal radiation that emits light at high temperatures. The density of the thermal image also changes depending on the temperature distribution of the furnace wall. This is because a concentration distribution occurs even in the attached region.
Therefore, in each embodiment shown below, based on the shape of the concentration histogram for each of a plurality of small regions obtained by dividing the furnace wall surface of the coking chamber of the coke oven, the state of each region obtained by dividing the furnace wall surface is a healthy brick surface state. It is determined which of the three states of the uniform carbon adhering state and the mottled carbon adhering state.

(First embodiment)
First, a first embodiment of the present invention will be described.
FIG. 2 is a diagram illustrating an example of a functional configuration of the furnace wall surface state determination device 200 of the coking chamber of the coke oven. Note that the hardware of the coke oven coking chamber surface condition determination apparatus 200 includes, for example, a CPU, ROM, RAM, HDD, image input / output board, and an information processing apparatus including various interfaces and dedicated hardware. It is realizable by using. Moreover, in the following description, the furnace wall surface state determination device of the coking chamber of the coke oven is abbreviated as a furnace wall surface state determination device as necessary.

[Wall surface image acquisition unit 201]
The wall surface image acquisition unit 201 acquires thermal image data of the furnace wall surface of the carbonization chamber imaged by the furnace wall observation device. The thermal image data on the furnace wall surface of the carbonization chamber is data on the distribution of thermal radiance on the furnace wall surface of the carbonization chamber, and is composed of pixels having density values corresponding to the thermal radiance.

In the present embodiment, the furnace wall observation apparatus captures 256-gradation thermal image data with a resolution of 1 to 2 mm. In addition, since the furnace wall observation apparatus is described in patent document 1 etc., the detailed description is abbreviate | omitted. Moreover, in the following description, the thermal image of the furnace wall surface of the carbonization chamber imaged by the furnace wall observation device is referred to as “furnace wall thermal image” as necessary, and the thermal chamber image of the carbonization chamber imaged by the furnace wall observation device. The data of the furnace wall thermal image is referred to as “furnace wall thermal image data” as necessary.
FIG. 3 is a diagram (photograph) showing an example of a furnace wall surface thermal image.
The wall surface image acquisition unit 201 is realized by using, for example, a CPU, a ROM, a RAM, an HDD, and a communication interface.

[Shading correction unit 202]
The shading correction unit 202 performs shading correction on the furnace wall surface thermal image data acquired by the wall surface image acquisition unit 201. As shown in FIG. 3, since the density unevenness (brightness unevenness) due to the temperature difference exists in the furnace wall surface thermal image, this density unevenness is removed.
Specifically, the shading correction unit 202 performs averaging filtering using a kernel having a large pixel size (for example, a kernel having a kernel size of 128 × 128). The shading correction unit 202 calculates the difference between the pixel values of the pixels corresponding to the furnace wall thermal image data on which the averaging filtering has been performed and the furnace wall thermal image data acquired by the wall image acquisition unit 201, in the corresponding pixel. A difference image as a value is derived. Then, the shading correction unit 202 generates image data obtained by adding a predetermined density (for example, 128 which is half of 256 gradations) to the pixel value of each pixel of the difference image as the furnace wall thermal image data after the shading correction. To do. The reason why the predetermined density is added to the pixel value of each pixel of the difference image is that the pixel value of the difference image is small, and if the difference image is used as it is, a dark image is obtained.
FIG. 4 is a diagram (photograph) showing an example of a furnace wall thermal image after shading correction.
The shading correction unit 202 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[First region dividing unit 203]
The first region dividing unit 203 divides the image region of the post-shading-corrected furnace wall thermal image data generated by the shading correction unit 202 into rectangular small regions having a preset size. In the following description, this small area is referred to as a “block” as necessary. If the state is a healthy brick surface, the size of the block is set so that one block always has a size that includes joints or cracks. That is, the size of the block is set so that the size of one block is equal to or larger than the size of one refractory brick (the length and width of the surface exposed to the furnace wall). Set.
The first area dividing unit 203 is realized by using, for example, a CPU, a ROM, a RAM, an HDD, and a user interface.

[Second area dividing unit 204]
The second region dividing unit 204 divides the image region of the furnace wall surface thermal image data acquired by the wall surface image acquiring unit 201 into a plurality of blocks. The size and shape of this block are the same as the size and shape of the block when the image region of the post-shading corrected furnace wall thermal image data generated by the shading correction unit 202 is divided by the first region dividing unit 203. To be. That is, the division position by the second region dividing unit 204 and the division position by the first region dividing unit 203 are made the same.
The second area dividing unit 204 is realized by using, for example, a CPU, a ROM, a RAM, an HDD, and a user interface. Note that the function of the second region dividing unit 204 may be included in the first region dividing unit 203.

[Unmeasurable area determination unit 205]
The non-measurable region determination unit 205 derives, as the non-measurable region, a region having an extremely low concentration and a region where the concentration is saturated in the image region of the furnace wall thermal image data acquired by the wall surface image acquisition unit 201. . In the present embodiment, the unmeasurable area determination unit 205 determines whether each of the blocks divided by the second area dividing unit 204 is an unmeasurable area. A specific example of this determination method will be described. First, the non-measurable area determination unit 205 calculates an average density in one block. Then, the non-measurable area determination unit 205 determines that the block is an unmeasurable area when the obtained average density is larger than a preset first threshold value or smaller than a second threshold value. If not, it is determined that the block is not a non-measurable area. For example, 240 can be employed as the first threshold, and 20 can be employed as the second threshold, for example.
The non-measurable area determination unit 205 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Density Histogram Deriving Unit 206]
The density histogram deriving unit 206 derives a density histogram of the furnace wall thermal image data after shading correction for each block that has not been determined as an unmeasurable area by the unmeasurable area determination unit 205. The density histogram shows the relationship between the density of each pixel and the appearance frequency of the density (the number of pixels having the density).
The density histogram deriving unit 206 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Threshold deriving unit 207, peak determining unit 208, pixel number ratio deriving unit 209, wall surface state determining unit 210]
The thermal radiance from the furnace wall of the carbonization chamber is mostly the thermal radiance of refractory bricks and the thermal radiance of carbon. Therefore, in the density histogram of the furnace wall thermal image data after shading correction, peaks due to these thermal radiances occur. That is, when the block is in the state of a sound brick surface, one peak due to the thermal radiance of the refractory brick is obtained in the density histogram of the furnace wall thermal image data after shading correction. In addition, when the block is in a uniform carbon adhesion state, one peak due to the thermal radiance of carbon is obtained. In addition, when the block is in the state of adhering carbon, two peaks are obtained: a peak due to thermal radiance and a peak due to carbon. Thus, in the density histogram of the furnace wall thermal image data after shading correction, one large peak always appears.

The threshold deriving unit 207 derives a threshold based on the Otsu method from the density histogram derived by the density histogram deriving unit 206.
The peak determination unit 208 determines whether there are two peaks in the density histogram derived by the density histogram deriving unit 206.
When the peak determination unit 208 determines that there are not two peaks in the density histogram (that is, when it is determined that there is only one peak in the density histogram), the pixel number ratio deriving unit 209 In the density histogram derived by the deriving unit 206, a ratio r of appearance frequencies of pixels having a density equal to or lower than the threshold value by the Otsu method (= the number of pixels having a density equal to or lower than the threshold value / the total number of pixels) is derived.

  The wall surface state determination unit 210 determines that the block having two peaks in the density histogram derived by the density histogram deriving unit 206 is a spot carbon adhering state region. Further, the wall surface state determination unit 210 determines that a block in which the ratio r derived by the pixel number ratio deriving unit 209 is equal to or greater than a predetermined value (0.7 in the present embodiment) is an area of a healthy brick surface state. . Further, the wall surface state determination unit 210 determines that a block in which the ratio r derived by the pixel number ratio deriving unit 209 is less than a predetermined value is a region having a uniform carbon adhesion state.

  As described in Non-Patent Document 1, Otsu's method assumes that the histogram can be classified into two classes, and determines the threshold value so that the ratio of interclass variance to intraclass variance is maximized. It is. The inventors determined that the threshold value according to the method of Otsu is such that the state of the furnace wall surface of the coking chamber of the coke oven is one of three states: a healthy brick surface state, a uniform carbon adhesion state, and a mottled carbon adhesion state. We have found empirically that it is useful to determine if there is. Otsu's method is generally an algorithm used to determine a threshold for binarization of an image. Therefore, the threshold usage method by the Otsu method in this embodiment is different from the general threshold usage method by the Otsu method.

<In the case of mottled carbon adhesion>
FIG. 5 shows an image (FIG. 5 (a)) obtained by cutting out a block in a mottled carbon adhering state from the furnace wall thermal image data after shading correction, a density histogram in the block, and a threshold value according to the method of Otsu (FIG. 5 (b)). )) Is a diagram showing an example.
As shown in FIG. 5 (b), the present inventors have found that, in the block in the mottled carbon adhering state, there are two peaks in the density histogram with a density near 128 in between. Since there are two peaks in the density histogram, the threshold according to the method of Otsu is obtained near the boundary between these two peaks. In the example shown in FIG. 5B, the threshold value according to the method of Otsu is 135, and there are peaks on both the low concentration side and the high concentration side of the threshold value.

As described above, in the present embodiment, as described above, the threshold deriving unit 207 derives a threshold based on the Otsu method from the density histogram, and the peak determining unit 208 determines whether there are two peaks in the density histogram. Then, the wall surface state determination unit 210 determines that the block having two peaks in the density histogram is the region of the mottled carbon adhesion state.
In this embodiment, it is determined as follows that there are two peaks in the density histogram. First, the peak determining unit 208, the density histogram, the pixel number I _T at the threshold by Otsu method, the maximum value I _L of the number of pixels in the region of lower density side than the threshold value according to the method of Otsu, by Otsu's method A maximum value I _{H of the} number of pixels in the range on the higher density side than the threshold is derived. The peak determination unit 208 can determine that there are two peaks in the density histogram when I _T <I _L and I _T <I _H.

<In case of sound brick surface>
FIG. 6 shows an image (FIG. 6 (a)) obtained by cutting out a block that is in a state of a healthy brick surface from the furnace wall thermal image data after shading correction, a density histogram in the block, and a threshold (FIG. 6 (b) )) Is a diagram showing an example.

  As shown in FIG. 6 (b), the present inventors have found that in a block in a healthy brick surface state, the density is about 128 in most pixels, and the density histogram has one high peak. It was. In addition, the present inventors have found that the density histogram appears over a wider area in the high density area than in the low density area from the peak due to the presence of high density pixels in the area corresponding to the joint. The knowledge that the frequency exceeds 0 (zero) was obtained (see the region 601 shown in FIG. 6B). Therefore, the threshold value according to the method of Otsu is higher than the concentration showing the peak. In the example shown in FIG. 6B, the density indicating the peak is 128, whereas the threshold value according to Otsu's method is 135. As a result, the ratio r of the appearance frequency of pixels having a density equal to or lower than the threshold value according to the method of Otsu was 0.857. The inventors examined this ratio r for a plurality of different blocks. As a result, it was found that the ratio r takes a large value of about 0.8 to 0.9 in the block in the state of a sound brick surface.

  As described above, in the present embodiment, as described above, the threshold deriving unit 207 derives a threshold based on the Otsu method from the density histogram, and the peak determining unit 208 determines whether there are two peaks in the density histogram. The pixel number ratio deriving unit 209 derives the appearance frequency ratio r of pixels having a density equal to or lower than the threshold value by the Otsu method in the density histogram when the density histogram has two peaks (one). Then, the wall surface state determination unit 210 determines that a block having the ratio r of 0.7 or more is a healthy brick surface state region.

<In the case of uniform carbon adhesion>
FIG. 7 shows an image (FIG. 7A) obtained by cutting out a block having a uniform carbon adhesion state from the furnace wall thermal image data after shading correction, a density histogram in the block, and a threshold value according to the method of Otsu (FIG. 7 ( It is a figure which shows an example of b)).
As shown in FIG. 7 (b), the present inventors have a small change in the concentration in the uniformly carbon-attached block. Therefore, the concentration distribution is concentrated in the vicinity of 128 in the concentration histogram, and the substantially symmetrical peak is 1. It was found that there is one. In the example shown in FIG. 7B, the threshold value by the Otsu method is 128, and the appearance frequency ratio r of pixels having a density equal to or lower than the threshold value by the Otsu method is 0.541. The inventors examined this ratio r for a plurality of different blocks. As a result, it was found that the ratio r has a value of about 0.5 in the block in a uniform carbon adhesion state.

  As described above, in the present embodiment, as described above, the threshold deriving unit 207 derives a threshold based on the Otsu method from the density histogram, and the peak determining unit 208 determines whether there are two peaks in the density histogram. The pixel number ratio deriving unit 209 derives the appearance frequency ratio r of pixels having a density equal to or lower than the threshold value by the Otsu method in the density histogram when the density histogram has two peaks (one). Then, the wall surface state determination unit 210 determines that a block having the ratio r of less than 0.7 is a region having a uniform carbon adhesion state.

As described above, since there is always one large peak in the density histogram, in this embodiment, if there are not two peaks in the density histogram, it is assumed that there is only one peak in the density histogram (automatically However, it may be determined as follows, for example, that the density histogram has one peak. First, the peak determining unit 208, the density histogram, the pixel number I _T at the threshold by Otsu method, the maximum value I _L of the number of pixels in the region of lower density side than the threshold value according to the method of Otsu, by Otsu's method A maximum value I _{H of the} number of pixels in the range on the higher density side than the threshold is derived. Then, the peak determination unit 208 determines that the density histogram has one peak when I _T > I _L or I _T > I _H. Here, employ "I _T> I _L or I _T> I _H" instead of "I _T ≧ I _L or I _T> I _H" or "I _T> I _L or I _T ≧ I _H" Also good.

The threshold deriving unit 207, the peak determining unit 208, the pixel number ratio deriving unit 209, and the wall surface state determining unit 210 are the non-measurable region determining unit 205 among the blocks divided by the second region dividing unit 204. This is performed individually for all the blocks that have not been determined as non-measurable areas (that is, all the blocks derived by the density histogram deriving unit 206). As a result, among the blocks divided by the second region dividing unit 204, the carbon adhesion state in all the blocks that have not been determined as non-measurable regions by the non-measurable region determining unit 205 is the sound brick surface state, uniform It is possible to specify which state is the carbon adhesion state or the mottled carbon adhesion state.
The threshold deriving unit 207, the peak determining unit 208, the pixel number ratio deriving unit 209, and the wall surface state determining unit 210 are realized by using, for example, a CPU, a ROM, a RAM, and an HDD.

[Area Ratio Deriving Unit 211]
The region ratio deriving unit 211 determines that the non-measurable region determination unit 205 determines the non-measurable region from among the image regions of the furnace wall thermal image data acquired by the wall surface image acquisition unit 201 based on the determination result in the wall surface state determination unit 210. The ratio of the area of the sound brick surface state to the area excluding the designated area (= area of the sound brick surface area ÷ the area of the image area of the furnace wall thermal image data excluding the area determined to be unmeasurable) To do. In the following description, this ratio is referred to as the area ratio of the area in the state of a healthy brick surface as necessary.
Similarly, the region ratio deriving unit 211 is a region other than the region determined as the non-measurable region by the non-measurable region determining unit 205 among the image regions of the furnace wall surface thermal image data acquired by the wall surface image acquiring unit 201. Percentage of areas with uniform carbon adhesion (= area of uniform carbon adhesion area / area of image area of furnace wall thermal image data excluding areas determined as non-measurable areas) and area of mottled carbon adhesion state (= Area of mottled carbon adhering area / area of image area of furnace wall thermal image data excluding area determined to be unmeasurable area). In the following description, these ratios are referred to as the area ratio of the uniform carbon adhering area and the area ratio of the mottled carbon adhering area as necessary.
The area ratio deriving unit 211 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Wall surface state display unit 212]
The wall surface state display unit 212 includes information on the adhesion state of carbon in each block determined by the wall surface state determination unit 210, information on unmeasurable regions determined by the unmeasurable region determination unit 205, and a region ratio deriving unit. Generates display data for displaying the area ratio of the area in the state of a healthy brick surface, the area ratio of the area in the uniform carbon adhesion state, and the area ratio in the area of the mottled carbon adhesion state, derived by 211. Then, the display data is displayed on the computer display.
FIG. 8 is a diagram showing an example of a display screen of the carbon adhesion state on the furnace wall surface of the coking chamber of the coke oven.

  In the example shown in FIG. 8, on the furnace wall thermal image, the area of the healthy brick surface state, the area of the uniform carbon adhesion state, the area of the mottled carbon adhesion state, and the non-measurable area are displayed, and the state of the healthy brick surface state is displayed. The area ratios of the area (healthy brick area), the mottled carbon adhesion state area (the mottled carbon adhesion area), and the uniform carbon adhesion state area (the uniform carbon adhesion area) are displayed as numbers.

In the example shown in FIG. 8, the region surrounded by the broken line is a region with a uniform carbon adhesion state, the region surrounded by the solid line is a region with a patchy carbon adhesion state, and the region beyond the one-dot chain line is a non-measurable region. Yes, and other areas are areas with a sound brick surface. In addition, the method of displaying the area | region of a healthy brick surface state, the area | region of a uniform carbon adhesion state, the area | region of a mottled carbon adhesion state, and a measurement impossible area | region on a furnace wall surface thermal image is not limited to such a method. For example, you may arrange | position the translucent color which changes for every area | region on a furnace wall surface thermal image.
The wall surface state display unit 212 is realized by using, for example, a CPU, ROM, RAM, HDD, image memory, communication interface, and computer display.

  In addition, information on the adhesion state of carbon in each block, information on non-measurable regions, and information on the area ratio of each region (region of healthy brick surface state, region of mottled carbon adhesion state, region of uniform carbon adhesion state) May not be displayed on the furnace wall surface state determination device 200. For example, the furnace wall surface state determination device 200 may output the information by transmitting the information to an external device or by storing the information in a portable storage medium. In such a case, such information can be displayed by an apparatus other than the furnace wall surface state determination apparatus 200.

[Operation flowchart]
Next, an example of operation | movement of the furnace wall surface state determination apparatus 200 is demonstrated, referring the flowchart of FIG.
First, in step S901, the wall surface image acquisition unit 201 acquires furnace wall surface thermal image data (see FIG. 3).
Next, in step S902, the shading correction unit 202 performs shading correction on the furnace wall thermal image data acquired in step S901, and generates furnace wall thermal image data after shading correction (see FIG. 4).

Next, in step S903, the second region dividing unit 204 divides the image region of the furnace wall thermal image data acquired in step S901 into a plurality of blocks.
Next, in step S904, the unmeasurable area determination unit 205 determines whether the block is an unmeasurable area for each of the plurality of blocks divided in step S903. That is, the unmeasurable area determination unit 205 derives an unmeasurable area from the image area of the furnace wall surface thermal image data acquired in step S901.

  Next, in step S905, the first area dividing unit 203 divides the image area of the post-shading corrected furnace wall thermal image data generated in step S902 into a plurality of blocks. Here, the division position in step S903 and the division position in step S905 are the same. Note that the process of step S905 is not necessarily performed after steps S903 and S904 as long as it is after the process of step S902. For example, in step S903, the process of step S905 may be performed in parallel.

Next, in step S906, the density histogram deriving unit 206 sets “1” to a variable i that designates a block that has not been determined to be an unmeasurable region in step S904 among the plurality of blocks divided in step S905. .
Next, in step S907, the density histogram deriving unit 206 derives a density histogram in a block corresponding to the variable i in the image area of the post-shading corrected furnace wall thermal image data generated in step S902.

Next, in step S908, the threshold deriving unit 207 derives a threshold based on the Otsu method from the density histogram in the block corresponding to the variable i.
Next, in step S909, the peak determining unit 208, the density histogram derived in step S907, the pixel number I _T at the threshold by Otsu method, the number of pixels in the range of the low density side than the threshold value according to the method of Otsu and the maximum value I _L of, than the threshold by Otsu's method to derive the maximum value I _H of the number of pixels in the region of the high concentration side.

Next, in step S910, the peak determination unit 208 determines whether I _T <I _L and I _T <I _H. If it is determined that I _T <I _L and I _T <I _H are not satisfied, the process proceeds to step S914 described later. On the other hand, if I _T <I _L and I _T <I _H , the process proceeds to step S911. In step S911, the wall surface state determination unit 210 determines that there are two peaks in the density histogram derived in step S907, and determines that the block corresponding to the variable i is a patchy carbon adhesion state region. .

  Next, in step S912, the wall surface state determination unit 210 determines whether or not the variable i is N. N is a number equal to the total number of blocks that have not been determined as non-measurable areas in step S904 among the plurality of blocks divided in step S905.

As a result of the determination, if the variable i is N, the determination of the carbon adhesion state is performed for all of the blocks that were not determined as the non-measurable area in step S904 among the plurality of blocks divided in step S905. It ’s gone. Accordingly, the process proceeds to step S918 described later.
On the other hand, when the variable i is not N, the carbon adhesion state is not determined for all the blocks that are not determined to be unmeasurable regions in step S904 among the plurality of blocks divided in step S905. In step S913, the density histogram deriving unit 206 adds “1” to the variable i. And the process after step S907 is performed about the block corresponding to the following variable i.

As described above, if I _T <I _L and I _T <I _H are not satisfied in step S910, it is determined that there is one peak in the density histogram derived in step S907, and the process proceeds to step S914. In step S914, the pixel number ratio deriving unit 208 derives the appearance frequency ratio r of pixels having a density equal to or lower than the threshold value by the Otsu method derived in step S908 in the density histogram derived in step S907. .

Next, in step S915, the wall surface state determination unit 210 determines whether the ratio r derived in step S914 is 0.7 or more. As a result of the determination, if the ratio r derived in step S914 is 0.7 or more, the process proceeds to step S916. If it progresses to step S916, the wall surface state determination part 210 will determine with the block corresponding to the variable i being the area | region of a healthy brick surface state. Then, the process proceeds to step S912 described above.
On the other hand, if the ratio r derived in step S914 is not 0.7 or more, the process proceeds to step S917. In step S917, the wall surface state determination unit 210 determines that the block corresponding to the variable i is a region having a uniform carbon adhesion state. Then, the process proceeds to step S912 described above.

As described above, in step S912, it is determined that the variable i is N, and among the plurality of blocks divided in step S905, all of the blocks that have not been determined as unmeasurable regions in step S904 are attached with carbon. When the state is determined, the process proceeds to step S918.
In step S918, the area ratio deriving unit 211 derives the area ratio of the area in the healthy brick surface state, the area ratio of the area in the uniform carbon adhesion state, and the area ratio of the area in the mottled carbon adhesion state.

Next, in step S919, the wall surface state display unit 212 includes information on the carbon adhesion state in each block determined in steps S911, S916, and S917, and information on the unmeasurable region derived in step S904, Display data for displaying the area ratio information of each region (region with a healthy brick surface state, region with a uniform carbon adhesion state, region with a mottled carbon adhesion state) derived in step S918 is generated.
Next, in step S920, the wall surface state display unit 212 displays an image based on the display data generated in step S919 on the computer display. And the process by the flowchart of FIG. 9 is complete | finished.

[Summary]
As described above, in the present embodiment, the furnace wall thermal image data after shading correction is divided into a plurality of blocks, and the density histogram of the furnace wall thermal image data after shading correction is derived and derived for each of the divided blocks. The threshold value by the method of Otsu is derived from the density histogram. Then, the block having two peaks in the density histogram is determined to be in the state of adhering carbon. If there are not two peaks in the density histogram, it is determined that a block whose appearance frequency ratio of pixels having a density equal to or lower than the threshold value by the Otsu method is 0.7 or more in the density histogram is in a sound brick surface state. To do. On the other hand, in the density histogram, a block in which the ratio of the appearance frequency of pixels having a density equal to or lower than the threshold according to the method of Otsu is less than 0.7 is determined to be in a uniform carbon adhesion state. And while displaying the area | region of each state on a furnace wall surface thermal image, the area ratio of the area | region of each state is displayed.

  Therefore, it is possible to automatically and accurately grasp the adhesion state and position of carbon on the furnace wall surface of the coking chamber of the coke oven. Since it becomes possible to quantify the carbon adhesion region much more quantitatively than the conventional judgment by visual observation of the furnace wall thermal image, it is possible to improve the accuracy of management of the carbon adhesion state. The amount of carbon deposited gradually varies over the course of the day depending on operating conditions such as furnace temperature and coal blending. If the carbon adhesion region tends to increase after the change of the operating conditions, the carbon will grow excessively and become uneven, and the extrusion load may increase and clogging may occur. On the contrary, if the tendency of carbon to continue decreasing, even the carbon which filled the recessed part by the thinning of a refractory brick or a fine corner chip | tip will lose | disappear, and an extrusion load will rise. Based on the information indicating in which region of the furnace wall the carbon adhesion state is based on the method described in the present embodiment, for example, the time of carbon incineration in which air is blown into the carbonization chamber between operations If is adjusted, the extrusion load can be stabilized accurately.

[Modification]
As in the present embodiment, it is preferable to determine the unmeasurable region and determine the carbon adhesion state in the region excluding this non-measurable region because the carbon adhesion state on the entire furnace wall surface can be appropriately determined. However, it is not always necessary to determine the unmeasurable area. For example, a region assumed as a non-measurable region may be set in advance, and the carbon adhesion state in a region excluding this region may be determined from the image region of the furnace wall thermal image data. In addition, in the case where the carbon adhesion state is determined only for a region that is not assumed as a non-measurable region, the non-measurable region may not be set.

  In addition, if the shading correction is performed on the furnace wall surface thermal image data as in the present embodiment, the carbon adhesion state can be appropriately determined even if the furnace wall surface thermal image data has uneven density. It is preferable because it is possible. However, when the influence of such density unevenness is small, the shading correction is not necessarily performed on the furnace wall surface thermal image data.

[Example]
The furnace wall surface thermal image of the carbonization chamber of the coke oven in operation was obtained, and the carbon adhesion state was determined by the method described in this embodiment. The coke oven to be used for determining the carbon adhesion state was an aging furnace that had been operating for over 40 years. The furnace wall surface of the coking chamber of the coke oven has a depth of about 16 m and a height of about 6 m. This furnace wall thermal image of the entire furnace wall surface was subjected to determination with a 256-gradation furnace wall thermal image with a resolution in the depth direction of 2 mm and a resolution in the height direction of 1.3 mm. The furnace wall surface thermal image and the shading-corrected furnace wall surface thermal image generated from the furnace wall surface thermal image are divided into rectangular blocks each having 128 pixels in the vertical direction and 128 pixels in the horizontal direction. The determination was made according to the flowchart of FIG. Such a measurement was performed on the furnace wall surfaces of two different carbonization chambers. FIG. 10 is a diagram (photograph) showing an example of the result of determination of the adhesion state of carbon on the furnace wall surface of the carbonization chamber. In FIG. 10, as in FIG. 8, a region surrounded by a broken line is a region with a uniform carbon adhesion state, a region surrounded by a solid line is a region with a patchy carbon adhesion state, and a region beyond the one-dot chain line is an unmeasurable region. The other areas are defined as areas with a sound brick surface.
8 and 10, it can be seen that both the wide-area carbon adhesion state and the mottled carbon adhesion state that can be seen by humans are almost correctly extracted.

(Second Embodiment)
Next, a second embodiment of the present invention will be described. In the first embodiment, the case where the adhesion state of carbon in each block of the furnace wall thermal image data after shading correction is determined as an example using the threshold value by the method of Otsu as a determination criterion has been described as an example. On the other hand, in the present embodiment, shading correction is performed using the distance between the center of density distribution on each of the lower density side and the higher density side than the density showing the peak in the density histogram and the density showing the peak as a criterion. The case where the adhesion state of carbon in each block of the rear furnace wall surface thermal image data is determined will be described. As described above, the present embodiment and the first embodiment are different in a part of the method for determining the carbon adhesion state in each block of the furnace wall thermal image data after shading correction. Therefore, in the description of the present embodiment, the same parts as those in the first embodiment are denoted by the same reference numerals as those in FIGS.

  FIG. 11 is a diagram illustrating an example of a functional configuration of the furnace wall surface state determination device 1100 of the coking chamber of the coke oven. The hardware of the coke oven coking chamber furnace wall surface state determination device 1100 includes, for example, a CPU, ROM, RAM, HDD, image input / output board, and an information processing device including various interfaces and dedicated hardware. It is realizable by using. Moreover, in the following description, the furnace wall surface state determination device of the coking chamber of the coke oven is abbreviated as a furnace wall surface state determination device as necessary.

[Wall surface image acquisition unit 201, shading correction unit 202, second region dividing unit 204, non-measurable region determining unit 205, first region dividing unit 203, density histogram deriving unit 206]
The wall surface image acquisition unit 201, the shading correction unit 202, the second region dividing unit 204, the non-measurable region determining unit 205, the first region dividing unit 203, and the density histogram deriving unit 206 are the furnace wall according to the first embodiment. This is the same as that in the surface state determination apparatus 200.

The outline will be described again. The wall surface image acquisition unit 201 acquires furnace wall surface thermal image data (see FIG. 3). The shading correction unit 202 performs shading correction on the furnace wall surface thermal image data acquired by the wall surface image acquisition unit 201 (see FIG. 4).
The first region dividing unit 203 divides the image region of the post-shading corrected furnace wall thermal image data generated by the shading correction unit 202 into a plurality of blocks. The second region dividing unit 204 divides the image region of the furnace wall surface thermal image data acquired by the wall surface image acquiring unit 201 into a plurality of blocks. The non-measurable region determination unit 205 derives, as the non-measurable region, a region having an extremely low concentration and a region where the concentration is saturated in the image region of the furnace wall thermal image data acquired by the wall surface image acquisition unit 201. .
The density histogram deriving unit 206 calculates the shading-corrected furnace wall thermal image data for each block that has not been determined by the non-measurable region determination unit 205 among the blocks divided by the second region dividing unit 204. A density histogram is derived.

[Peak concentration deriving unit 1101]
The peak density deriving unit 1101 derives a peak density C _p that is the density at which the number of pixels is maximized in the density histogram derived by the density histogram deriving unit 206.
The peak concentration deriving unit 1101 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Peak determination unit 1102]
Peak determining unit 1102 determines, in the region of the high density side than the peak concentration C _p, whether or not there is a maximum value.
12, in the region of the high-density side than the peak concentration C _p, is a diagram conceptually illustrating an example of a method of determining whether there is a maximum value.
For example, as shown in FIG. 12, the peak determining unit 1102 increases the concentration by minute concentration intervals starting from the peak concentration C _p (see the arrow line shown in FIG. 12), and the increased concentration (see FIG. 12). The change in the number of pixels in the density of the portion surrounded by 12 circles) is read. Then, the peak determination unit 1102 determines whether or not there is a region where the number of pixels increases as the density increases from the change in the number of pixels read.

As shown in FIG. 12, if there is a region on the higher density side than the peak density C _{p and the} number of pixels increases as the density increases, the density histogram is displayed in the higher density side area than the peak density C _p. There will be a local maximum. On the other hand, as shown in FIGS. 13A and 13B, which will be described later, if there is no area where the number of pixels increases as the density increases, the density histogram of the density density area is higher than the peak density C _p . The number of pixels decreases monotonously.
The method for determining whether or not there is a maximum value in the density histogram is not limited to the method described above. For example, there is a region where the number of pixels increases as the density is higher in a region on the higher density side than the peak density C _p , and a region where the number of pixels decreases as the density is higher in a region on the higher density side than that region. In some cases, it may be determined that a maximum value exists in the density histogram.
Alternatively, a threshold value for the increase in the number of pixels may be set, and it may be determined that the number of pixels has increased only when the increase in the number of pixels is equal to or greater than the threshold value.
The peak determination unit 1102 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Center of gravity deriving unit 1103]
The center-of-gravity deriving unit 1103 has no maximum value in the region on the higher density side than the peak concentration C _p by the peak determining unit 1102 (the number of peaks in the density histogram derived by the density histogram deriving unit 206 is one). when it is determined that, than the peak concentration C _p derived by the peak concentration derivation section 1101 for the density histogram to derive the concentration of the center of gravity of the density distribution in the region of the low concentration side. In the following description, this density is referred to as a low density side gravity center density G _L as necessary.
The center-of-gravity deriving unit 1103 derives a density that is the center of density distribution in a region on the higher density side than the peak density C _p derived by the peak density deriving unit 1101 for the density histogram. In the following description, this concentration is referred to as a high concentration side gravity center concentration _GH as necessary.
The low density side centroid density G _L is a density that divides the area of the area on the low density side from the peak density C _p in half in the density histogram area, and the high density side centroid density _GH is the density histogram. Is a concentration that divides the area of the region on the higher concentration side than the peak concentration C _p in half.

FIG. 13 is a diagram illustrating an example of the peak density C _p , the low density side centroid density G _L , and the high density side centroid density _{GH in} the density histogram. Specifically, FIG. 13A shows a peak density C _p , a low density side centroid density G _L , and a high density side centroid density in the density histogram of the block shown in FIG. _GH is shown. FIG. 13B shows the peak concentration C _p , the low concentration side centroid concentration G _L , and the high concentration side centroid concentration in the concentration histogram of the block shown in FIG. _GH is shown.
In the example shown in FIG. 13A, the peak density C _p = 126, the low density side centroid density G _L = 12.3, and the high density side centroid density G _H = 134.0. In the example shown in FIG. 13B, the peak density C _p = 127, the low density side centroid density G _L = 12.9, and the high density side centroid density G _H = 132.3.

Next, the gravity center deriving unit 1103 derives a low concentration side distance D _L that is a distance between the low concentration side gravity center concentration G _L and the peak concentration C _p . Further, the centroid deriving unit 1103 derives a high concentration side distance D _H that is a distance between the high concentration side centroid concentration G _H and the peak concentration C _p . The center-of-gravity deriving unit 1103 derives a distance difference (= D _H −D _L ) that is a value obtained by subtracting the low concentration side distance D _L from the high concentration side distance D _H. Specifically, in the example shown in FIG. 13A, the distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H, is “4.3. " In the example shown in FIG. 13B, the distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H, is “1.2”. Become.

Next, the center-of-gravity deriving unit 1103 determines whether or not the distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H , is equal to or less than the threshold value ΔD (D _H− D _L ≦ ΔD is satisfied). The threshold value ΔD is a tolerance for considering that the low density side gravity center density D _L and the high density side gravity center density D _H are in positions symmetrical with respect to the peak density C _p . The threshold value ΔD is, for example, based on a result obtained by comparing a density histogram obtained from the result of actual operation with the shading-corrected furnace wall thermal image data when the density histogram is obtained for a number of density histograms. It can be determined as appropriate. As the threshold ΔD, for example, “3” can be adopted.
In the density histogram of the area in the state of the healthy brick surface, the high density side distance D _H is larger than the low density side distance D _L due to the presence of high density pixels in the area corresponding to the joint. Therefore, the center-of-gravity deriving unit 1103 determines whether or not a distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H , is equal to or less than the threshold value ΔD.
The center-of-gravity deriving unit 1103 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Wall surface state determination unit 1104]
The wall surface state determination unit 1104 determines a block that has been determined by the peak determination unit 1102 to have a maximum value in the concentration histogram (that is, there are two peaks) in an area higher than the peak concentration C _p . It is determined that the region is a spotted carbon adhesion state. This is because, as described above, in the block in the mottled carbon adhering state, it has been found that there are two peaks in the density histogram with a density near 128 (see FIGS. 5B and 12).
Further, the wall surface state determination unit 1104 has a distance difference (= D _H −D _L ) that is a value obtained by subtracting the low concentration side distance D _L from the high concentration side distance D _H by the gravity center deriving unit 1102 being equal to or less than the threshold value ΔD. It is determined that the block determined to be a region having a uniform carbon adhesion state. As described in the first embodiment, since the change in the concentration in the block in the uniform carbon adhesion state is small, the concentration distribution is concentrated in the vicinity of 128 in the concentration histogram, and there is one approximately symmetrical peak. (See FIGS. 7 (b) and 13 (b)).

Further, the wall surface state determination unit 1104 has a distance difference (= D _H −D _L ) that is a value obtained by subtracting the low concentration side distance D _L from the high concentration side distance D _H by the gravity center deriving unit 1102 being equal to or less than the threshold value ΔD. Blocks that are not determined as (that is, blocks whose distance difference (= D _H −D _L ) is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H ) is a healthy brick It is determined that this is a state area. As described above, in the block that is in the state of a healthy brick surface, the density is about 128 in most pixels, and the area on the high density side is wider than the area on the low density side than the peak. This is because the appearance frequency exceeds 0 (zero) (see FIG. 6B and FIG. 13A).
The wall surface state determination unit 1104 is realized by using, for example, a CPU, ROM, RAM, and HDD.

  The peak concentration deriving unit 1101, the peak determining unit 1102, the center of gravity deriving unit 1103, and the wall surface state determining unit 1104 are determined by the non-measurable region determining unit 205 among the blocks divided by the second region dividing unit 204. This is performed individually for all blocks that are not determined to be unmeasurable regions (that is, all blocks derived by the density histogram deriving unit 206). As a result, among the blocks divided by the second region dividing unit 204, the carbon adhesion state in all the blocks that have not been determined as non-measurable regions by the non-measurable region determining unit 205 is the sound brick surface state, uniform It can be specified which state is the carbon adhesion state or the mottled carbon adhesion state.

[Area Ratio Deriving Unit 1105]
From the determination result in the wall surface state determination unit 1104, the region ratio deriving unit 1105 is similar to the region ratio deriving unit 211 of the furnace wall surface state determination device 200 of the first embodiment, and the area ratio of the area in the healthy brick surface state Then, the area ratio of the area in the uniform carbon adhesion state and the area ratio of the area in the mottled carbon adhesion state are derived.
The area ratio deriving unit 1105 is realized by using, for example, a CPU, ROM, RAM, and HDD.

[Wall surface state display unit 1106]
The wall surface state display unit 1106 includes information on the adhesion state of carbon in each block determined by the wall surface state determination unit 1104, information on the unmeasurable region determined by the unmeasurable region determination unit 205, and an area ratio deriving unit. Generate display data for displaying the area ratio of the area in the state of the sound brick surface, the area ratio of the area in the uniform carbon adhesion state, and the area ratio in the area of the mottled carbon adhesion state derived by 1105 Then, the display data is displayed on the computer display.

The wall surface display unit 1106 can also display the same information as the wall surface state display unit 212 of the furnace wall surface state determination device 200 of the first embodiment.
In addition, information on the carbon adhesion state in each block, information on non-measurable areas, and information on the area ratio of each area (area with healthy brick surface, area with mottled carbon adhesion, area with uniform carbon adhesion) May not be displayed on the furnace wall surface state determination device 1100. For example, the furnace wall surface state determination device 1100 may output the information by transmitting the information to an external device or storing the information in a portable storage medium. In such a case, such information can be displayed by an apparatus other than the furnace wall surface state determination apparatus 1100.

[Operation flowchart]
Next, an example of operation | movement of the furnace wall surface state determination apparatus 1100 is demonstrated, referring the flowchart of FIG.
First, in step S1401, the wall surface image acquisition unit 201 acquires furnace wall surface thermal image data (see FIG. 3).
Next, in step S1402, the shading correction unit 202 performs shading correction on the furnace wall thermal image data acquired in step S1401, and generates furnace wall thermal image data after shading correction (see FIG. 4).

Next, in step S1403, the second region dividing unit 204 divides the image region of the furnace wall thermal image data acquired in step S1401 into a plurality of blocks.
Next, in step S1404, the unmeasurable area determination unit 205 determines whether the block is an unmeasurable area for each of the plurality of blocks divided in step S1403.

  Next, in step S1405, the first area dividing unit 203 divides the image area of the post-shading corrected furnace wall thermal image data generated in step S1402 into a plurality of blocks. Note that the processing in step S1405 is not necessarily performed after steps S1403 and S1404 as long as it is after the processing in step S1402. This is the same as the processing in step S905 in the flowchart of FIG.

Next, in step S1406, the density histogram deriving unit 206 sets “1” to a variable i that designates a block that has not been determined to be an unmeasurable region in step S1404 among a plurality of blocks divided in step S1405. .
Next, in step S1407, the density histogram deriving unit 206 derives a density histogram in a block corresponding to the variable i in the image region of the post-shading corrected furnace wall thermal image data generated in step S1402.

Next, in step S1408, the peak density deriving unit 1101 derives the peak density C _p of the density histogram in the block corresponding to the variable i.
Next, in step S1409, the peak determining unit 1102, than the peak concentration C _p of the density histogram in the block corresponding to the variable i in the region of the high-density side, whether there is a maximum value.

The result of this determination, in the area of the high concentration side than the peak concentration C _p of the density histogram in the block corresponding to the variable i, determined when there is no local maximum value, the number of peaks in the density histogram is one Then, the process proceeds to step S1413 described later. On the other hand, if there is a maximum value in the region on the higher density side than the peak density C _p of the density histogram in the block corresponding to the variable i, the process proceeds to step S1410. When the processing proceeds to step S1410, the wall surface state determination unit 1104 determines that the block corresponding to the variable i is an area of the mottled carbon adhesion state.
Next, in step S1411, the wall surface state determination unit 210 determines whether or not the variable i is N. N is a number equal to the total number of blocks that have not been determined as unmeasurable areas in step S1404 among the plurality of blocks divided in step S1405.

As a result of the determination, if the variable i is N, the determination of the carbon adhesion state is performed for all of the blocks divided in step S1405 that have not been determined as unmeasurable regions in step S1404. It ’s gone. Accordingly, the process proceeds to step S1419 described later.
On the other hand, when the variable i is not N, the carbon adhesion state is not determined for all the blocks that are not determined as the non-measurable region in step S1404 among the plurality of blocks divided in step S1405. In step S1412, the density histogram deriving unit 206 adds “1” to the variable i. And the process after step S1407 is performed about the block corresponding to the following variable i.

In step S1409, as described above, in the region of the high density side than the peak concentration C _p of the density histogram in the block corresponding to the variable i, if it is determined that there is no local maximum value, the process proceeds to step S1413. In step S1413, the centroid derivation unit 1103 derives the high density side centroid density _GH and the low density side centroid density _GL of the density histogram in the block corresponding to the variable i.
Next, in step S1414, the center of gravity derivation unit 1103, and the high density side distance D _H which is a distance of the high concentration side centroid concentration G _H and the peak concentration C _p, of the low density side center of gravity concentration G _L and the peak concentration C _p A low concentration side distance D _L which is a distance is derived.

Next, in step S1415, the center-of-gravity deriving unit 1103 derives a distance difference (= D _H −D _L ) that is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H.
Next, in step S1416, the center-of-gravity deriving unit 1103 determines whether or not the distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H , is equal to or less than the threshold value ΔD. Determine whether. As a result of the determination, if the distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H , is equal to or less than the threshold value ΔD, the process proceeds to step S1417.

In step S1417, the wall surface state determination unit 1104 determines that the block corresponding to the variable i is a region having a uniform carbon adhesion state. Then, the process proceeds to step S1411 described above.
On the other hand, if it is determined in step S1416 that the distance difference (= D _H −D _L ), which is a value obtained by subtracting the low density side distance D _L from the high density side distance D _H , is not less than or equal to the threshold value ΔD, step S1418 is performed. Proceed to If it progresses to step S1418, the wall surface state determination part 1104 will determine with the block corresponding to the variable i being the area | region of a healthy brick surface state. Then, the process proceeds to step S1411 described above.

As described above, in step S1411, it is determined that the variable i is N, and out of the plurality of blocks divided in step S1405, all of the blocks that are not determined as non-measurable regions in step S1404 are attached with carbon. When the state is determined, the process proceeds to step S1419.
In step S1419, the area ratio deriving unit 1105 derives the area ratio of the area in the healthy brick surface state, the area ratio of the area in the uniform carbon adhesion state, and the area ratio of the area in the mottled carbon adhesion state.

Next, in step S1420, the wall surface state display unit 1106 displays information on the carbon adhesion state in each block determined in steps S1410, S1417, and S1418, information on the unmeasurable area derived in step S1404, Display data for displaying the area ratio information of each region (region with a healthy brick surface state, region with a uniform carbon adhesion state, region with a mottled carbon adhesion state) derived in step S1419 is generated.
Next, in step S1421, the wall surface state display unit 1106 displays an image based on the display data generated in step S1420 on the computer display. And the process by the flowchart of FIG. 14 is complete | finished.

[Summary]
As described above, in the present embodiment, a block having two peaks in the density histogram is determined to be a patchy carbon adhesion state region. Further, the low concentration side distance D _L that is the distance between the low concentration side gravity center concentration G _L and the peak concentration C _p is subtracted from the high concentration side distance D _H that is the distance between the high concentration side gravity center concentration _GH and the peak concentration C _p. A block having a distance difference (= D _H −D _L ) equal to or less than a threshold value ΔD is determined to be a region having a uniform carbon adhesion state. Further, is one peak in the density histogram, from the high concentration side distance D _H which is a distance of the high concentration side centroid concentration G _H and the peak concentration C _p, the low concentration side distance of the center of gravity concentration G _L and the peak concentration C _p A block having a distance difference (= D _H −D _L ) that is a value obtained by subtracting the low density side distance D _L being greater than the threshold value ΔD is determined to be a healthy brick surface area. And while displaying the area | region of each state on a furnace wall surface thermal image, the area ratio of the area | region of each state is displayed.
Therefore, also in this embodiment, the same effect as that described in the first embodiment can be obtained.

[Modification]
Also in this embodiment, as described in the first embodiment, it is not always necessary to determine an unmeasurable region. Further, it is not always necessary to perform shading correction on the furnace wall thermal image data.

In the first embodiment, in the density histogram, the appearance frequency ratio r of pixels having a density equal to or lower than the threshold by the method of Otsu is 0.7 or more, and the state of the block to be processed is a sound brick surface. It has been described as an example of the first condition for determining that the state is in effect. Moreover, I _T <I _L and I _T <I _H have been described as an example of the second condition for determining that the state of the block to be processed is a mottled carbon adhesion state.

On the other hand, in the second embodiment, the distance difference, which is the value obtained by subtracting the low concentration side distance D _L from the high concentration side distance D _H described above, exceeds the threshold ΔD, and the state of the block to be processed is sound brick. It has been described as an example of the first condition for determining that the surface state is present. Further, in the region of the high density side than the peak concentration C _p of the density histogram, that there is a maximum value, as an example of the second condition for the state of the block to be processed is determined to be a patchy carbon adhesion state explained.

  However, the first condition is a condition that indicates that the number of pixels at a density value farther from the peak is higher on the higher density side than the peak in the density histogram as compared with the lower density side than the peak. If present, the first condition is not limited to the above-described condition. Further, the second condition is not limited to the above-described condition as long as there are two peaks in the density histogram.

The embodiment of the present invention described above can be realized by a computer executing a program. Further, a computer-readable recording medium in which the program is recorded and a computer program product such as the program can also be applied as an embodiment of the present invention. As the recording medium, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.
In addition, the embodiments of the present invention described above are merely examples of implementation in carrying out the present invention, and the technical scope of the present invention should not be construed as being limited thereto. Is. That is, the present invention can be implemented in various forms without departing from the technical idea or the main features thereof.

  200, 1100: Furnace wall surface state determination device for coking chamber of coke oven, 201: Wall surface image acquisition unit, 202: Shading correction unit, 203: First region dividing unit, 204: Second region dividing unit, 205: Unmeasurable area determination unit, 206: density histogram deriving unit, 207: threshold deriving unit, 208: peak determining unit, 209: pixel number ratio deriving unit, 210, 1104: wall surface state determining unit, 211, 1105: area ratio deriving unit 212, 1106: wall surface state display unit, 1101: peak concentration deriving unit, 1102: peak determining unit, 1103: barycentric deriving unit

Claims (8)

A wall surface image acquisition means for acquiring a distribution of thermal radiance of a furnace wall surface of a coking chamber of a coke oven as a thermal image composed of pixels having a concentration value corresponding to the thermal radiance;
First area dividing means for dividing the thermal image acquired by the wall surface image acquiring means into a plurality of areas;
A density histogram deriving unit for deriving a density histogram indicating a relationship between the density and the number of pixels of each pixel of the thermal image for each of the plurality of regions divided by the first region dividing unit;
Based on the shape of the concentration histogram derived by the concentration histogram deriving means, the state of the furnace wall surface is one of a healthy brick surface state, a mottled carbon adhesion state, and a uniform carbon adhesion state. And wall surface state determining means for determining each of the plurality of regions,
The sound brick surface state is a state in which the furnace wall surface is exposed, and is a state in which carbon is clogged linearly in joints and cracks of the refractory brick constituting the furnace wall surface,
The uniform carbon adhering state is a state in which carbon is adhering substantially uniformly so as to cover the furnace wall surface.
The mottled carbon adhering state is a state in which carbon having a width equal to or greater than the width of the joints and cracks is attached to a region including the joints and cracks of the refractory bricks constituting the furnace wall surface,
The wall surface state determining means determines that the state of the region that satisfies the condition that the peak in the density histogram is one and satisfies the first condition is the healthy brick surface state,
The condition of the region that satisfies the condition that there is one peak in the concentration histogram and does not satisfy the first condition is determined as the uniform carbon adhesion state,
The state of the region that satisfies the second condition that there are two peaks in the density histogram is determined as the mottled carbon adhesion state,
The first condition is a condition indicating that the number of pixels at a density value farther from the peak is higher on the higher density side than the peak in the density histogram as compared with the lower density side than the peak. A furnace wall surface state determination device for a coking chamber of a coke oven. Threshold deriving means for deriving a concentration threshold from the density histogram derived by the density histogram deriving means by the method of Otsu;
Peak determination means for determining whether or not one peak in the density histogram derived by the density histogram deriving means exists on each of the high density side and the low density side than the threshold value according to the method of Otsu;
In the density histogram derived by the density histogram deriving unit, the number of pixels having a density equal to or lower than the threshold value by the Otsu method derived by the threshold deriving unit is divided by the number of all pixels in the density histogram. A pixel number ratio deriving unit for deriving a value;
The first condition is a condition that the value derived by the pixel number ratio deriving unit is equal to or greater than a threshold value.
2. The condition according to claim 1, wherein the second condition is a condition in which one peak in the density histogram exists on each of a higher density side and a lower density side than a threshold value according to the method of the Otsu method. For determining the furnace wall surface state of the coking chamber of a coke oven. Peak concentration deriving means for deriving a peak concentration which is a concentration indicating a peak in the density histogram derived by the concentration histogram deriving means;
Peak determination for determining whether the density histogram derived by the density histogram deriving means has a maximum value in a region on the higher density side than the peak density derived by the peak density deriving means for the density histogram Means,
From the density histogram derived by the density histogram deriving means, the high density side centroid density and the peak density, which are the density that is the centroid of the density distribution on the higher density side than the peak density derived by the peak density deriving means. The high concentration side distance that is the distance to the peak concentration, and the low concentration side centroid concentration that is the concentration that is the centroid of the concentration distribution on the lower concentration side than the peak concentration derived by the peak concentration deriving means and the peak concentration A center-of-gravity deriving unit that derives a low-concentration side distance that is a distance and determines whether or not a distance difference that is a value obtained by subtracting the low-concentration side distance from the derived high-concentration side distance is equal to or less than a threshold value. ,
The first condition is a condition that the distance difference is greater than a threshold value,
The furnace wall surface state determination apparatus for a coke oven in a coke oven according to claim 1, wherein the second condition is a condition that the maximum value exists. Further comprising shading correction means for performing shading correction on the thermal image acquired by the wall surface image acquisition means;
The first region dividing unit divides the thermal image after the shading correction is performed by the shading correction unit into a plurality of regions,
The density histogram deriving unit includes a density and a pixel number for each pixel of the thermal image after the shading correction is performed by the shading correction unit for each of a plurality of regions divided by the first region dividing unit. The apparatus according to claim 1, wherein a concentration histogram indicating the relationship between the coke ovens is derived. Second area dividing means for dividing the thermal image acquired by the wall surface image acquiring means into a plurality of areas;
Whether or not the area is a non-measurable area based on the density value of the thermal image acquired by the wall surface image acquisition means and based on the density value of the area divided by the second area dividing means And a non-measurable area determination means that performs the determination for each of the plurality of areas divided by the second area dividing means,
The division position of the thermal image after the shading correction by the first area dividing means and the division position of the thermal image by the second area dividing means are the same,
The density histogram deriving unit performs the shading correction for each of the plurality of regions divided by the second region dividing unit, for each region that has not been determined to be the non-measurable region by the non-measurable region determining unit. 5. A coke oven furnace according to claim 4, wherein a density histogram indicating a relationship between the density and the number of pixels of the thermal image after the shading correction is performed by means is derived. Wall surface condition determination device.   Based on the result of the determination by the wall surface state determination means, the area ratio of the areas of the healthy brick surface state, the mottled carbon adhesion state, and the uniform carbon adhesion state on the furnace wall surface of the coking chamber of the coke oven is calculated. The furnace wall surface state determination device for a coking chamber of a coke oven according to any one of claims 1 to 5, further comprising an area ratio deriving unit for deriving. A wall surface image acquisition step of acquiring a distribution of thermal radiance of the coke oven carbonization chamber wall surface as a thermal image composed of pixels having density values corresponding to the thermal radiance, and
A first region dividing step of dividing the thermal image acquired by the wall surface image acquiring step into a plurality of regions;
A density histogram deriving step for deriving a density histogram indicating a relationship between the density and the number of pixels of each pixel of the thermal image for each of the plurality of regions divided by the first region dividing step;
Based on the shape of the concentration histogram derived by the concentration histogram deriving step, the state of the furnace wall surface is any of a healthy brick surface state, a mottled carbon adhesion state, and a uniform carbon adhesion state. And a wall surface state determining step for determining for each of the plurality of regions,
The sound brick surface state is a state in which the furnace wall surface is exposed, and is a state in which carbon is clogged linearly in joints and cracks of the refractory brick constituting the furnace wall surface,
The uniform carbon adhering state is a state in which carbon is adhering substantially uniformly so as to cover the furnace wall surface.
The mottled carbon adhering state is a state in which carbon having a width equal to or greater than the width of the joints and cracks is attached to a region including the joints and cracks of the refractory bricks constituting the furnace wall surface,
The wall surface state determining step satisfies the condition that there is one peak in the density histogram, and determines the state of the region that satisfies the first condition as the healthy brick surface state,
The condition of the region that satisfies the condition that there is one peak in the concentration histogram and does not satisfy the first condition is determined as the uniform carbon adhesion state,
The state of the region that satisfies the second condition that there are two peaks in the density histogram is determined as the mottled carbon adhesion state,
The first condition is a condition indicating that the number of pixels at a density value farther from the peak is higher on the higher density side than the peak in the density histogram as compared with the lower density side than the peak. A method for determining a furnace wall surface state of a coking chamber of a coke oven.   A program that causes a computer to function as each means of the furnace wall surface state determination device for a coking chamber of a coke oven according to any one of claims 1 to 6.