JP2843185B2 - Threshold value determination method and device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for binarizing a grayscale image in a visual sensor or the like in a binarization process which is a typical process for separating an object in an image from a background. And a device for determining an optimum threshold value of the threshold value.

[0002]

2. Description of the Related Art Heretofore, with respect to this type of threshold value determination, a method based on a density histogram representing the distribution of the appearance frequency of the density at each point of an image has been generally used. For example, a mode method in which a threshold value is determined at the position of a valley of a density histogram, or a threshold value is determined such that when the density histogram is divided into two classes, the sum of the variances of the densities in each class is minimized. Discriminant analysis and the like.

In addition to such a method based on the density histogram, an evaluation function indicating the quality of the binarization result based on the characteristic amount measured from each binarized image when the binarization is performed while changing the threshold value. After calculating the threshold value, there is a method of determining an optimal threshold value by judging from a change state of the evaluation function with respect to the threshold value. This method has a problem that the amount of calculation increases in proportion to the number of thresholds to be changed. The present inventor has already proposed in Japanese Patent Application No. 2-202802 a "threshold value determining device" which solves the above problem.

[0004] The present invention relates to the improvement of the invention, and therefore the above-mentioned prior application will be described first. It is considered that the optimum threshold value is a value at which the connected component of “1” is extracted as a “coherent (the boundary is not jagged)” area on the binary image. Now, regarding the point of “1” in the binary image, if the number of points of “1” in the vicinity of eight is defined as the number of neighbors, as shown in FIG. It takes a value of 0 to 7 at the boundary point. Therefore, if the sum of the number of neighbors is obtained for the entire binary image, and a value obtained by dividing the sum by the total number (area) of points of “1” is defined as the average number of neighbors, the better the united area is, the more the area occupies the area. Since the ratio of interior points is high and the ratio of boundary points is low, the average number of neighbors shows a high value.

Here, the sum of the number of adjacent pixels in the binary image with respect to the threshold value T is represented by A (T), and the area of the point "1" is represented by S (T).
The average number of neighbors R (T) is

[0006]

(Equation 1)

[0007] Further, if the number of neighbors is i (i =
If the number of 3 × 3 patterns (0,..., 8) is represented by ai, the total sum A (T) and area S (T) of the adjacent numbers can be expressed as follows.

[0008]

(Equation 2)

[0009]

(Equation 3)

In general, when an object and a background in an image have a density distribution as shown in FIG. 9, the manner in which the average number of adjacent pixels R (T) changes with respect to the threshold value T is as shown in FIG. In this case, when the binarization is performed using the value at which the average number of neighbors is maximized as a threshold value, the region is extracted as a united region, and a good binarization result is obtained.

FIG. 5 shows an embodiment of the invention of the prior application. In the figure, reference numeral 1 denotes a local data extraction unit for extracting a density G0 of a point of interest from a grayscale image and density G1,...
The selection signal n (n = n) is selected from among the densities G0,.
0,..., 8), the local data selection unit 3 selects one of the outputs Gn (n = 0,
, 8) and the density G0 of the point of interest output from the local data extraction unit 1, and the smaller value Fn (n = 0,
, 8), a histogram counting unit 4 for counting the frequency of appearance of the output Fn of the minimum value selecting unit 3, and a histogram counting unit 5 obtained by the histogram counting unit 4.
From the nine histograms for n, the threshold T is 2
After calculating the total sum A (T) and the area S (T) of the number of neighbors in the binary image at the time of binarization, and calculating the average number of neighbors R (T),
This is a threshold value determination unit that determines an optimum threshold value based on a change state of the average number of neighbors R (T) relative to the threshold value T.

Next, the operation will be described. First, as shown in FIG. 6, for example, the local data extraction unit 1
It is composed of one and two shift registers 12, and extracts the density G0 of the target point and the densities G1 to G8 of the eight neighboring points from the grayscale image to be binarized. Next, the local data selecting unit 2 selects a selection signal n (n = 0,..., 9) from the nine density data G0,.
One data Gn is selected according to 8). Next, the minimum value selector 3 compares the output Gn of the local data selector 2 with the density G0 of the point of interest sent from the local data extractor 1, and outputs the smaller value Fn. Now, change Fn to G0
, ..., G8

[0013]

(Equation 4)

## EQU1 ## By changing the selection signal n from 0 to 8 and scanning the grayscale image nine times, Fn is output from the minimum value selection unit 3 and the histogram counting unit 4 counts the frequency of its appearance, thereby obtaining nine types of histograms. Is obtained.

In the histogram of the result Fn of the result of the minimum value selection processing, the number of pixels having the density value k is represented by hn (k)
(K = 0,..., N-1; n = 0,..., 8). Here, N is the number of density levels representing the grayscale image. Therefore,

[0016]

(Equation 5)

When the number of pixels is expressed as Hn (T),

[0018]

(Equation 6)

## EQU1 ## Here, the output Fn (n = 1,..., 8) of the minimum value selection processing is the density G0 of the point of interest and the density Gn (n = 1,.
Hn (T) (n = 1,...,
8), when the grayscale image is binarized by the threshold value T, the point of interest is "1", the point corresponding to Gn in the eight neighborhoods is "1", and the other seven points are "0". 3x3 which becomes "" or "1"
Represents the number of patterns. Hn (T) (n = 1,...,
8) includes the number of 3 × 3 patterns in which the number of neighbors is 1 to 8 in the binary image with respect to the threshold value T, and the sum of all of these includes the number of neighbors i (i = i =
1,..., 8) is the total sum of i times the 3 × 3 pattern number ai. That is, the sum A (T) of the number of neighbors in the binary image with respect to the threshold value T is

[0020]

(Equation 7)

## EQU1 ## The area S (T) in the binary image with respect to the threshold value T is the number of pixels equal to or larger than the threshold value T in the histogram of F0 = G0.

[0022]

(Equation 8)

## EQU1 ## As described above, the sum A (T) of the number of neighbors and the area S (T) are obtained by the minimum value selection processing result Fn (n = 0,
…, 8) are obtained from 9 types of histograms,
Based on these, the average adjacent number R (T) can be calculated.

In the above example, Fn (n = 0,
, 8), the case where the grayscale image is scanned 9 times and counted while changing the selection signal n from 0 to 8 has been described. As shown in FIG. 7, the minimum value selecting unit 3 and the histogram counting unit 4 are used. Are arranged in parallel, and (G0, G0),
By giving (G1, G0),..., (G8, G0), nine types of histograms for Fn (n = 0,..., 8) can be obtained by scanning the grayscale image only once. Can also. In this case, the local data selection unit 2 becomes unnecessary. Needless to say, the minimum value selection unit 3 having (G0, G0) as two inputs can be omitted by directly inputting G0 to the histogram counting unit 4.

[0025]

Since the conventional threshold value determining apparatus is configured as described above, the histogram of the output Fn (n = 0,..., 8) of the minimum value selecting section is converted to the selection signal n. It is necessary to scan and count the grayscale image nine times while changing the value from 0 to 8, which requires processing time. Also, Fn
Nine histograms for (n = 0,..., 8)
Although it is possible to obtain a grayscale image by scanning only once, it is necessary to arrange nine minimum value selecting units and nine histogram counting units in parallel, and there is a problem that the circuit scale becomes large. Was.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide a method and an apparatus for determining a threshold value capable of reducing the number of scans to reduce the processing time and the circuit scale. The purpose is to get.

[0027]

According to the threshold value determining method of the present invention, the density G0 of a point of interest from a grayscale image and its density G0
Among the neighboring points, four points of density G1,..., G4 selected so as not to include two points symmetrical with respect to the point of interest are extracted, and the density G0,. Grayscale image 1
One is selected for each scan, and the selected density G
n (n = 0,..., 4) and the smaller value Fn of the density G0 of the target point are selected, and the appearance frequency of each Fn obtained by scanning the grayscale image five times is counted to obtain five types of values. A histogram is generated, and a total sum A (T) and an area S (T) of the number of neighbors are obtained based on the generated histogram to obtain an average number of neighbors R (T).
Is calculated, and then the optimum threshold value is determined based on the state of change of the average number of neighbors R (T) with respect to the threshold value T.

A threshold value determining method according to another invention of the present application is arranged so that a density G0 of a point of interest from a grayscale image and two points symmetrical with respect to the point of interest among eight points near the density G0 are not included. , G4 at the four selected points are extracted, and the smaller value Fn (n = 0,...) Of each of the extracted densities GO,. ,
4) is selected, the appearance frequency of each Fn obtained by scanning the grayscale image once is counted, and five types of histograms are generated. Based on the generated histograms, the total sum A (T) and the area of the number of neighbors are calculated. After calculating S (T) and calculating the average number of neighbors R (T), an optimum threshold value is determined based on the state of change of the average number of neighbors R (T) with respect to the threshold value T.

On the other hand, the threshold value determining apparatus according to the present invention selects the density G0 of the target point from the grayscale image and two points near the density G0 so as not to include two points symmetrical with respect to the target point. .., G4 of the four points, and a selection signal n (n = 0,..., 4) from the five points of density GO,. A first selector for selecting and outputting one, a second selector for selecting and outputting the smaller value Fn of the output Gn of the first selector and the density G0 of the point of interest, A histogram counting unit for counting the frequency of appearance of each output Fn of the second selecting unit; and a sum A (T) and an area S (T) of the number of neighbors based on a histogram obtained for each Fn from the histogram counting unit. Is calculated and the average number of neighbors R (T) is calculated. It is obtained by a threshold value determination unit that determines an optimal threshold Hazuki.

Further, the threshold value determining apparatus according to another invention of the present application provides a five-point density G0,.
A selection section for selecting and outputting the smaller value Fn (n = 0,..., 4) of each of G4 and the density G0 of the point of interest, and simultaneously counting the frequency of appearance of each output Fn of the selection section. And five histogram counters provided to the threshold determiner.

[0031]

According to the threshold value determining method of the present invention, the density G0 of the target point from the grayscale image and the four points selected from the eight neighboring points so as not to include two points symmetrical with respect to the target point. By extracting the densities G1,..., G4, the optimum threshold value can be determined from the five types of histograms.

The threshold value determining method according to another aspect of the present invention is a method of determining the threshold values of five points GO,.
, And the smaller value Fn (n = 0,..., 4) of the density G0 of the target point is selected, and the frequency of appearance of each Fn obtained by scanning the grayscale image once is counted. Since the different types of histograms are generated, only one scan of the grayscale image is required, and the circuit scale in that case can be reduced.

On the other hand, the extraction unit of the threshold value determining apparatus according to the present invention determines the density G0 of the point of interest from the grayscale image and its density G0.
The density G1,..., G4 of four points selected so as not to include two points symmetrical with respect to the point of interest among the neighboring points is extracted, and
The selection section of the five points, the density GO, sent from the extraction section,
, G4, one is selected according to the selection signal n (n = 0,..., 4), and the output Gn is used as one input of the second selection unit.

Further, in the threshold value determining device according to another invention, five (or four) selection units corresponding to the second selection unit are arranged in parallel, and (G0, G0, G0), (G1, G0),..., (G4,
G0), and by providing five histogram counters for simultaneously counting the appearance frequency of each output Fn and providing the same to the threshold determination unit, the grayscale image only needs to be scanned once. The selection unit becomes unnecessary, and the circuit scale of the selection unit and the histogram counting unit can be reduced to about half of the conventional size.

[0035]

[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 configuration of an embodiment. In the drawing, reference numeral 1 denotes a density G0 of a point of interest from a grayscale image and two points symmetrical with respect to the point of interest among eight neighboring points. Density G1, ..., G of four points selected not to be
, G4 output from the local data extraction unit 1 in accordance with a selection signal n (n = 0,..., 4). The local data selection unit (first selection unit) 3 for selecting one of the three is the output Gn (n = 0,..., 4) of the local data selection unit 2 and the point of interest output from the local data extraction unit 1. The minimum value selector (second selector) which outputs a smaller value Fn (n = 0,..., 4) as compared with the density G0, and 4 is the frequency of appearance of the output Fn of the minimum value selector 3. , A histogram counting unit 5 for counting the number
From the five types of histograms for n, the threshold T is 2
After calculating the total sum A (T) and the area S (T) of the number of neighbors in the binary image at the time of binarization, and calculating the average number of neighbors R (T),
This is a threshold value determination unit that determines an optimum threshold value based on a change state of the average number of neighbors R (T) with respect to the threshold value T.

Next, the operation will be described. First, the local data extraction unit 1 includes five latches 1 as shown in FIG.
The shift register 12 is composed of one and one shift register 12. The density G0 of the point of interest from the grayscale image to be binarized and the density of the eight points near the point do not include two points symmetrical with respect to the point of interest. , G4 of the four points selected in step (1) are extracted. Next, the local data selecting unit 2 selects one of the five density data G0,..., G4 output from the local data extracting unit 1 in accordance with the selection signal n (n = 0,. Select Gn. Next, the minimum value selector 3 compares the output Gn of the local data selector 2 with the density G0 of the point of interest sent from the local data extractor 1, and outputs the smaller value Fn. Now, if Fn is represented using G0, ..., G4,

[0037]

(Equation 9)

## EQU4 ## By changing the selection signal n from 0 to 4 and scanning the grayscale image five times, each Fn is output from the minimum value selection unit 3 and the histogram counting unit 4 counts the frequency of its appearance to obtain five types of Fn. A histogram is obtained.

Considering a 3 × 3 local area 13 and a local area 14 shifted by one pixel to the right as shown in FIG. 3A, the densities (G0, G1) of two points in the local area 13 are considered. And the densities (G5, G0) of the two points in the local region 14 are the same. Therefore, F1 = min (G0,
G1) and F5 = min (G5, G0) of the local area 14 have the same value. In other words, looking at the whole image, F1 and F5
Have the same histogram.

As shown in FIG. 3B, considering a 3 × 3 local area 13 and a local area 15 shifted one pixel to the right and one pixel downward, the density of two points in the local area 13 is considered. (G0,
G2) and the density (G6, G0) at two points in the local area 15 are
Each is the same. Therefore, F2 = m in the local region 13
in (G2, G0) and F6 = min (G
6, G0). That is, the histograms of F2 and F6 are the same in the whole image.

Similarly, the histogram of F3 and F7,
The histograms for F4 and F8 are the same. From the above, F
In the histogram of n, the number of pixels hn having the density value k
(k)

[0042]

(Equation 10)

The following relationship is established. Also, the number of pixels Hn (T) satisfying the expression (5) is

[0044]

[Equation 11]

The following relationship is established. Therefore, when equation (11) is applied to equation (7), the total sum A (T) of the number of neighbors becomes

[0046]

(Equation 12)

## EQU1 ## That is, the sum A of adjacent numbers
(T) and the area S (T) are the result Fn (n =
0,..., 4) are obtained from five types of histograms, and based on these, the average number of neighbors R (T) can be calculated.

Embodiment 2 FIG. By the way, in the above embodiment, F
The case where the histogram of n (n = 0,..., 4) is scanned and counted five times while changing the selection signal n from 0 to 4 and counted, as shown in FIG. 5 and 5 histogram counters 4 are arranged in parallel, and (G0, G0), (G1) are input to the respective minimum value selectors 3 constituting the selector 30 of another invention of the present application.
, G0),..., (G4, G0), five types of histograms for Fn (n = 0,..., 4) can be obtained by scanning the grayscale image only once. . In this case, the local data selection unit 2 becomes unnecessary. Needless to say, the minimum value selection unit 3 having (G0, G0) as two inputs can be omitted by directly inputting G0 to the histogram counting unit 4.

[0049]

As described above, according to the threshold value determining method and apparatus according to the present invention, the density G0 of the target point from the grayscale image and the eight points near the target point are symmetrical with respect to the target point. Four points of density G1, ..., G4 selected not to include two points
, G4 are selected one by one for each scan of the grayscale image from the extracted five points of density Go,..., G4, and the selected density Gn (n = 0,. Concentration G
The smaller value Fn of 0 is selected, the appearance frequency of each Fn obtained by scanning the grayscale image five times is counted, and five types of histograms are generated, and the optimal threshold value is determined based on the histogram. This eliminates the necessity of scanning the grayscale image nine times as in the related art, so that only five scans are required and the processing time can be shortened.

According to another embodiment of the present invention, there is provided a method and apparatus for determining a threshold value.
Each of the point densities GO,..., G4 and the point of interest density G0
Since the smaller value Fn (n = 0,..., 4) is selected and the frequency of appearance of each Fn obtained by scanning the grayscale image once is counted to generate five types of histograms, The scanning of the grayscale image can be performed only once, and there is no need to arrange nine minimum value selecting units and nine histogram counting units in parallel as in the related art, so that the circuit scale can be greatly reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing one embodiment of the present invention.

FIG. 2 is a configuration diagram illustrating an embodiment of a local data extraction unit according to the present invention.

FIG. 3 is an explanatory diagram showing the identity of a minimum value selection process.

FIG. 4 is a configuration diagram showing one embodiment of another invention.

FIG. 5 is a configuration diagram showing a conventional example.

FIG. 6 is a configuration diagram showing a local data extraction unit in a conventional example.

FIG. 7 is a configuration diagram showing another conventional example.

FIG. 8 is an explanatory diagram of the number of neighbors.

FIG. 9 is an explanatory diagram showing a density distribution of a target and a background in an image.

FIG. 10 is an explanatory diagram showing how the average number of neighbors changes with respect to a threshold.

[Explanation of symbols]

 Reference Signs List 1 local data extraction unit 2 local data selection unit (first selection unit) 3 minimum value selection unit (second selection unit) 4 histogram counting unit 5 threshold value determination unit 30 selection unit

Continuation of the front page (56) References JP-A-4-86959 (JP, A) JP-A-3-68084 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04N 1 / 40-1/409 H04N 1/46 H04N 1/60

Claims (4)

(57) [Claims] When a certain point of a binary image with respect to a threshold value T is "1" with respect to determination of a threshold value at the time of binarizing a grayscale image, a point near the point 8 is adjacent to the point of interest. The sum A (T) of adjacent numbers indicating the number of “1” points and “1”
Based on the area S (T) indicating the total number of points
(T) = A (T) / S (T), and a threshold value determining method for determining an optimum threshold value based on a change state of the average number of neighbors R (T) with respect to the threshold value T. The density G0 of the point of interest from the image and the densities G1 of the four points selected so as not to include two points symmetrical with respect to the point of interest among the eight neighboring points.
G4 is extracted, and one density is selected from the five extracted densities GO,..., G4 for each scan of the grayscale image, and the selected density Gn (n = 0,..., 4) is noted. The smaller value Fn of the point density G0 is selected, the appearance frequency of each Fn obtained by scanning the grayscale image five times is counted, and five types of histograms are generated. Based on the generated histogram, After calculating the sum A (T) of the number of neighbors and the area S (T) of the point of “1” to calculate the average number of neighbors R (T),
A threshold value determining method characterized in that an optimum threshold value is determined based on a change state of an average number of neighbors R (T) with respect to a threshold value T. 2. A threshold value for binarizing a grayscale image, when a point of the binary image with respect to the threshold value T is “1”, the point near the point 8 is adjacent to the point of interest. The sum A (T) of adjacent numbers indicating the number of “1” points and “1”
Based on the area S (T) indicating the total number of points
(T) = A (T) / S (T), and a threshold value determining method for determining an optimum threshold value based on a change state of the average number of neighbors R (T) with respect to the threshold value T. The density G0 of the point of interest from the image and the densities G1 of the four points selected so as not to include two points symmetrical with respect to the point of interest among the eight neighboring points.
G4 is extracted, and the smaller value Fn (n = 0,..., 4) of each of the extracted five points of density GO,..., G4 and the density of the point of interest G0 is selected. The number of appearances of each Fn obtained by scanning twice is counted to generate five types of histograms. Based on the generated histograms, the sum A (T) of the adjacent numbers and the area S (T ) To calculate an average number of neighbors R (T), and then determine an optimum threshold value based on a change state of the average number of neighbors R (T) with respect to the threshold value T. . 3. A method for determining a threshold value for binarizing a grayscale image, wherein when a certain point of the binary image with respect to the threshold value T is "1", a point near the point 8 is adjacent to the point of interest. The sum A (T) of adjacent numbers indicating the number of “1” points and “1”
Based on the area S (T) indicating the total number of points
(T) = A (T) / S (T), and a threshold value determining apparatus for determining an optimum threshold value based on a change state of the average number of neighbors R (T) with respect to the threshold value T. The density G0 of the point of interest from the image and the densities G1 of the four points selected so as not to include two points symmetrical with respect to the point of interest among the eight neighboring points.
An extracting unit for extracting G4, and 5 output from the extracting unit.
A selection signal n (n = 0, n = 0) from among the point densities GO,.
, 4), and selects and outputs the smaller value Fn of the output Gn of the first selector and the density G0 of the point of interest. A second selecting unit, a histogram counting unit for counting the frequency of appearance of each output Fn of the second selecting unit, and a sum A () of the adjacent numbers based on the histogram obtained for each Fn from the histogram counting unit T) and the area S (T) of the point "1" are calculated to calculate the average number of adjacent neighbors R (T). A threshold value determining unit for determining the threshold value. 4. A method for determining a threshold value for binarizing a grayscale image, when a certain point of the binary image with respect to the threshold value T is “1”, the point near the point 8 is adjacent to the point of interest. The sum A (T) of adjacent numbers indicating the number of “1” points and “1”
Based on the area S (T) indicating the total number of points
(T) = A (T) / S (T), and a threshold value determining apparatus for determining an optimum threshold value based on a change state of the average number of neighbors R (T) with respect to the threshold value T. The density G0 of the point of interest from the image and the densities G1 of the four points selected so as not to include two points symmetrical with respect to the point of interest among the eight neighboring points.
An extracting unit for extracting G4, and 5 output from the extracting unit.
, G4 of the point and the density G0 of the point of interest
A selection unit for selecting and outputting the smaller value Fn (n = 0,..., 4) of each of the above, five histogram counting units for simultaneously counting the appearance frequency of each output Fn of the selection unit, Based on the histograms obtained simultaneously for each Fn from each histogram counter, the sum A
(T) and the area S (T) of the point of “1” are obtained, and the average number of neighbors R
After calculating (T), the average number of neighbors R for the threshold T
(T) a threshold value determining unit that determines an optimum threshold value based on the change state.