JP6923180B2 - Uneven distribution detection method for granular matter in the observation object - Google Patents

Description

The present invention relates to a method for detecting uneven distribution of granules in an object to be observed.

In the fields of pharmaceuticals and foods, tablets obtained by mixing a plurality of raw materials and molding the mixture are used. As a raw material for tablets, excipients are appropriately added in addition to the active ingredient. Excipients are additives that are added to improve handleability, moldability, or make it easier to take. In addition, as necessary, the raw materials for tablets include binders, auxiliaries (for example, solubilizers, solubilizers, disintegrants, etc.), flavoring agents, colorants, pharmaceutical additives, antioxidants, preservatives, etc. Shading agents, antioxidants, fragrances, sweeteners, fluidizing agents, flavoring agents, lubricants, etc. are added.

When the raw materials of the tablets are not chemically soluble, the tablets are a mixture when observed macroscopically, but the granules which are the raw materials are dispersed when observed microscopically. It is preferable that the active ingredients and the like in the tablets are uniformly dispersed, but in reality, they may be unevenly distributed. Therefore, it is required to quantitatively evaluate the degree of uneven distribution of active ingredients and the like. If the degree of uneven distribution can be evaluated quantitatively, a standard value can be set. Therefore, for example, in quality control, review the mixing of raw materials and kneading conditions that affect the degree of uneven distribution in the manufacturing process based on the quantitative result. It is considered that the uneven distribution of active ingredients and the like can be reduced.

By the way, as a method for evaluating the dispersibility of an object, a technique for Fourier transforming an image visualized by imaging or the like is generally known (for example, Non-Patent Documents 1 and 2). The Fourier transform is a conversion formula between the spatial domain (waveform) and the spatial frequency domain (waveform) using the Fourier integral. By the Fourier transform, the fineness and uniformity of an object can be evaluated, and a specific frequency component can be evaluated. The intensity (peak) of the above or the average value in a specific frequency region can be used as a quantitative value. Therefore, it is conceivable to evaluate the degree of uneven distribution of granules such as active ingredients contained in tablets by using the Fourier transform.

Masaru Terada, Hiroyuki Aoki, Masatoshi Chika, Seiji Yokozawa, "Evaluation of Dispersivity of Bright Material in Metallic Coating Film by Texture Analysis Method Using Two-Dimensional Fourier Transform", Color Material, 2004, 77 (1), p. 7-12 Manabu Kanematsu, Ryoma Kitagaki, Takafumi Noguchi, Fumiki Tomozawa, "Basic Research on Evaluation Method of Concrete Stain by Two-Dimensional Fourier Transform", Annual Proceedings of Concrete Engineering, 2000, Vol. 22, No. 1

The Fourier transform is a method of calculating an infinite region with respect to the spatial direction, but since the actual data to be calculated is a finite range, the calculation is performed assuming that the finite range is repeated in the spatial direction. Will be done. Therefore, when the visualized image having the uneven distribution portion is Fourier transformed, the calculation is performed based on the assumption that the uneven distribution portion is repeated infinitely, and the uneven distribution portion is calculated (detected) as a low frequency frequency component. However, such low-frequency frequency components obtained by the Fourier transform include components based on the presence of uneven distribution, components based on low uniformity of particles (large variation in particle size), and particles. Information on the coarseness of the object is also detected. Therefore, based on the result obtained by the Fourier transform, the same result is obtained regardless of whether the fine particles are unevenly distributed or the coarse particles are uniformly dispersed, and these cannot be distinguished. Therefore, the method of quantifying the degree of uneven distribution using the Fourier transform on an object in which particles of various sizes are mixed, such as a tablet, is inferior in accuracy.

The present invention has been made by paying attention to the above circumstances, and an object of the present invention is to provide a method for detecting uneven distribution of granules without being affected by the particle size of granules in an observation object. be. Another object of the present invention is to provide a program for causing a computer to execute the uneven distribution detection method, and a computer-readable recording medium in which the program is stored.

The first method according to the present invention that has solved the above problems is a method of detecting uneven distribution of granules in an observation object, and at least a part of the outline of the observation object is present in the field of view. In the first step of preparing the image data to be used, the observation object in the visual field is scanned along a predetermined line, and the granules intersecting the predetermined line are counted by a predetermined reference to obtain the number n of appearances of the granular matter. It is characterized by having a second step of obtaining and a third step of obtaining a standardized number of appearances n / L by dividing the number of appearances n by the length L of the object to be observed scanned by the predetermined line. There is.

In the first method, following the third step, the fourth step of obtaining the normalization constant N and the number of occurrences of normalization n / L are divided by the normalization constant N, and the parameter value n / (L ×). It may further have a fifth step of obtaining N), and as the standardization constant N, the total number of granules in the observation object in the visual field can be used.

The second method according to the present invention that has solved the above problems is a method of detecting uneven distribution of granules in an observation object, and at least a part of the outline of the observation object is present in the field of view. Appearance of granules by the first step of preparing the image data to be performed, scanning the observation object in the visual field along a plurality of predetermined lines, and counting the granules intersecting each predetermined line with a predetermined reference. The second step of obtaining the numbers n ₁ , n ₂ , ..., n _m (where m is an integer of 2 or more. The same applies hereinafter), and the number of occurrences n ₁ , n ₂ , ..., n _m. a predetermined line is scanning the observation object length L _1, L _2, · · ·, number of occurrences normalized by dividing by _{L m n 1 / L 1,} n 2 / L 2, ·· -It is characterized in that it has a third step of obtaining _nm / L _m.

In the second method, following the third step, the fourth step of obtaining the normalization constant N and the number of occurrences of the normalization n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _A fifth obtained by dividing m by the normalization constant N _{to obtain parameter values n 1} / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m × N). steps and may further comprise, as is the normalization constant n, the number of occurrences n _1, n ₂ of granules that intersects the predetermined line, · · ·, n sum of _m .SIGMA.n _i or the standard, The average value of the number of appearances n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _{m can be used.}

In the second method, following the fifth step, the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m × N) It may further have a sixth step of calculating at least one of the standard deviation σ, kurtosis, or skewness of.

In the second method, following the third step, the standard deviation σ, kurtosis, or skewness of the _{normalized appearance numbers n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _m. It may further have a seventh step of calculating at least one degree. Further, following the seventh step, the eighth step of obtaining the normalization constant N and the ninth step of dividing the standard deviation σ calculated in the seventh step by the normalization constant N are further provided. At best, as the normalization constant n, the number of occurrences n _1, n ₂ of granules crossed the predetermined line, ..., the sum of n _m .SIGMA.n _i or the normalized number of occurrences n ₁ / L _1, , N ₂ / L ₂ , ..., The average value of _{n m} / L _{m can be used.}

The predetermined line preferably includes a line segment or a curve. The predetermined line may be a line segment extending radially in the field of view. In the first step, it is preferable to prepare image data in which the entire observation object exists in the field of view. The granular material may be counted by the edge count.

The present invention also includes a program for causing a computer to execute the above method, and a computer-readable recording medium in which the program is stored.

According to the present invention, the number of granules in the observation object is counted and the uneven distribution of the granules is detected by normalizing the number, so that the particles are not affected by the particle size. , The uneven distribution of granules can be detected, and the degree of uneven distribution can be quantified. Further, according to the present invention, it is possible to provide a program for causing a computer to execute such an uneven distribution detection method, and a computer-readable recording medium in which the program is stored.

FIG. 1 is a flowchart of a program that causes a computer to execute the method of the present invention. FIG. 2 is a schematic diagram for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. FIG. 3 is another schematic diagram for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. FIG. 4 is another schematic diagram for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. FIG. 5 is another schematic diagram for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. FIG. 6 is a drawing substitute photograph for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. FIG. 7 is another drawing-substituting photograph for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. FIG. 8A is another drawing-substituting photograph for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. .. B in FIG. 8 is the result of drawing a predetermined line in the x-axis direction with respect to the observation object shown in a in FIG. 8, counting the number of granules, and normalizing along the circular distribution. It is a graph which shows. C in FIG. 8 is the result of drawing a predetermined line in the y-axis direction with respect to the observation object shown in a in FIG. 8, counting the number of granules, and normalizing along the circular distribution. It is a graph which shows. FIG. 9D in FIG. 9 is another drawing-substituting photograph for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules. .. In FIG. 9, e is an angle and a granular object in the circumferential direction when the observation object shown in d in FIG. 9 is scanned with a line segment extending radially from the center of the observation object as a predetermined line. It is a graph which shows the result of counting the number of. FIG. 10 is a drawing-substituting photograph of the standard sample used in the examples. FIG. 11 is a drawing-substituting photograph of the sample A used in the examples. FIG. 12 is a drawing-substituting photograph of the sample B used in the examples. A in FIG. 13 is a drawing-substituting photograph of the sample A used in the examples. B in FIG. 13 is the result of drawing a predetermined line in the x-axis direction with respect to the observation object shown in a in FIG. 13, counting the number of granules, and normalizing along the circular distribution. It is a graph which shows. C in FIG. 13 is the result of drawing a predetermined line in the y-axis direction with respect to the observation object shown in a in FIG. 13, counting the number of granules, and normalizing along the circular distribution. It is a graph which shows. Reference numeral (e) in FIG. 13 indicates an angle and granular matter in the circumferential direction when the observation object shown in FIG. 13 is scanned with a line segment extending radially from the center of the observation object as a predetermined line. It is a graph which shows the result of counting the number of. A in FIG. 14 is a drawing-substituting photograph of the sample B used in the examples. B in FIG. 14 is the result of drawing a predetermined line in the x-axis direction with respect to the observation object shown in a in FIG. 14, counting the number of granules, and normalizing along the circular distribution. It is a graph which shows. C in FIG. 14 is the result of drawing a predetermined line in the y-axis direction with respect to the observation object shown in a in FIG. 14, counting the number of granules, and normalizing along the circular distribution. It is a graph which shows. E in FIG. 14 is an angle and granular matter in the circumferential direction when a line segment extending radially from the center of the observation object is scanned as a predetermined line with respect to the observation object shown in a in FIG. It is a graph which shows the result of counting the number of. FIG. 15 is a diagram showing model data used in the examples. FIG. 16 is a graph showing the result of calculating the standard deviation σ of the parameter value using the total number of granules existing in the image as the normalization constant N. Figure 17 is a normalization constant N, is a graph showing a result of calculating the standard deviation σ of the parameter value using the number of occurrences of the sum .SIGMA.n _i of granules crossed the 1777 present or 888 of a predetermined line. FIG. 18 is a graph showing the result of calculating the standard deviation σ of the parameter value using _{the average value [Σ (ni} / _Li )] / m of the number of occurrences of normalization as the normalization constant N.

The present inventors have made extensive studies in order to provide a method capable of detecting uneven distribution of granules without being affected by the particle size of the granules in the observation object. As a result, they have found that uneven distribution of granules can be detected by counting the number of granules intersecting a predetermined line and standardizing the number based on a predetermined procedure, and completed the present invention. Hereinafter, the method of the present invention will be described step by step with reference to FIGS. 1 to 5. FIG. 1 is a flowchart of a program that causes a computer to execute the method of the present invention. 2 to 5 are schematic views for explaining a procedure for detecting granules in an observation object in a visual field based on the method of the present invention and measuring the number of appearances of the granules.

In the first method and the second method of the present invention, the first to third steps are performed in order. In the first method, one predetermined line is scanned for the observation object, whereas in the second method, a plurality of predetermined lines are scanned for the observation object. Is the same.

[First step]
The first step is common to the first method and the second method.

In the first step, image data is prepared in which at least a part of the outline of the observation object exists in the field of view. The first step will be described with reference to FIG. In FIG. 2, 1 indicates image data, and the image data corresponds to a field of view. In FIG. 2, 2 shows the outline of the observation object, and in the schematic diagram shown in FIG. 2, the entire observation object is present in the visual field. In FIGS. 2 and 3, 3 shows the granules in the observation object.

By allowing at least a part of the outline 2 of the observation object to be present in the visual field, the place where the granular matter 3 is unevenly distributed can be grasped as a relative position to the observation object, and the observation object can be further defined. It can be observed as a whole.

The shape of the observation object is not particularly limited, and examples thereof include an ellipse, a triangle, a rectangle (square, rectangle), and a rhombus in addition to the circle shown in FIG. Further, each vertex of the observation object may be rounded and may be composed of a curved line.

In the image data 1, it is preferable that the entire observation object is present in the field of view.

The image data 1 can be prepared by imaging with, for example, an optical microscope including a Raman microscope, a fluorescence microscope, a phase contrast microscope, or a scanning electron microscope.

As a specific embodiment of the preparation of the image data 1, for example, the data is used as a CD-R, a CD-RW, a DVD-RAM, a DVD-R, a DVD-R DL, a DVD-RW, a DVD + R, a DVD + R DL, a DVD + RW, and the like. Storage in a storage medium such as BD-R, BD-RE, BD-LTH, etc. can be mentioned. In addition, for example, an auxiliary storage device such as a hard disk drive (HDD) or a solid state drive (SDD) in a computer, an auxiliary storage device of a computer connected via a network, and the like can be mentioned.

[Second step]
In the second step of the first method, one predetermined line is scanned against the observation object in the visual field, and the granules intersecting the predetermined line are counted by a predetermined reference, whereby the number of appearances of the granular substances n is increased. can get.

On the other hand, in the second step of the second method, a plurality of predetermined lines are scanned against the observation object in the visual field, and the granules intersecting each predetermined line are counted by a predetermined reference to cause the appearance of the granular substances. The numbers n ₁ , n ₂ , ..., N _m (where m is an integer of 2 or more. The same applies hereinafter) can be obtained.

In the present specification, scanning means performing an operation of counting particles intersecting a predetermined line, and this operation does not have to be performed in order from the end of the predetermined line.

The second step will be described with reference to FIGS. 3 to 5. In FIGS. 3 to 5, the same parts as those in FIG. 2 are designated by the same reference numerals.

FIG. 3 is a schematic view showing a state in which a predetermined line Lx is drawn in the x-axis direction with respect to the outline 2 of the observation object in the image data 1. FIG. 3 shows the x-axis, the y-axis, and the radius R of the outline 2 of the observation object for convenience of explanation.

FIG. 4 is a schematic view showing a state in which a predetermined circular line Lc, which is a concentric circle, is drawn with respect to the outline 2 of the observation object in the image data 1. For convenience of explanation, FIG. 4 shows an angle and a radius r of the circular predetermined line Lc instead of the coordinate axes in order to represent the position on the circular predetermined line Lc.

FIG. 5 is a schematic view showing a state in which a line segment extending radially is drawn as a predetermined line Lz with respect to the outline 2 of the observation object in the image data 1. For convenience of explanation, FIG. 5 shows an arrow indicating the rotation direction when the predetermined line Lz is drawn at the 0 o'clock position of the clock and the predetermined line Lz is rotated counterclockwise and scanned. Further, instead of the coordinate axes, an angle (rotation angle of a predetermined line Lz) is shown.

First, in measuring the number of appearances of the granules, the image obtained by imaging in the first step is, for example, binarized with a predetermined brightness as a threshold value to match the base material of the observation object and the degree of uneven distribution. It is preferable to clearly distinguish the granules 3 to be detected. The image processing software used for binarization is not particularly limited, but general-purpose software such as "ImageJ" and Microsoft Excel (registered trademark) can also be used.

There is no particular limitation on the predetermined standard for counting the number of appearances of the granular material 3a intersecting the predetermined line. For example, the number of appearances of all the granules 3a that intersect the predetermined line by even one pixel may be counted, or the appearance of the granular matter 3a whose particle size (intersection length with the predetermined line Lx) satisfies the predetermined range. You may count the number.

The count of the granules may be performed by the edge count. The edge count is a method of counting the portion where the granular matter intersects a predetermined line and the black changes to white (or the portion where white changes to black) as 1 when scanning on the observation object. .. In the edge count, 1 count is calculated when black changes to white and 1 count is calculated when white changes to black for one granular material, so the total count number is 2. Therefore, when calculating the number of granules by the edge count, the count number obtained by the edge count may be divided by 2.

The position, shape, and number of predetermined lines for scanning the observation object can be arbitrarily selected. For example, a predetermined line having a shape corresponding to the shape of the observation object may be drawn. For example, when the uneven distribution tendency of the granular material can be grasped in advance, the uneven distribution can be detected more efficiently by setting a predetermined line suitable for grasping the uneven distribution.

The shape of the predetermined line for scanning the observation object is not particularly limited, and may include a line segment or a curve.

When the predetermined line is a line segment, as shown in FIG. 3, the predetermined line may be scanned in the x-axis direction with respect to the observation object in the visual field, or the predetermined line may be scanned in the y-axis direction, although not shown. It may be scanned. Further, the scanning may be performed in an oblique direction with respect to the field of view (for example, in the diagonal direction of the field of view).

The shape of the predetermined line may be a polygonal line in which line segments are combined, or may be a closed system in which line segments are combined. Closed systems include, for example, shapes such as triangles, rectangles (squares, rectangles), rhombuses, polygons, and the like.

When the predetermined line is a curve, it may be a closed system such as a circle or an ellipse, or an arc-shaped curve along the outline of the object to be observed. For example, when the outer shape of the observation object is a circle, the predetermined line may be circular and scanned so as to be concentric with the observation object as shown in FIG. 4, or the center may be shifted although not shown. You may scan so that they do not become concentric circles. Similarly, when the outer line of the observation object is circular, oval, rugby ball, oblong, or the like, a predetermined line having a shape similar to each outer line may be used.

The predetermined line segment or curved line may be one line or a plurality of two or more lines with respect to the observation object in the visual field. For example, one of the x-axis direction, the y-axis direction, and the circular shape may be scanned to count the number of granules 3a intersecting the predetermined line, or the x-axis direction or the shape may be counted. A plurality of predetermined lines parallel to the y-axis direction may be scanned to count the number of granules 3a intersecting the predetermined lines. Further, the number of granules 3a intersecting the predetermined lines may be counted by scanning a plurality of circular predetermined lines having different diameters. When a plurality of predetermined lines having a similar shape to the outer line are used, predetermined lines having different similarity ratios can be used. Alternatively, two or more types are selected from the group consisting of a line segment in the x-axis direction, a line segment in the y-axis direction, and a circular shape, the number of granules 3a intersecting each predetermined line is counted, and the granules are unevenly distributed. May be detected and evaluated. By making a plurality of predetermined lines or combining predetermined lines in different directions or different shapes, it is possible to accurately detect and evaluate the uneven distribution of granules in the observation object.

When scanning two or more of the predetermined lines, as shown in FIG. 5, the line segments extending radially in the visual field may be scanned as the predetermined line Lz. In FIG. 5, 18 predetermined lines Lz are drawn at intervals of 20 °, but the number and intervals of the predetermined lines Lz are not limited to this. Further, the predetermined line Lz may be scanned radially at once and count the particles intersecting each predetermined line with a predetermined reference, or as shown in FIG. 5, one predetermined line Lz may be counterclockwise. Granules that are rotated around and intersect the predetermined line Lz at predetermined angles or continuously may be counted by a predetermined reference.

The measured number of appearances n or the number of appearances n ₁ , n ₂ , ..., N _m (where m is an integer of 2 or more; the same shall apply hereinafter) may be stored in the recording medium or recording device. can.

[Third step]
In the third step of the first method, the normalized appearance number n / L is calculated by dividing the appearance number n measured in the second step by the length L obtained by scanning the observation object with the predetermined line. do.

On the other hand, in the third step in the second method, the number of appearances n ₁ , n ₂ , ..., N _m measured in the second step is measured by the predetermined line and the length L _{1 obtained by scanning the observation object.} , L ₂ , ···, L _m to calculate the normalized number of occurrences n ₁ / L ₁ , n ₂ / L ₂ , ···, n _m / L _m .

The calculated number of occurrences of the standardization n / L or the number of occurrences of the standardization n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m is the granularity existing per unit length. The number of objects 3a is shown.

Granules in the observation object based on the normalized appearance number n / L or the normalized appearance number n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _{m calculated in this way.} Can detect uneven distribution.

That is, if the granules are completely and uniformly dispersed in the observation object, one predetermined line is scanned for each sample, and the standardization calculated for the predetermined line drawn for each sample is performed. The number of appearances n / L becomes constant (same value). If the granules are completely and uniformly dispersed in the object to be observed, the number of normalized appearances calculated for each predetermined line by scanning a plurality of predetermined lines n ₁ / L ₁ , n ₂ / L ₂ , ..., n _m / L _m becomes constant (same value). Further, the standardized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m calculated by the second method are set to the positions where the observation object is scanned with a predetermined line, for example. On the other hand, by plotting, it is possible to visually detect and evaluate the uneven distribution of granules in the observation object based on the shape of the graph. Further, the uneven distribution of granules in the object may be evaluated by obtaining the difference between the maximum value and the minimum value of the graph or by obtaining the component value of a specific spatial frequency by Fourier transform.

The length of scanning the observation object can be calculated, for example, by the following procedure.

As shown in FIG. 3, the length L when a predetermined line Lx is drawn in the x-axis direction with respect to the outline 2 of the observation object in the image data 1 is an arbitrary point on the outer diameter line 2. Is expressed by (x1, y1) and the radius of the outer diameter line 2 is expressed by R, it can be calculated by ^{2 × √ (R 2-} y1 ^2). Although not shown, the length L when a predetermined line Lx is drawn in the y-axis direction with respect to the outline 2 of the observation object in the image data 1 is 2 × √ (R ² −x1 ² ). Can be calculated with.

As shown in FIG. 4, the length L when a circular predetermined line Lc which is a concentric circle is drawn with respect to the outline 2 of the observation object in the image data 1 is the radius of the predetermined line Lc as r. Expressed, it can be calculated by 2πr.

_{As shown in FIG. 5, the lengths L 1} , L ₂ , ..., L _m when the line segments extending radially with respect to the observation object in the image data 1 are drawn as predetermined lines Lz are When the radius of the outer diameter line 2 is represented by R, it becomes the radius R itself.

In the first method, the standardized appearance number n / L may be calculated by inputting a numerical value of the length L into a personal computer and reading out the appearance number n stored in advance.

On the other hand, in the second method, the standardized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m are the lengths L ₁ , L ₂ , ... ..., It may be calculated by _{inputting a numerical value of L m} and reading out the above-mentioned appearance numbers n ₁ , n ₂ , ..., N _{m stored in advance.}

In the first method and the second method, following the third step, the following fourth and fifth steps may be performed in this order. Further, in the second method, the following seventh step may be performed following the third step.

[4th step]
The fourth step is common to the first method and the second method, and is a step of obtaining a standardization constant N.

As the standardization constant N, in the first method, the total number of granules 3 in the observation object in the visual field may be obtained.

On the other hand, in the second method, as normalization constant N, the number of occurrences n _1, n _2, · · ·, or obtaining the sum .SIGMA.n _i of n _m, the normalized number of occurrences n ₁ / L _1, n ₂ The average value of / L ₂ , ..., N _m / L _{m may be obtained.} The average value of the above-mentioned normalized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m can be obtained by [Σ ( _ni / _Li )] / m. Σ is the sum from i = 1 to i = m (the same applies hereinafter). Further, in the second method, the total number of granules 3 in the observation object in the visual field may be obtained as the normalization constant N. Further, in the second method, as normalization constant N, the number of occurrences n _1, n _2, · · ·, the average value of the sum .SIGMA.n _i of n _m may also be determined (Σn _i) / m, the The sum of Σni of the number of appearances n ₁ , n ₂ , ..., N _{m is the sum of Σ l i} of _{the lengths L 1} , L ₂ , ..., L _{m in which} the predetermined line scans the observation object. The divided value Σn _i / ΣL _i may be obtained.

In both the first method and the second method, if standardization is not required, the normalization constant N = 1 may be set.

The field of view means the field of view observed in the first step.

When the total number of granules 3 in the observation object in the visual field is obtained as the normalization constant N, the number of appearances n or the number of appearances n ₁ , n ₂ , ... _nm was a method of scanning a predetermined line in one dimension and counting, but since the total number of granules 3 in the fourth step is the number count in the two-dimensional space, it is usually counted in the second step. The count is performed by a method different from the method. For example, in the fourth step, the particles having a diameter equivalent to a circle of 3a of a predetermined value or more can be counted as one. Of course, the numerical standard for counting the size of the granular material 3a as one (for example, the standard of 1 μm or more) is the same as the count standard of the second step and the count standard of the fourth step. You may.

The measured normalization constant N can be stored in the recording medium or recording device.

[Fifth step]
The fifth step is also common to the first method and the second method, and is a step of dividing the number of occurrences of normalization obtained in the third step by the normalization constant N obtained in the fourth step.

That is, in the fifth step in the first method, the normalized appearance number n / L calculated in the third step is divided by the normalization constant N calculated in the fourth step, and the parameter value n / (L). × N) is calculated.

On the other hand, in the fifth step in the second method, the normalized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m calculated in the third step are converted into the fourth step. Divide by the normalization constant N calculated in the above to calculate the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ···, _nm / (L _m × N). ..

The parameter value n / (L × N) or the parameter value n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m × N) calculated in this way. ), The uneven distribution of granules in the observation object can be detected.

That is, if the granules are completely and uniformly dispersed in the observation object, the parameter value n / (L × N) or the parameter value n ₁ / (L ₁ × N), n ₂ / (L _2). × N), ···, _nm / (L _m × N) becomes constant (same value). Further, the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m × N) calculated by the second method are specified, for example. By plotting the observation object with a line against the scanned position, it is possible to visually detect and evaluate the uneven distribution of the particles in the observation object based on the shape of the graph. Further, the uneven distribution of granules in the object may be evaluated by obtaining the difference between the maximum value and the minimum value of the graph or by obtaining the component value of a specific spatial frequency by Fourier transform.

In the fifth step of the second method, the average value of the _{normalized appearance numbers n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _{m calculated in the third step above [Σ (} The _{parameter value [Σ (ni} / _Li )] / (m × N) is calculated by _{dividing n i} / _Li )] / m by the normalization constant N calculated in the fourth step above, and this is calculated. The uneven distribution of granules in the observation object may be detected based on the parameter value [Σ ( _ni / _{Li)] / (m × N).} Alternatively, the parameter value n _ave / (L _ave × N) is obtained from the average value n _{ave of the} number of occurrences and the average value L _{ave of the scanning length, and based on this parameter value n ave} / (L _ave × N), The uneven distribution of granules in the observation object may be detected.

In the first method, the parameter value n / (L × N) may be obtained by calculating the normalized appearance number n / L stored in advance in the personal computer and the normalized constant N.

Further, in the second method, the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m × N) are applied to the personal computer. The number of normalization occurrences n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m stored in advance and the normalization constant N may be calculated and obtained.

[6th step]
In the second method, the following sixth step may be performed after the first to fifth steps are performed.

_{In the sixth step, the parameter values n 1} / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m ) calculated in the fifth step in the second method. × N) Calculate at least one of the standard deviation σ, kurtosis (sometimes called “pointing”), and skewness (sometimes called “distortion” or “strain”). By calculating the standard deviation σ, kurtosis, or skewness, the uniformity (variation) of the appearance state of the granular material can be evaluated.

The calculated standard deviation σ, kurtosis, or skewness can be stored in the recording medium or recording device.

[7th step]
In the second method, the following seventh step may be performed following the third step.

In the seventh step, at least one of the standard deviation σ, kurtosis, or skewness of the _{normalized appearance numbers n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _{m is calculated.} By calculating the standard deviation σ, kurtosis, or skewness, the uniformity (variation) of the appearance state of the granular material can be evaluated.

[8th step]
In the second method, after performing the seventh step, the following eighth and ninth steps may be performed in this order.

The eighth step is the same as the fourth step, and is a step of obtaining the normalization constant N. That is, the normalization constant N, the number of occurrences n _1, n _2, · · ·, or obtaining the sum .SIGMA.n _i of n _m, the normalized number of occurrences _{n 1 / L 1, n 2} / L 2, ··・, The average value of _nm / L _{m may be obtained.} Further, in the second method, the total number of granules 3 in the observation object in the visual field may be obtained as the normalization constant N. Further, in the second method, as normalization constant N, the number of occurrences n _1, n _2, · · ·, the average value of the sum .SIGMA.n _i of n _m may also be determined (Σn _i) / m, the The sum of Σni of the number of appearances n ₁ , n ₂ , ..., N _{m is the sum of Σ l i} of _{the lengths L 1} , L ₂ , ..., L _{m in which} the predetermined line scans the observation object. The divided value Σn _i / ΣL _i may be obtained.

[9th step]
In the 9th step, the standard deviation σ of the above-mentioned normalized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m calculated in the 7th step is obtained in the 8th step. This is a step of dividing by the normalization constant N. The smaller the value obtained by dividing the standard deviation σ by the normalization constant N, the more it can be evaluated that the granules are uniformly dispersed in the observation object.

The calculation in the ninth step may be performed using the standard deviation σ stored in advance in the personal computer and the normalization constant N.

Examples of the observation target include tablets used in the fields of pharmaceuticals and foods, and examples of the granules include active ingredients contained in tablets.

As the tablet, the uncoated tablet may be used as it is as an observation target, or the uncoated tablet may be cut to smooth the observation surface and used as the observation object.

Next, a procedure for detecting granules in the observation object in the visual field based on the method of the present invention and measuring the number of appearances of the granules is specifically described with reference to the drawings (FIGS. 6 to 9). explain. The procedure is not limited to the drawings.

A tablet with a circular dosage form (uncoated tablet) was prepared, cut and the surface was smoothed, and the appearance was photographed with a Raman microscope at the same magnification. A drawing substitute photograph is shown in FIG. In FIG. 6, the entire tablet was photographed so as to fit within the field of view (photograph).

FIG. 7 shows a drawing substitute photograph obtained by binarizing the drawing substitute photograph shown in FIG. 6 using the image analysis software “ImageJ”. The binarization was performed so that the granules to be detected for the presence or absence of uneven distribution and other components were not covered. The white dots shown in FIG. 7 indicate granules.

A drawing substitute photograph in which a predetermined line is drawn in each of the x-axis direction and the y-axis direction with respect to the tablet shown in FIG. 7 is shown in FIG. 8a. In FIG. 8, the line segment Lx indicates a predetermined line scanned in the x-axis direction, and the line segment Ly indicates a predetermined line scanned in the y-axis direction.

The result of counting the number of granules (number of appearances) intersecting with the line segment Lx and dividing by the length of the line segment Lx is shown in FIG. 8b. Further, the result of counting the number of granules (number of appearances) intersecting with the line segment Ly and dividing by the length of the line segment Ly is shown in FIG. 8c. The horizontal axes of b and c in FIG. 8 indicate the scanning positions corresponding to the observation objects, and the vertical axis indicates the value obtained by dividing the counted number (appearance number) by the length of the predetermined line.

Further, a drawing substitute photograph in which a predetermined line Lz corresponding to the radius of the observation object is drawn with respect to the tablet shown in FIG. 7 is shown in d of FIG. The predetermined line Lz was rotated 360 ° clockwise in the direction indicated by the arrow, and when the position of the first predetermined line Lz was 0 °, the number of granules intersecting the predetermined line Lz at each angle was counted. The result obtained by dividing the number of granules counted at each angle by the length of the predetermined line Lz (that is, corresponding to the radius R of the observation object) is shown in FIG. 9e. The horizontal axis of FIG. 9e shows the rotation angle of the predetermined line, and the vertical axis shows the value obtained by dividing the number of particles (the number of appearances) at each angle by the length of the predetermined line Lz.

It is considered that the value of the number of granules / the length of the predetermined line counted when the granules are uniformly dispersed in the observation object shows a constant value depending on the scanning position. On the other hand, when the granules are unevenly distributed, the value of the number of granules to be counted / the length of the predetermined line differs depending on the scanning position.

The procedure for detecting granules in the observation object in the visual field based on the method of the present invention and measuring the number of appearances of the granules has been described above with reference to the drawings.

The present invention includes a program for causing a computer to execute the above method.

The present invention also includes a computer-readable recording medium in which the above program is stored.

Hereinafter, the present invention will be described in more detail with reference to Examples, but the present invention is not limited by the following Examples, and the present invention shall be carried out with modifications to the extent that it can be adapted to the gist of the above and the following. Of course, they are all within the technical scope of the present invention.

Experiment 1
[Sample preparation]
Commercially available tablets (bare tablets) were prepared as standard samples. The dosage form of the prepared tablet was circular, and the tablet was cut and the surface was smoothed. The appearance was observed with a Raman microscope at the same magnification, and the entire tablet was photographed so as to fit in the field of view (photograph). The radius R of the outline of the tablet is 5 mm. The photograph taken was image-analyzed using the image analysis software "ImageJ" and binarized. The binarization was performed so that the granules to be detected for uneven distribution and other components were not covered. A binarized drawing substitute photograph is shown in FIG. The white dots shown in FIG. 10 indicate granules. As is clear from FIG. 10, it can be seen that in the standard sample, the granules are dispersed almost uniformly.

Since the granules contained in the standard sample are dispersed almost uniformly, the sample in which the granules are unevenly distributed is simulated, and the drawing substitute photograph obtained by binarizing the photograph taken of the standard sample is image-processed and two samples ( Sample A and Sample B) were prepared. A drawing-substituting photograph of Sample A is shown in FIG. 11, and a drawing-substituting photograph of Sample B is shown in FIG. As shown in FIG. 11, in Sample A, the granules are unevenly distributed in the lower right corner of the tablet, and the granules are hardly unevenly distributed in the upper left corner of the tablet. As shown in FIG. 12, in the sample B, the granules are unevenly distributed on the outer peripheral side of the tablet, and the granules are hardly present in the central portion.

Using the above three tablets, uneven distribution of granules in the tablets was detected and evaluated by the following procedure.

[Comparison example]
As a conventional method, the uneven distribution of granules was detected and evaluated using the Fourier transform.

The drawing-substituting photographs of FIGS. 10 to 12 were Fourier-transformed, and a value obtained by averaging the frequency regions of 0.33 to 0.5 / mm was calculated. The calculation results are shown in Table 1 below. In the Fourier transform, the smaller the value, the more uniformly the particles are dispersed. As is clear from Table 1 below, the average value calculated using the Fourier transform is almost the same for the standard sample in which the granules are uniformly dispersed, and in the sample A and the sample B in which the granules are unevenly distributed. Yes, it can be seen that uneven distribution of granules cannot be detected even based on this average value.

[Invention Example]
Uneven distribution of granules was detected and evaluated based on the method of the present invention.

Using the drawing-substituting photographs of FIGS. 10 to 12, a predetermined line Lx in the x-axis direction, a predetermined line Ly in the y-axis direction, and a predetermined line Lz so as to be able to scan radially are drawn, and granules intersecting each predetermined line are drawn. Edge counting was performed, and the number of appearances of granules n ₁ , n ₂ , ..., N _m (where m is an integer of 2 or more. The same applies hereinafter) was measured.

The length of the predetermined line Lx drawn in the x-axis direction is determined by ^{2 × √ (25-x1 2} ) mm, the length of the predetermined line Ly drawn in the y-axis direction is ^{2 × √ (25-y1 2} ) mm .. In addition, x1 and y1 mean the coordinates (x1, y1) of an arbitrary point on the circumference of a tablet.

Next, for each predetermined line, the number of appearances n ₁ , n ₂ , ..., N _m is set, and the length L ₁ , L ₂ , ..., L _m (that is, Lx, _{The number of normalized occurrences n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _m was calculated by dividing by (length of Ly).

Graphs with the scanning position corresponding to the observation object on the horizontal axis and the number of normalized occurrences calculated at each scanning position on the vertical axis are shown in b and c in FIG. 8, b and c in FIG. 13, and b and c in FIG. Each is shown. 8 shows the result of using the standard sample shown in FIG. 10, FIG. 13 shows the result of using the sample A shown in FIG. 11, and FIG. 14 shows the result of using the sample B shown in FIG. In FIGS. 8, 13, and 14, b shows the result of scanning the predetermined line Lx in the x-axis direction, and c shows the result of scanning the predetermined line Ly in the y-axis direction.

On the other hand, the predetermined line Lz is drawn so as to overlap the radius of the tablet, and the length of the predetermined line Lz is 5 mm. The predetermined line Lz was rotated 360 ° with the center of the tablet as an end point and scanned at each rotation angle. That is, when the initial position of the predetermined line Lz is 0 °, the number of granules intersecting the predetermined line Lz at each angle when the predetermined line Lz is rotated is counted.

Next, for each predetermined line, the number of appearances n ₁ , n ₂ , ..., N _{m is set, and the lengths L 1} , L ₂ , ..., L _m (that is, Lz) obtained by scanning the tablet by the predetermined line. The number of normalized occurrences n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m was calculated by dividing by (length).

Graphs with the rotation angle of the predetermined line on the horizontal axis and the number of normalized occurrences calculated at each angle on the vertical axis are shown in E of FIG. 9, e of FIG. 13, and e of FIG. 14, respectively.

As is clear from b, c in FIG. 8, e in FIG. 9, b, c, e in FIG. 13, b, c, e in FIG. 14, the number of normalized occurrences n ₁ / L ₁ , n ₂ / L ₂ , ..., The uneven distribution of granules in the observation object can be detected by the shape of the graph based on _nm / L _m.

Next, the standard deviation σ of the normalized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _{m calculated above was obtained.} The calculated standard deviation σ is shown in Table 1 below.

Next, as the standardization constant N, the total number of granules in the tablet in the visual field was measured under the same conditions as when the number of granules intersecting the predetermined line was counted. The obtained standardized constant N is shown in Table 1 below.

Next, the standard deviation σ was divided by the normalization constant N to calculate σ / N. The calculated σ / N is shown in Table 1 below.

As is clear from Table 1 below, the standard sample tablets have a small σ / N and almost no uneven distribution of granules regardless of whether a predetermined line is drawn in the x-axis direction, the y-axis direction, or the radial direction. Can be evaluated.

It can be evaluated that the tablet of sample A has a large σ / N and the granules are unevenly distributed regardless of whether a predetermined line is drawn in the x-axis direction, the y-axis direction, or the radial direction.

In the tablet of sample B, since σ / N is large when a predetermined line is drawn in the x-axis direction and the y-axis direction, the granules are unevenly distributed in the x-axis direction and the y-axis direction, but the predetermined lines are radially distributed. Since the σ / N when subtracted is small, it can be evaluated that the granules are not unevenly distributed in the radial direction and are uniformly dispersed.

From the above results, according to the present invention, by changing the drawing direction of the predetermined line and the shape of the predetermined line according to the uneven distribution state of the granular material, the influence of the size of the granular material in the observation object is affected. Uneven distribution of granules can be detected and evaluated without receiving.

Experiment 2
In Experiment 2, the difference in the degree of uneven distribution of granules detected depending on the type of normalization constant N was examined.

Images A to C shown in FIG. 15 were prepared as model data. A is a Gaussian distribution in which the center of the distribution is at the center of an image of 400 × 400 pixels, and is an image in which particles having a radius of 2 pixels are randomly arranged. B is an image in which the center of the distribution of A is shifted by 50 pixels, and C is an image in which the center of the distribution of A is shifted by 100 pixels.

Using the image shown in FIG. 15, uneven distribution of granules was detected and evaluated based on the method of the present invention. Specifically, as a predetermined line, a plurality of line segments extending radially from the center of the image are drawn, and the number of granules intersecting each predetermined line is counted, so that the number of appearances of the granules n ₁ , n ₂ , ...・, N _m was calculated. The number of predetermined lines was 1777 or 888. The number of granules was counted by the edge count.

As normalization constant N, the total number of particulates present in the image, 1777 present or 888 number of occurrences of the sum .SIGMA.n _i of granules crosses a predetermined line, and the normalized number of occurrences of the average value [Σ (n _i / L _i )] / m is obtained, and the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ···, _nm / (L) are obtained using the respective normalization constants N. _m × N) was calculated.

Next, the standard deviation σ of the obtained parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ..., N _m / (L _m × N) was calculated. The calculation results are shown in Table 2 below.

Further, FIG. 16 shows the result of calculating the standard deviation σ of the parameter value using the total number of granules existing in the image as the normalization constant N. As is clear from FIG. 16, when the standard deviation σ of the parameter value was calculated using the total number of granules existing in the image as the normalization constant N, the standard deviation σ did not show a monotonous increase. That is, the standard deviation σ is considered to be A <B <C according to the bias of the center of the distribution, but when the total number of granules existing in the image is used as the normalization constant N, such a tendency is not shown. rice field. It is considered that this is because the number of appearances of the granules intersecting the predetermined line changes with the uneven distribution of the granules. That is, it is considered that the influence of the change in the number of appearances of the granules intersecting the predetermined line cancels out the change in the standard deviation σ due to the uneven distribution.

Next, as normalization constant N, the result of calculating the standard deviation σ of the parameter value using the number of occurrences of the sum .SIGMA.n _i of granules crossed the 1777 present or 888 of a predetermined line in FIG. In FIG. 17, the dotted line (a) shows the result when the number of predetermined lines is calculated as 1777, and the solid line (b) shows the result when the number of predetermined lines is calculated as 888. As is clear from FIG. 17, when calculating the standard deviation σ of the parameter value using the number of occurrences of the sum .SIGMA.n _i of granules crossed the 1777 present or 888 of predetermined lines as a normalization constant N, the standard deviation σ showed a monotonous increase, and uneven distribution of granules could be detected. However, the calculated standard deviation σ is approximately doubled when the number of predetermined lines is 1877 and when the number of predetermined lines is 888, and may depend on the number of predetermined lines. I understand.

Next, FIG. 18 shows the result of calculating the standard deviation σ of the parameter value using _{the average value [Σ (ni} / _Li )] / m of the number of occurrences of normalization as the normalization constant N. In FIG. 18, the dotted line (a) shows the result when the number of predetermined lines is calculated as 1777, and the solid line (b) shows the result when the number of predetermined lines is calculated as 888. As is clear from FIG. 18, when the standard deviation of the parameter value is calculated using _{the average value [Σ (ni} / _Li )] / m of the number of appearances of standardization as the standardization constant N, it is as shown in FIG. Similarly, the standard deviation σ showed a monotonous increase, and uneven distribution of granules could be detected. Further, the result of the calculated standard deviation σ is almost the same regardless of whether the number of predetermined lines is 1777 or 888, and it can be seen that it does not depend on the number of predetermined lines.

If the count number of the granular material changes due to the uneven distribution of the granular material according to the result of Experiment 2, the total number of appearances of the granular material intersecting the predetermined line _Σ ni is used as the normalization constant N, or the standardized appearance number is used. It can be seen that it is better to use the average value [Σ ( _ni / _{Li)] / m.} Further, it can be seen that in order to obtain a result independent of the number of predetermined lines, it is better to use the _{average value [Σ (ni} / _{Li)] / m of the number of normalized occurrences as the normalization constant N.}

1 Image data 2 Outline line 3 Granules 3a Granules that intersect the predetermined line Lx Predetermined line drawn in the x-axis direction Ly y Predetermined line drawn in the axis direction Lc Predetermined line drawn in a circle Lz Predetermined line drawn in a radial direction L Predetermined line length

Claims (15)

It is a method of preparing multiple samples and detecting the uneven distribution of granules in the observation object of each sample.
The first step of preparing image data in which at least a part of the outline of the observation object exists in the field of view, and
The second step of scanning the observation object in the visual field along a predetermined line and counting the granules intersecting the predetermined line with a predetermined reference to obtain the number n of appearances of the granules.
The third step of obtaining the normalized number of appearances n / L by dividing the number of appearances n by the length L obtained by scanning the object to be observed by the predetermined line.
Have,
When the normalized appearance number n / L calculated for each sample is constant, it is detected that the granules in the observation object are uniformly dispersed, and when it is not constant, it is in the observation object. wherein the benzalkonium be detected and granules are unevenly distributed. It is a method of preparing multiple samples and detecting the uneven distribution of granules in the observation object of each sample.
The first step of preparing image data in which at least a part of the outline of the observation object exists in the field of view, and
The second step of scanning the observation object in the visual field along a predetermined line and counting the granules intersecting the predetermined line with a predetermined reference to obtain the number n of appearances of the granules.
The third step of obtaining the normalized number of appearances n / L by dividing the number of appearances n by the length L obtained by scanning the object to be observed by the predetermined line.
Have,
Following the third step,
The fourth step of finding the normalization constant N and
The fifth step of dividing the normalized appearance number n / L by the normalized constant N to obtain a parameter value n / (L × N), and
With more
The normalization constant N is the total number of granules in the observation object in the field of view.
When the parameter value n / (L × N) calculated for each sample is constant, it is detected that the granules in the observation object are uniformly dispersed, and when it is not constant, it is in the observation object. A method characterized by detecting that the granules in the above are unevenly distributed. It is a method of detecting the uneven distribution of granules in the observation object.
The first step of preparing image data in which at least a part of the outline of the observation object exists in the field of view, and
The number of appearances of granules n ₁ , n ₂ , ... The second step to obtain n _m (however, m is an integer of 2 or more. The same applies hereinafter), and
Normalized appearance by dividing the number of appearances n ₁ , n ₂ , ..., N _{m by the lengths L 1} , L ₂ , ..., L _m of the predetermined line scanning the observation object. The third step of obtaining the numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., n _m / L _{m, and}
Have,
Following the third step,
The fourth step of finding the normalization constant N and
The number of occurrences of the standardization n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m is divided by the normalization constant N, and the parameter values n ₁ / (L ₁ × N), n ₂ The fifth step of obtaining / (L ₂ × N), ···, _nm / (L _{m × N), and}
With more
Following the fifth step,
At least one of the standard deviation σ, kurtosis, or skewness of the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ···, _nm / (L _{m × N).} Further has a sixth step of calculating one
The predetermined line is a circular line segment in the x-axis direction with respect to the observation object in the visual field, a line segment in the y-axis direction with respect to the observation object in the visual field, and a circular circle concentric with the outer line of the observation object. Curve, line segment extending radially in the field of view, line segment diagonal to the field of view, arcuate curve along the outline, or line having a shape similar to the outline, or a combination of line segments It is either a folded line, a closed system line that combines line segments, or a closed system curve.
Wherein said normalization constant N is the number of occurrences n _1, n ₂ of granules crossed the predetermined line, ..., the sum .SIGMA.n _i of n _m. It is a method of detecting the uneven distribution of granules in the observation object.
The first step of preparing image data in which at least a part of the outline of the observation object exists in the field of view, and
The number of appearances of granules n ₁ , n ₂ , ... The second step to obtain n _m (however, m is an integer of 2 or more. The same applies hereinafter), and
Normalized appearance by dividing the number of appearances n ₁ , n ₂ , ..., N _{m by the lengths L 1} , L ₂ , ..., L _m of the predetermined line scanning the observation object. The third step of obtaining the numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., n _m / L _{m, and}
Have,
Following the third step,
The fourth step of finding the normalization constant N and
The number of occurrences of the standardization n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m is divided by the normalization constant N, and the parameter values n ₁ / (L ₁ × N), n ₂ The fifth step of obtaining / (L ₂ × N), ···, _nm / (L _{m × N), and}
With more
Following the fifth step,
At least one of the standard deviation σ, kurtosis, or skewness of the parameter values n ₁ / (L ₁ × N), n ₂ / (L ₂ × N), ···, _nm / (L _{m × N).} Further has a sixth step of calculating one
The predetermined line is a circular line segment in the x-axis direction with respect to the observation object in the visual field, a line segment in the y-axis direction with respect to the observation object in the visual field, and a circular circle concentric with the outer line of the observation object. Curve, line segment extending radially in the field of view, line segment diagonal to the field of view, arcuate curve along the outline, or line having a shape similar to the outline, or a combination of line segments It is either a folded line, a closed system line that combines line segments, or a closed system curve.
A method characterized in that the normalization constant N is _{an average value of the normalization appearance numbers n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _m. It is a method of detecting the uneven distribution of granules in the observation object.
The first step of preparing image data in which at least a part of the outline of the observation object exists in the field of view, and
The number of appearances of granules n ₁ , n ₂ , ... The second step to obtain n _m (however, m is an integer of 2 or more. The same applies hereinafter), and
Normalized appearance by dividing the number of appearances n ₁ , n ₂ , ..., N _{m by the lengths L 1} , L ₂ , ..., L _m of the predetermined line scanning the observation object. The third step of obtaining the numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., n _m / L _{m, and}
Have,
Following the third step,
Further, there is a seventh step of calculating at least one of the standard deviation σ, kurtosis, or skewness of the normalized appearance numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., N _m / L _m. death,
The predetermined line is a circular line segment in the x-axis direction with respect to the observation object in the visual field, a line segment in the y-axis direction with respect to the observation object in the visual field, and a circular circle concentric with the outer line of the observation object. Curve, line segment extending radially in the field of view, line segment diagonal to the field of view, arcuate curve along the outline, or line having a shape similar to the outline, or a combination of line segments A method characterized by being either a folded line, a closed system line that combines line segments, or a closed system curve. Following the seventh step,
The eighth step of finding the normalization constant N and
In the ninth step, which divides the standard deviation σ calculated in the seventh step by the normalization constant N,
With more
The method of claim 5 wherein the normalization constant N is the number of occurrences n _1, n ₂ of granules crossed the predetermined line, ..., the sum .SIGMA.n _i of n _m. Following the seventh step,
The eighth step of finding the normalization constant N and
In the ninth step, which divides the standard deviation σ calculated in the seventh step by the normalization constant N,
With more
The method according to claim 5, wherein the normalization constant N is _{an average value of the normalization appearance numbers n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _m. It is a method of detecting the uneven distribution of granules in the observation object.
The first step of preparing image data in which at least a part of the outline of the observation object exists in the field of view, and
The number of appearances of granules n ₁ , n ₂ , ... The second step to obtain n _m (however, m is an integer of 2 or more. The same applies hereinafter), and
Normalized appearance by dividing the number of appearances n ₁ , n ₂ , ..., N _{m by the lengths L 1} , L ₂ , ..., L _m of the predetermined line scanning the observation object. The third step of obtaining the numbers n ₁ / L ₁ , n ₂ / L ₂ , ..., n _m / L _{m, and}
Have,
_{Multiple standardized appearance numbers n 1} / L ₁ , n ₂ / L ₂ , ..., N _m / L _m calculated for the above plurality of predetermined lines are set with respect to the position where the observation object is scanned along the predetermined lines. A method characterized in that a plotted graph is created and the shape of the graph is visually evaluated, or the uneven distribution of granules in an observation object is evaluated by analyzing the graph. The method according to any one of claims 3 to 8, wherein the predetermined line is a line segment extending radially in the field of view. The method according to any one of claims 1, 2 and 8, wherein the predetermined line includes a line segment. The method according to any one of claims 1, 2 and 8, wherein the predetermined line includes a curve. The method according to any one of claims 1 to 11, wherein in the first step, image data in which the entire observation object is present in the field of view is prepared. The method according to any one of claims 1 to 12, wherein the counting of the granules is performed by edge counting. A program for causing a computer to execute the method according to any one of claims 1 to 13. A computer-readable recording medium in which the program according to claim 14 is stored.

Images