JP3150778B2 - Threshold determination device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention determines an optimal threshold value for binarizing a grayscale image in a binarization process which is a typical process for separating a target object in an image from a background. The present invention relates to a threshold value determining device.

[0002]

2. Description of the Related Art FIG. 25 is a block diagram showing a conventional threshold value determining apparatus. In FIG.
The histogram processing unit 3 counts the frequency of appearance of each density in. The threshold processing unit 3 determines a threshold value based on the processing result of the histogram processing unit 2, and the histogram processing unit 4 includes the histogram processing unit 2 and the threshold determination processing unit 3. The threshold value determination devices 5 and 5 are binarization processing units that binarize the grayscale image based on the threshold value determined by the threshold value determination device 4 to obtain a binary image.

Next, the operation will be described. For example, when a grayscale image 1 having the number of gradations m (having a value of 0 to m-1) is input to the histogram processing unit 2, for each density value t,
The number of pixels having the density is counted, and the appearance frequency Ht for each density is obtained.

Next, in the threshold value determination processing section 3, for example, P
The threshold value is determined using a means such as a tile method, a mode method, and a discriminant analysis method. First, in the P tile method, a density level at which the cumulative frequency Pt from the higher density of the density histogram is equal to or greater than a preset value P percent is determined as a threshold value.

[0005]

(Equation 1)

In the mode method, on the premise that the shape of the density histogram indicates bilaterality, a valley between two peaks is found, and this density is determined as a threshold. Further, in the discriminant analysis method, when the density histogram is divided into two by a certain density, the density at which the variance between the classes of the density histogram divided into two is the maximum is determined as the threshold value.

Using the threshold value t obtained by the above-described method, the binarization processing unit 5 sets a pixel having a density equal to or greater than the threshold value t to 1 and a pixel not having the density to 0 in the grayscale image 1. To obtain a binary image.

[0008]

Since the conventional threshold value determining apparatus is configured as described above, when the P tile method is used, it is necessary to know in advance the ratio of the target object in the grayscale image. In the case of using the mode method, it is difficult to determine the threshold value when the density histogram is not smooth or the valley portion is not conspicuous, and when the discriminant analysis method is used, the ratio of the target object and the background in the image is extremely large. However, when the difference is different or when the density difference between the target object and the background is small, an appropriate threshold cannot be determined. Also, no matter which method is used, the threshold value obtained for an image in which the target object does not exist is not meaningful because it is not the threshold value obtained after determining that the target object does not exist. Was.

The present invention has been made in order to solve the above-mentioned problems, and can be applied to a case where conditions such as an extremely different ratio of a target object and a background in an image are severe or a case where no target object exists. It is an object of the present invention to obtain a threshold value determining device capable of determining an optimum threshold value.

[0010]

According to a first aspect of the present invention, there is provided a threshold value determining apparatus for each threshold value, wherein the number of pixels counted by the first counting section is compared with the number of pixels counted by the second counting section. Calculate the ratio of the number, if there is a value whose ratio becomes a local maximum due to the change of the threshold, select the threshold when the ratio becomes the local maximum as the optimal threshold, and if there is no value that becomes the local maximum,
The threshold at which the ratio becomes zero is selected as the optimum threshold.

Further, the threshold value determining apparatus according to the second aspect of the present invention further comprises a setting unit for setting a maximum value and a minimum value of the number of pixels of the target object, wherein the number of pixels counted by the first counting unit is: Only when the ratio is larger than the minimum value set by the setting unit and smaller than the maximum value, if the ratio of the number of pixels calculated by the calculation unit has a maximum value due to a change in the threshold, the ratio is the maximum value Is selected as the optimum threshold, and if there is no value that has the maximum value, the threshold at which the ratio becomes zero is selected as the optimum threshold.

Further, the threshold value determining apparatus according to the third aspect of the present invention, for each threshold value, counts the number of pixels having a density higher than the threshold value for each threshold value based on the processing result of the first histogram processing unit. A count processing unit, and a minimum value filter processing unit that sequentially sets each pixel in the grayscale image as a target pixel, and changes a density value of a pixel having the lowest density among the target pixel and its neighboring pixels to a density value of the target pixel. A second histogram processing unit that counts the number of pixels for each density in the grayscale image whose density value has been changed by the minimum value filter processing unit, and a threshold value based on the processing result of the second histogram processing unit. And a second counting section for counting the number of pixels having a density higher than the threshold value for each time.

[0013]

According to a first aspect of the present invention, there is provided a threshold value determining apparatus that, when there is a value at which the ratio of the number of pixels calculated by the calculation unit has a maximum value due to a change in the threshold value, sets a threshold value at which the ratio becomes a maximum value to an optimum value. If the threshold is selected as a threshold, and if there is no maximum value, by providing a threshold determination processing unit that selects the threshold when the ratio becomes zero as the optimal threshold, the ratio of the target object and the background in the image Without being affected, a threshold is chosen.

Further, the threshold value determining apparatus according to the second aspect of the present invention includes a setting unit for setting a maximum value and a minimum value of the number of pixels of the target object, and the number of pixels counted by the first counting unit is: Only when the ratio is larger than the minimum value set by the setting unit and smaller than the maximum value, if the ratio of the number of pixels calculated by the calculation unit has a maximum value due to a change in the threshold, the ratio is the maximum value Is selected as an optimal threshold when there is no maximum value, and by providing a threshold determination processing unit that selects a threshold when the ratio becomes zero as an optimal threshold, the In the case where there are a target object to be extracted and an object not to be extracted, a threshold value capable of cutting out only the target object to be extracted is selected.

Further, the threshold value determining apparatus according to the third aspect of the present invention includes a first counting means for counting the number of pixels having a density higher than the threshold value for each threshold value based on the processing result of the first histogram processing unit. A binarization process is provided by providing a processing unit and a second counting unit that counts the number of pixels having a density higher than the threshold value for each threshold value based on the processing result of the second histogram processing unit. The threshold can be obtained at high speed without any processing.

[0016]

[Embodiment 1] An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a threshold value determining apparatus according to an embodiment of the present invention. In the figure, the same reference numerals as those in the related art denote the same or corresponding parts, and a description thereof will be omitted. 6 is a threshold value t (t = 0, 1, 2,.
., 9) are sequentially changed, and a binarization processing section is provided to obtain a binary image by binarizing a grayscale image (number of gradations m = 10) for each threshold value.

An area counting section (first counting section) 7 counts the number of pixels having a density higher than the threshold value for each threshold value based on the processing result of the binarization processing section 6 and obtains the area St. , 8 counts the number of pixels surrounded by pixels having a density higher than the threshold among pixels having a density higher than the threshold for each threshold based on the processing result of the binarization processing unit 6, An internal point counting unit (second counting unit) for obtaining the internal point It.

Reference numeral 9 denotes a ratio of the internal point It calculated by the internal point counter 8 to the area St calculated by the area counter 7 for each threshold value, that is, an evaluation function Et = It /
The evaluation function calculation unit (calculation unit) 10 that calculates St has a value at which the evaluation function Et calculated by the evaluation function calculation unit 9 has a maximum value due to a change in the threshold value t, and the evaluation function Et has a maximum value. This is a threshold determination processing unit that selects the threshold value at this time as an optimal threshold value, and if there is no maximum value, selects the threshold value at which the evaluation function Et becomes zero as the optimal threshold value.

Next, the operation will be described. First, the binarization processing unit 6 performs a binarization process of setting a pixel having a density equal to or higher than the threshold t to 1 and a pixel having a density not higher than 0 to 0 with respect to the threshold image t, and obtain. Next, the area counting unit 7 calculates the area St by counting the number of pixels having a value of 1 in the entire screen.

On the other hand, the internal point counting section 8 has a value of 1
Of surrounding pixels (neighboring pixels)
Those having a value of 1 are all counted to obtain the internal score It. Here, for example, in FIG. 2, when the pixel e is a target pixel in FIG. 2 (the pixel e has a value of 1 ), the pixels located vertically, horizontally, and obliquely to the pixel e, Pixels a, b, c, d, f, g, h, and i. Therefore, the internal point counting unit 8 calculates a, b, c, d,
If all of the pixels of f, g, h, and i have a value of 1 , counting is performed.

Next, the evaluation function calculator 9 calculates an evaluation function Et = It / St at the threshold value t using the area St and the number of internal points It obtained above. The operation up to this point will be described below with reference to a screen example.

First, FIG. 3 shows a gray-scale image having a number of gradations of m = 10 and a size of 10 * 10. This image is a general gray-scale image including a target object in the background. FIGS. 4 to 13 show the results of the binarization process by sequentially increasing the threshold value from 0 to 9 for this grayscale image.

For example, when binarization is performed with the threshold value t = 4, pixels having values equal to or greater than the threshold value t = 4 are counted on the gray scale screen of FIG. 3 to obtain an area S4 = 42 (see FIG. 8). . Then, the number of internal points (the target pixel in which all the target pixel and its neighboring pixels have a value of 4 or more) is counted, and I4 = 17 is obtained (see FIG. 8). However, the internal point can be obtained by subtracting a boundary point composed of one pixel around the area S4 from the area S4. The above operation is performed from the threshold values t = 0 to 9, and the evaluation function Et = It / St is calculated for each threshold value. FIG. 14 shows a graph of the evaluation function with respect to the threshold value.

Hereinafter, the operation of the threshold value determination processing section 10 for obtaining the optimum threshold value from the evaluation function graph shown in FIG. 14 will be described. As shown in FIG. 14, the evaluation function graph is
It takes the maximum value E0 = 1.0 at the threshold value t = 0, decreases as the threshold value is increased, increases once, decreases again, and finally becomes zero. In FIG. 14, since the evaluation function graph becomes a maximum point when the threshold value is t = 6, the threshold value t = 6 is selected as an optimum threshold value (if there is no maximum point in FIG.
= 0 is selected as the optimal threshold), but the reason for selecting the threshold that is the maximum point is as follows.

In the case of a grayscale image in which a target object exists as shown in FIG. 3, when binarization is performed with the threshold value t = 0, all points are extracted, and all extracted points are also internal points. Therefore, the evaluation function reaches the maximum value (E0 = 1.0) at t = 0. Next, when the threshold value is increased, a region that is not extracted starts to appear in the background portion, and the evaluation function indicating the goodness of unity of the extracted object decreases. Further, as the threshold value is increased, only the target object is extracted in a coherent manner, so that the evaluation function increases. Further, when the threshold value is further increased, a portion that is not extracted starts to appear in the target object itself, so that the unity of the target object deteriorates, the evaluation function decreases, and the evaluation function becomes 0 when there are no more internal points. From the above, it can be seen that the unity of the target object is best when the evaluation function takes the maximum value. Therefore, if binarization is performed using this threshold value, the target object can be extracted in the best unity.

On the other hand, when the target object does not exist, the background collapses, but there is no target object to be extracted in a coherent manner. Therefore, the evaluation function does not have a maximum value, monotonically decreases, and finally becomes 0. Become. Therefore, when the evaluation function has no local maximum value, a value at which the evaluation function Et becomes 0 is selected as the threshold value. This is an optimal threshold value after it is determined that there is no target object to be extracted in the grayscale image. In this case, it can be seen that the target object does not exist without performing the binarization process thereafter.

Embodiment 2 FIG. FIG. 15 is a block diagram showing a threshold value determining device according to another embodiment of the present invention.
11 is a setting unit for setting the maximum value (maximum area Sx) and the minimum value (minimum area Sn) of the number of pixels of the target object, 12 is a maximum area setting unit for setting the maximum area Sx of the target object, 13
Is a minimum area setting unit for setting the minimum area Sn of the target object;
Reference numeral 14 denotes an evaluation function Et calculated by the evaluation function calculation unit 9 only when the area St obtained by the area counting unit 7 is larger than the minimum area Sn set by the setting unit 11 and smaller than the maximum area Sx. If there is a value that has a maximum value due to a change in the threshold value t, the threshold value at which the evaluation function Et becomes a maximum value is selected as an optimal threshold value. If there is no value that has a maximum value, the evaluation function Et becomes zero. This is a threshold value determination processing unit that selects a threshold value at this time as an optimal threshold value.

Next, the operation will be described. The grayscale image 1 is input and the evaluation function Et = It / St is calculated by the evaluation function calculation unit 9
The process is the same as that of the first embodiment up to the point determined by. In this embodiment, a case is considered where the area S of the target object to be extracted is not less than the minimum area Sn and not more than the maximum area Sx. Sn ≦ S ≦ Sx (2)

That is, the maximum area Sx and the minimum area Sn are set in advance in the maximum area setting section 12 and the minimum area setting section 13, and the threshold value determination processing section 14 sets, for example, a threshold value t
The area St extracted by binarization is compared with the preset maximum and minimum areas Sx and Sn, and only when the expression (2) is satisfied, the output result of the evaluation function calculator 9 is evaluated. Choose the optimal threshold.

As described above, since the output result of the evaluation function calculation unit 9 is evaluated only when the expression (2) is satisfied, there are a target object to be extracted and an object not to be extracted in the grayscale image. In this case, there is an effect that a threshold for properly cutting out only an object to be extracted can be selected.

Embodiment 3 FIG. FIG. 16 is a block diagram showing a threshold value determining device according to another embodiment of the present invention.
Reference numeral 15 denotes a histogram processing unit (first histogram processing unit) for counting the number of pixels for each density in the grayscale image 1;
Reference numeral 16 denotes an accumulation processing unit (first counting processing unit) that counts the number of pixels having a density higher than the threshold value for each threshold value based on the processing result of the histogram processing unit 15.

Reference numeral 17 denotes a minimum value filter for sequentially setting each pixel in the grayscale image 1 as a target pixel, and changing a density value of a pixel having the lowest density among the target pixel and its neighboring pixels to a density value of the target pixel. A processing unit 18 is a histogram processing unit ( second histogram processing unit) that counts the number of pixels for each density in the grayscale image 1 whose density value has been changed by the minimum value filtering unit 17, and 19 is a histogram processing unit 18. Is a cumulative processing unit (second counting processing unit) that counts, for each threshold value, the number of pixels having a density higher than the threshold value based on the processing result.

Next, the operation will be described. The grayscale image 1 of 10 gradations (density values 0 to 9) is converted to a histogram processing unit 1
5, the frequency Ht (the number of pixels) for each density value t (0 to 9) is obtained. The accumulation processing unit 16 accumulates density frequencies equal to or higher than the density value t of the density histogram. That is,
Each threshold t based on the processing result of the histogram processing unit 15
Each time, the number of pixels having a density higher than the threshold value t is counted. The accumulated frequency number equal to or greater than the density value t is determined by the area S of the target object extracted by binarizing the grayscale image 1 with the threshold value t.
t.

[0034]

(Equation 2)

On the other hand, the grayscale image 1 is also input to the minimum value filter processing unit 17, and the minimum value of the target pixel and its neighboring pixels is obtained. That is, each pixel in the gray-scale pixel 1 is sequentially set as a target pixel, and the density value of a pixel having the lowest density among the target pixel and its neighboring pixels is changed to the density value of the target pixel. For example, in FIG. 3, when a pixel whose i (horizontal direction) is 3 and whose j (vertical direction) is 4 is a target pixel e, the target pixel e and its neighboring pixels a, b, c, d, f, and
The pixel with the lowest density among g, h, and i has i = 2 and j
Is a pixel of 4 and its density value is “1”, so that the density value of the target pixel e is changed to “1” (see FIG. 20).

The grayscale image 1 whose density value has been changed by the minimum value filter processing unit 17 is input to the histogram processing unit 18 and the frequency NHt for each density value t (0 to 9)
(The number of pixels) is obtained. In addition, in the accumulation processing unit 19,
Density frequencies equal to or higher than the density value t of the density histogram are accumulated. That is, the number of pixels having a density higher than the threshold value t is counted for each threshold value t based on the processing result of the histogram processing unit 18. The accumulated frequency number is equal to the internal score It of the target object extracted by binarizing the grayscale image 1 with the threshold value t.

[0037]

(Equation 3)

The evaluation function calculator 9 calculates the evaluation function Et using the area St and the number of internal points It, and the threshold determination processor 10 calculates the optimum threshold value in the first embodiment.
Therefore, the description is omitted.

Here, the operation of the minimum value filter processing section 17 will be described in detail with reference to FIG. Shade image 1
Is one pixel at a time from j = 0 to i = 9, then i = 1
= 0 to i = 9,..., J = 9, and sequentially input to the minimum value filter processing unit 17 from i = 0 to i = 9. Consider a case in which a neighboring pixel g is input to the pixel storage unit 111. At this time, the one-line delay line 101
The image data of the previous row is stored, and the pixel storage unit 112
Indicates a neighboring pixel d (a pixel on one line above g), a pixel storage unit 11
3 stores the neighboring pixels a (pixels on the second row of g). In the minimum value selection unit 114a, the pixels g and d
Is compared with the pixel a, and the smaller data is output and stored in the pixel storage unit 115.

Here, the operation of the minimum value selectors 114a to 114d will be described with reference to FIG. The grayscale images x and y are the comparison unit 1
21 where the two data are compared,
According to the result, c = 0 or 1 is output and sent to the selection signal input s of the selection unit 122. The grayscale images x and y are also input to the selection unit 122, and the grayscale images x or y are output according to the value of the selection signal input s. The input conditions and output results of the comparison unit 121 and the input conditions and output results of the selection unit 122 are as shown in the table of FIG. For example, when y is smaller than x, the output result of the comparison unit 121 is c = 0, and the output of the selection unit 122 is the data y input to the b side.
Is selected and output. Therefore, the smaller data value is selectively output. The same applies to the opposite case.

Next, the operation of the minimum value filter processing unit 17 will be described in detail with reference to FIG. This figure is
This is the time when the grayscale image 1 is sequentially input and the neighboring pixel i is input to the pixel storage unit 111. In this case, the pixel storage unit 112 stores the neighboring pixel f (pixel on the first row of i), and the pixel storage unit 113 stores the neighboring pixel c (pixel on the second row of i).

Then, the minimum value selection units 114a and 114b
, The minimum value of the pixels i, f, and c is selected and stored in the pixel storage unit 115. At this time, the previously stored minimum values of the pixels a, d, and g are stored in the pixel storage unit 117, and the minimum values of the pixels b, e, and h are stored in the pixel storage unit 115.
Is stored. And the minimum value of these three data is
It is obtained by the minimum value selection unit 114d and the like. By the above processing, the pixel of interest e and its neighboring pixels a, b, c,
The minimum values of d, f, g, h, and i will be obtained.

Embodiment 4 FIG. In the above embodiment, the evaluation function Et is the internal score Et.
Is defined by the ratio of the internal point It to the area St, but as shown in FIG. 22, the external point Ot may be used instead of the internal point It. In this case, the evaluation function calculator 22
Then, Ft = St / Ot is used instead of the evaluation function Et. This evaluation function Ft is equal to the evaluation function Et of the above embodiment.
Since other shapes are shown, the other processes are the same as those in the above embodiment. The external score is obtained by increasing one pixel around the target object extracted by the binarization processing (inverse of the internal score). For example, it can be obtained by changing the target pixel to 1. Needless to say, in order to perform high-speed processing, the grayscale image 1 may be provided with a maximum value filter processing unit that takes the maximum value of the pixel of interest and its neighboring pixels.

Embodiment 5 FIG. Further, as shown in FIG. 23, an internal point counting section 8 and an external point counting section 21 may be provided, and Gt = It / Ot may be used as an evaluation function.
The same effects as in the above embodiment can be obtained.

Embodiment 6 FIG. In the third embodiment, two histogram processing units 15, 18 and two accumulation processing units 16,
19 has been described, but FIG.
As shown in FIG. 4, the histogram processing unit and the accumulation processing unit may be configured to be shared, and the same effects as in the third embodiment can be obtained. The selection processing unit 24 selects the grayscale image 1 or the grayscale image 1 that has been subjected to the minimum value filter process as the image data to be subjected to the histogram processing, and sends it to the histogram 25.

[0046]

As is evident from the foregoing description, according to the first aspect of the present invention, for each threshold, the number of pixels counted by the second counting unit relative to the number of pixels counted by the first counting portion The ratio is calculated, and if there is a value at which the ratio has a local maximum value due to a change in the threshold value, the threshold at which the ratio becomes the local maximum value is selected as the optimal threshold value. Is set as the optimal threshold, the optimal threshold is set even when conditions such as the ratio of the target object and the background occupying in the image are extremely different or when the target object does not exist. The effect is that it can be determined.

According to the second aspect of the present invention, a setting unit for setting the maximum value and the minimum value of the number of pixels of the target object is provided, and the number of pixels counted by the first counting unit is set to the setting unit. If the ratio of the number of pixels calculated by the calculation unit has a maximum value due to a change in the threshold value only when the ratio is a maximum value and is smaller than the maximum value set by Is selected as the optimal threshold,
If there is no maximum value, the threshold at which the ratio becomes zero is selected as the optimal threshold. In addition to the effect of claim 1, the target object to be extracted is extracted from the image. When there is an object that the user does not want to extract, there is an effect that a threshold value for extracting only the target object to be extracted can be selected.

Further, according to the third aspect of the present invention, based on the processing result of the first histogram processing unit, for each threshold value,
A first method for counting the number of pixels having a density higher than the threshold value
And a minimum value filter processing unit that sequentially sets each pixel in the grayscale image as a target pixel, and changes the density value of the lowest density pixel among the target pixel and its neighboring pixels to the density value of the target pixel. A second histogram processing unit that counts the number of pixels for each density in the grayscale image whose density value has been changed by the minimum value filter processing unit;
And a second counting unit that counts the number of pixels having a density higher than the threshold value for each threshold based on the processing result of the histogram processing unit. However, there is an effect that an optimum threshold can be obtained at high speed.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a threshold value determining device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of neighboring pixels.

FIG. 3 is an image shading diagram showing an example of a shading image.

FIG. 4 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 5 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 6 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 7 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 8 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 9 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 10 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 11 is an explanatory diagram illustrating an area and internal points after a binarization process.

FIG. 12 is an explanatory diagram illustrating areas and internal points after binarization processing.

FIG. 13 is an explanatory diagram illustrating an area and internal points after the binarization processing.

FIG. 14 is a graph showing an evaluation function.

FIG. 15 is a configuration diagram showing a threshold value determining device according to another embodiment of the present invention.

FIG. 16 is a configuration diagram showing a threshold value determining device according to another embodiment of the present invention.

FIG. 17 is an explanatory diagram of neighboring pixels.

FIG. 18 is a configuration diagram illustrating a minimum value filter processing unit.

FIG. 19 is a configuration diagram illustrating a minimum value selection unit of the minimum value filter processing unit.

FIG. 20 is an image shading diagram showing an example of a shading image after the binarization processing.

FIG. 21 is a graph showing an evaluation function.

FIG. 22 is a configuration diagram showing a threshold value determining device according to another embodiment of the present invention.

FIG. 23 is a configuration diagram showing a threshold value determining device according to another embodiment of the present invention.

FIG. 24 is a configuration diagram showing a threshold value determining device according to another embodiment of the present invention.

FIG. 25 is a configuration diagram showing a conventional threshold value determination device.

[Explanation of symbols]

DESCRIPTION OF REFERENCE NUMERALS 1 Shading image 6 Binary processing section 7 Area counting section (first counting section) 8 Internal point counting section (second counting section) 9, 22, 23 Evaluation function calculation section (calculation section) 10, 14, determination of threshold Processing unit 11 setting unit 15 histogram processing unit (first histogram processing unit) 16 accumulation processing unit (first counting unit) 17 minimum value filter processing unit 18 histogram processing unit (second histogram processing unit) 19 accumulation processing unit (Second counting section) 21 External point counting section (Second counting section)

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-2-24787 (JP, A) JP-A-57-26966 (JP, A) JP-A-62-23391 (JP, A) JP-A-5-296 115005 (JP, A) JP-A-6-326871 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 1/40-1/409 H04N 1/46 H04N 1/60 G06T 1/00 460

Claims (3)

(57) [Claims] 1. A binarization processing unit for sequentially changing a threshold value indicating a density value, binarizing a grayscale image for each threshold value to obtain a binary image, and a processing result of the binarization processing unit. A first counting unit that counts the number of pixels having a density higher than the threshold for each threshold based on the threshold, and a density higher than the threshold for each threshold based on the processing result of the binarization processing unit. A second counting unit that counts the number of pixels surrounded by pixels having a darker density than the threshold, among the pixels that have the threshold, and for each threshold, the number of pixels counted by the first counting unit The second
If there is a calculation unit that calculates the ratio of the number of pixels counted by the counting unit, and if the ratio of the number of pixels calculated by the calculation unit has a maximum value due to a change in the threshold, when the ratio becomes a maximum value And a threshold determination processing unit that selects a threshold when the ratio becomes zero as an optimal threshold if there is no maximum value. 2. A binarization processing section for sequentially changing a threshold value indicating a density value, binarizing a grayscale image for each threshold value to obtain a binary image, and a processing result of the binarization processing section. A first counting unit that counts the number of pixels having a density higher than the threshold for each threshold based on the threshold, and a density higher than the threshold for each threshold based on the processing result of the binarization processing unit. A second counting unit that counts the number of pixels surrounded by pixels having a darker density than the threshold, among the pixels that have the threshold, and for each threshold, the number of pixels counted by the first counting unit The second
A calculating unit that calculates the ratio of the number of pixels counted by the counting unit, a setting unit that sets the maximum value and the minimum value of the number of pixels of the target object, and the number of pixels counted by the first counting unit. Only when the ratio is larger than the minimum value set by the setting unit and smaller than the maximum value, if the ratio of the number of pixels calculated by the calculation unit has a maximum value due to a change in the threshold, the ratio is A threshold determination processing unit that selects a threshold when the local maximum value is obtained as an optimal threshold, and if there is no maximum local value, selects a threshold when the ratio becomes zero as an optimum threshold value. . 3. A first histogram processing unit for counting the number of pixels for each density in a grayscale image, and a density darker than the threshold for each threshold based on a processing result of the first histogram processing unit. A first counting unit for counting the number of pixels having the target pixel, and sequentially setting each pixel in the grayscale image as a target pixel, and determining a density value of a pixel having the lowest density among the target pixel and its neighboring pixels. A minimum value filter processing section for changing to a density value, a second histogram processing section for counting the number of pixels for each density in the grayscale image whose density value has been changed by the minimum value filter processing section,
A second counting unit that counts the number of pixels having a density higher than the threshold value for each threshold value based on the processing result of the histogram processing unit, and a first counting unit that counts the number of pixels for each threshold value. A calculating unit that calculates the ratio of the number of pixels counted by the second counting unit to the number of pixels obtained, and a value in which the ratio of the number of pixels calculated by the calculating unit reaches a maximum value due to a change in the threshold. For example, a threshold determination processing unit that selects a threshold when the ratio reaches a local maximum value as an optimal threshold, and selects a threshold when the ratio becomes zero as an optimal threshold if there is no value that reaches a local maximum value. Threshold determination device.