JPH01158577A - Method for background erasing and binarization processing for line graphic picture and its device and picture processor for fingerprint picture - Google Patents

Abstract

PURPOSE:To obtain a satisfactory binarization image in any case by dividing the whole picture into local areas in each of which line graphics can be regarded as parallel straight line groups and executing a specific processing for the whole picture. CONSTITUTION:After a picture is divided into the local areas in each of which the line graphics can be regarded as the parallel straight line groups, the direction of the line graphics is calculated for each local area, and a weighted mean value, the center of which is each picture element, is obtained for each local area using the mask pattern of each direction. Thereafter, a binarizing is executed by binarizing according to the comparison of the weighted mean value and a threshold, namely, by binarizing while judging whether the line graphics are continued in the direction or not using not only the information of the picture level of one picture element but also that of the picture levels of picture elements near the former picture element at the time of executing the binarizing. Thus, the binarizing processing can be executed which is resistant to a noise, and the satisfactory binary picture can be obtained even for a low-quality variable density picture.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application 1] The present invention relates to a method and apparatus for binarizing a line graphic image, and in particular to a method and apparatus for binarizing a line graphic image, and particularly for converting a gray scale image of a line graphic into a binary value without impairing the continuity of the lines. 2 Line image images suitable for use when converting
The present invention relates to a value processing method and device.

[Conventional technology]

Generally, when converting a digital image, that is, an image that is expressed as a set of units called pixels and gives each pixel a discrete value (image level) depending on its brightness, a threshold value is set in some way. However, a method is usually adopted in which one of the two values to be given is determined depending on whether or not the value of each pixel exceeds its threshold value.

However, when this method is applied to line figure images,
In areas where lines are blurred, the image level falls below the threshold, and in a binary image, lines may be interrupted at those areas. Conversely, in areas where two or more lines are close to each other, the image level in the area between the lines exceeds the threshold, and the normally possible multiple lines become one in a binary image. It also happens.

In contrast, for example, Japanese Patent Laid-Open No. 62-29267 discloses a binarization method for low-quality line graphic images. The binarization method proposed in JP-A No. 62-29267 regards a surface surrounded by adjacent pixels as one local plane, calculates the slope of the shading for each local plane, and labels the slope. A label image is created, line figure candidate pixels are extracted using the label image, a threshold value is determined based on the extracted line figure candidate pixels, and the image is binarized.

That is, this invention extracts a line figure, determines a threshold value based on its image level, applies the obtained threshold value only to the vicinity, and binarizes it, thereby processing a dirty image. The aim is to obtain a good binary image.

However, since binarization is performed without comparing the size of each pixel with the threshold value, it is still the same as in the past, so parts of the line figure that are below the threshold value due to noise If there is, the line will be broken. Furthermore, if there is a bridge-like portion between two lines that exceeds the threshold value, lines that should normally be possible may become connected. Furthermore, if the quality of the line figure is poor and the line is discontinuous, it is not possible to perform good line figure extraction with only information about adjacent pixels due to noise, and it is difficult to perform good binarization processing. It still had the drawback of not being able to do so.

When binarizing a digitized grayscale image containing line figures, whether the same threshold is applied to the entire image or locally different thresholds are applied, areas where the lines are blurred In this case, lines may become disconnected, or conversely, in areas where two or more lines are close to each other, lines that should normally be connected may become connected.

In addition, when trying to obtain a binarized image of a line figure when a background image is superimposed on an image containing a line figure, first erase the background image, extract the line figure, and then perform the binarization process. There is a need to do.

A typical example is when a fingerprint that has been revealed by some method is photographed using a television camera, but there is no device that can perform the background erasure and binarization processing described above continuously, and the photographed fingerprint cannot be photographed by a human being. The current situation is that it is being traced.

[Problem to be solved by the invention]

The present invention has been made in view of the above circumstances, and it is an object of the present invention to provide a method and apparatus for binarizing a line figure image, which makes it possible to obtain a good binarized image in any case. With the goal.

[Means for Solving the Problems 1] The first aspect of the present invention is a method for binarizing line figure pixels to obtain a binary image of a line figure image from a digitized A-made image containing line figures. It consists of the following steps.

(a) Divide the entire image into local regions where line figures can be regarded as a group of parallel straight lines within that region.

(b) Next, calculate the direction of the line figure regarded as a group of straight lines for each local area.

(c) Calculating a weighted average value centered on each pixel using a mask pattern determined for each direction.

(d) Depending on whether or not these weighted average values exceed a preset threshold, one of the two values is applied to the central pixel of the mask pattern, and this process is applied to each of the local areas. What to do for all pixels in each area.

Next, as a preferred specific means of the first aspect of the present invention, as a method of calculating the direction of the line figure for each local area, a mask pattern determined for each direction is sequentially used to calculate the direction of the line figure at the center of each pixel. For all pixels, calculate the direction of the line figure in the vicinity of that pixel based on the direction of the mask pattern where the result is the maximum, and calculate the direction of the line figure in the vicinity of that pixel for each local area. The number of directions calculated for each pixel is calculated, and the direction with the largest number of directions is set as the direction of the line figure in that local area.

Furthermore, in a certain local area, the maximum number of directions is compared with the average value of the number of each direction, and if the difference is not greater than a preset threshold, that local area is directionless. , it is determined that the area does not include a line figure, and one of the binary values is uniformly given to all pixels within that local area.

Furthermore, in this case, it is preferable to divide the area as necessary and repeat the same process.

For example, for an image such as a fingerprint image in which the curvature of a linear image differs depending on the location or there are branches, the binary processing method that judges the local area as having no direction is applied, and the entire image is - If you divide the area into local areas of various sizes and try to assign the direction of the line figure to each area, the direction of the line figure is not constant within the area, and therefore the maximum number of directions and The average value of the number in each direction does not exceed the threshold,
Therefore, the local area may be determined to be non-directional.

FIG. 18 shows a flowchart of the binarization processing procedure in such a case. That is, when obtaining a binary image of a line figure image from a digitized grayscale image containing a line figure, (a) the entire image is divided into local regions in which the line figure can be regarded as a group of parallel straight lines within that area;

(b) Sequentially using mask patterns determined for each direction, perform a sum-of-products operation between the mask pattern and the pixel, centering on each pixel, (c) A mask pattern that maximizes the result of the sum-of-products operation. (d) calculate the direction of the line figure for each pixel for each local area, and aggregate the results for each local area;

(e) As a result of the aggregation, compare the number of directions with the highest number with the average value of the number of each direction, and if the difference is greater than the preset threshold, the A weighted average value centered on each pixel is determined using a mask pattern, and depending on whether or not these weighted average values exceed a preset threshold value, one of the two values is applied to the mask. This process is applied to the pixel at the center of the pattern, and this process is performed for all pixels in each local area, and the difference between the maximum number of directions and the average value of the number of each direction is determined by the preset value. If it is not above the threshold, the local area is further divided into smaller areas.

(f) For each divided small region, the calculation results of the line figure for each pixel are aggregated in the same way as for the local region, and the maximum number of directions is the number of directions in each direction. If the difference is greater than or equal to a preset threshold, the direction with the highest frequency is set as the direction of the line figure in that small area, and each pixel is determined using the mask pattern in that direction. (g) Depending on whether these weighted average values exceed a preset threshold, one of the two values is determined as the center of the mask pattern. (h) If the average difference between the numbers in each direction becomes equal to or less than the threshold value again, the small area is further divided, and the same process is applied to the divided areas. Appropriate processing can be achieved by repeating the 14 operations described above until the divided region has a direction or the size of the divided region becomes less than or equal to a preset threshold.

A second invention of the present invention is an apparatus that can suitably implement the first invention, and is comprised of the following phases 1 to 1.

■An image input device that captures an image of a line figure.■A first image memory that stores a grayscale image of a line figure input to this image input device.■This grayscale image is divided into local areas, and the direction of the line figure is determined for each area. Direction calculation circuit for calculation ■ Direction recording memory for recording this calculated direction ■ Z binarization circuit for reading out this recorded direction and binarizing the grayscale image stored in the first image memory ■ This 2 A second image memory that stores the digitized signal. An output switching circuit that selects the output signals of the image input device, the first image memory, and the second image memory. The selected output signal is converted into an analog signal. Convert to D
/A converter [phase] Monitor for displaying this analog signal Next, the third aspect of the present invention will be explained.

Generally, CCD (charge coupled de
vice) When trying to capture the fluorescence emitted by a substance attached to an object with a camera, etc., the pattern on the surface of the attached object is inevitably superimposed as a background.

For example, in a fingerprint detection device using a laser fluorescence method, which irradiates a fingerprint-attached object with laser light and detects the fingerprint by the fluorescence emitted by the fingerprint, when trying to take a fingerprint image, the pattern of the fingerprint-attached object inevitably appears. will also be superimposed and photographed as a background image.

In this way, when the background image is superimposed on the line figure of interest, the above-described line figure image binarization processing method or line figure image binarization processing apparatus cannot be immediately applied.

In such a case, it is necessary to delete the background image, extract a line figure image, and apply the above-described line figure image binarization processing method or apparatus to the extracted line figure image.

A third aspect of the present invention is a method for background erasing and binarizing a line graphic image suitable for application in such cases, comprising:
As shown in the flowchart of the procedure in FIG. 19, a grayscale image in which the background is superimposed on the line figure to be extracted (hereinafter referred to as image A) and a grayscale image with only the background (hereinafter referred to as image B) are extracted. , a background erasure and binarization processing method for erasing the background, extracting only the line figure to be extracted, and further binarizing the gray scale image of the extracted line figure, which consists of the following steps.

(b) Separating a partial image of the background image from image A.

(b) Extracting a set C of an appropriate number of pixels from the separated partial image of the background image.

(c) Extracting a set of pixels corresponding to set C in image B.

(d) Performing a regression analysis on the gradations of pixels corresponding to the set C and the set, and determining a conversion coefficient between the gradations of the background part of the image A and the corresponding part of the image B from the result.

(E) While correcting image B using the conversion coefficient, image A
To subtract image B from and extract line figures.

(f) Divide the entire image into local regions in which line figures can be regarded as a group of parallel straight lines within that region.

(g) Calculating the direction of a line figure regarded as a group of straight lines for each local area.

(H) Calculating a weighted average value centered on each pixel using a mask pattern determined for each direction.

(1) Depending on whether or not these weighted average values exceed a preset threshold, one of the two values is applied to the central pixel of the mask pattern, and this process is applied to each of the local areas. What to do for all pixels in each area.

Here, as a means for separating a partial image of the background image from the image A, a threshold value may be set at the image level, and the portion below this threshold value may be used as the background).

Furthermore, as a means for separating a partial image of the background image from the image A, it is preferable to utilize the difference in texture between the line figure part and the background part.

Next, the fourth invention of the present invention is a background erasing and binarization processing device for line graphic images capable of suitably implementing the method of the third invention, which comprises the following i) to xivl. ing.

i) An image input device that captures images.

li) A first image memory that stores an image (image A) in which a background is superimposed on a line figure image input to this image input device.

1ii1 A second image memory that stores the background-only image (image B) input to the image input device.

iv) means for separating partial images of the background image of image A;

(2) Means for extracting a set C of an appropriate number of pixels from this partial image.

vi) means for extracting a set of pixels corresponding to set C in image B;

viil A regression analysis circuit that performs regression analysis on the gradations of pixels corresponding to set C and the set, and calculates a conversion coefficient between the gradations of the background part of image A and the corresponding part of image B from the result.

viii A background image subtraction circuit that performs four arithmetic operations to subtract image B from image A while correcting image B using a conversion coefficient, and stores the result in a first image memory or a second image memory.

ix) A direction calculation circuit that divides this background-removed grayscale image into local regions and calculates the direction of a line figure for each region.

X) Direction recording memory that records the calculated direction.

xi) A binarization circuit that reads out the recorded direction, binarizes the grayscale image stored in the first image memory or the second image memory, and stores it in the other image memory.

xii) An output switching circuit that selects the respective output signals of the image input device, the first image memory, and the second image memory.

xiii) A D/A converter that converts the selected output signal into an analog signal.

xivl Monitor 6 for displaying this analog signal Note that when the fourth inventive device of the present invention is applied to an image processing device for fingerprint images, the device includes a background erasing and binarization processing device for line figure images and its The direction calculation circuit of the device, which is composed of peripheral devices, takes into account the curvature and spacing of each line (referred to as a ridge) in the fingerprint image when dividing the entire image into local areas, and divides the entire image into local areas. The circuit is configured so that the ridges form a group of parallel straight lines within the area, and the size of the local area is made suitable for a fingerprint image.

[Effect 1 Below, the structure of the present invention will be specifically explained along with its operation.

FIG. 1 is a flowchart showing the procedure of the present invention.

Each item will be explained in detail below.

First, as shown in Figure 2, within the area of the line figure,
Divide into local areas that can be regarded as a group of parallel straight lines. That is, here, a small rectangle is one local area.

Pay attention to the curvature of the line figure and divide it into areas that can be considered straight lines even if the curvature is large.

Next, the direction of the line figure, which is regarded as a group of straight lines, is calculated for each local area. Although various known methods can be used to calculate the direction, the method described below is effective for low-quality images with broken lines or blurred lines.

First, the directions are quantized into four directions as shown in FIG. 3, or eight directions as shown in FIG. Here, a case will be described in which quantization is performed in four directions, but the concept is similar in the case of quantization in eight directions.

For example, four mask patterns corresponding to the four directions shown in FIG. 3 are created as shown in FIG. 5, respectively. In addition, the size and numerical value of the mask pattern depend on the line thickness of the line shape,
It may be changed appropriately depending on the curvature and spacing. Therefore, for each pixel in one local area, a weighted sum is obtained using four mask patterns, and the direction in which the absolute value is the maximum is determined as the direction of the line figure in the vicinity of that pixel.

This is done for all pixels within that local area. For example, if the local area is a rectangular area with 16 pixels in the vertical direction and 16 pixels in the horizontal direction, the direction is calculated for a total of 256 pixels. Among the four directions, the direction that occurs the most is determined as the direction of the line figure in that local area. This is performed for the entire screen for each local area. For example, the case of FIG. 2 becomes as shown in FIG. 6. In FIG. 6, the numerical values represent the calculated directions. In this case, if the difference between the maximum number of directions and the average value of the number of four directions is less than a preset threshold, the local area is directionless, that is, it is an area that does not include a line figure. It can be determined that

Here, in the direction of the line figure near the pixel, O,
If one of 1, 2, or 3 is always assigned, the average value of the numbers in the four directions will always be the number of pixels in the local area/4 (in this case, 256/4 = 64), but When determining the direction of neighboring line figures, some criteria (for example, the direction where the absolute value of the weighted sum is the maximum and the direction where the absolute value is the minimum,
If the value is less than a preset threshold value, etc.), and the pixel is given an attribute of no direction, and that pixel is excluded from the aggregation performed within the local area, then the number of pixels in the local area It will not be /4. Therefore, the expression "average value of numbers in four directions" was used.

Furthermore, in cases where the curvature of a line figure differs depending on the location, such as in a fingerprint image, or where there are branches, it may not be appropriate to divide the entire image into local areas of uniform size. For example, since the curvature is large in the lower right region of FIG.

In such a case, the difference between the maximum number of directions and the average value of the number of four directions is less than the threshold value, and it can be determined that there is no direction. ), the lower right local area ml is further divided into small areas, and within this small area, the directions of the line figures near each pixel are totaled, and the direction with the largest number is taken as the direction of that small area. By using an area as a processing unit in the next step, an appropriate area size can be automatically selected. In addition, even if the area is divided into small areas, if it is determined that the area is too large and has no direction, the division will proceed further (in this case, the criterion for stopping the division is that the divided area is Either the direction is no longer present or the size of the divided area is less than or equal to a preset threshold value can be used.

Next, using the mask patterns created for each direction,
For each pixel in the local area, a weighted average value centered around that pixel is calculated, and one of the two values is assigned to that pixel depending on whether or not it exceeds a preset threshold. The value given is performed for all pixels in each area.

In that case, in the previous step, for the region determined to have no direction, the calculation to obtain the weighted average value is not performed, and one of the two values is uniformly given depending on the case.
It is possible to prevent background noise from appearing in a binary image.

For example, when the directions are quantized in four directions, the mask pattern shown in FIG. 8 can be used. This can also be changed appropriately depending on the thickness, curvature, and spacing of the line figure. It is not necessary that all the numerical values are l and 0, and the weighted average value may be calculated using other numerical values depending on the case.

The operation of calculating this average value is equivalent to checking whether a line figure continues in the direction determined in the previous step, centering on a certain pixel, and for this reason, it is a binarization processing method that is resistant to noise. .

Furthermore, although the threshold used for binarization can be changed for each local area, sufficient effects can be obtained by implementing the present invention even if the same threshold is uniformly used for the entire screen;
A different threshold value is better in that it can prevent interruptions in figures and differences in line thickness due to differences in threshold values depending on location.

[Example] Embodiments of the present invention will be described below based on the drawings.

Example 1 FIG. 9 shows two line graphic images suitable for carrying out the present invention.
FIG. 2 is a block diagram showing an example of the configuration of a value conversion processing device.

In FIG. 9, ■ is an image input device. Depending on the case, CCD (charge coupled dev
ice) A camera, drum scanner, etc. can be used. The input gray scale image of the line figure is stored in the first image memory 2.

The direction calculation circuit 7 divides the image stored in the first image memory 2 into local regions, and calculates the direction of a line figure for each region. The calculated direction is recorded in the direction recording memory 10.

The binarization circuit 12 binarizes the grayscale image in the first image memory 2 while reading out the calculated direction from the direction recording memory 10, and stores the result in the second image memory 9.

The output switching circuit 3 connects the image input device 1 and the first image memory 2.
．． A signal from the second image memory 9 is selected, and this signal passes through the D/A converter 4 and is displayed on the monitor 5.

The address generation circuit 6 performs the following for each process of each circuit described above.
The data in the first image memory 2 is sent to each circuit in an orderly and appropriate timing, and the processed data is sent to the second image memory 2.
The data stored in the first image memory 2 or the second image memory 9 is sent to the switching circuit 3.

The timing controller 8 is for sending control signals to each circuit in order to perform the processing of each circuit in an orderly manner, and can be operated by commands from the operation switch 11. Note that the results of these operations can be viewed on the monitor 5.

Embodiment 2 Next, an example in which a grayscale image of a line figure is binarized by implementing the present invention will be described.

FIG. 1O shows 7×7 pixels extracted from the image to be processed, and each numerical value represents an image level quantized into 256 gradations. The direction of the local area including this pixel is 0 in FIG. 3, and is 1.3.4.6 from the top.
The level of the seventh row is high, and if we use the high level part as a figure, we can see that there are three lines in the horizontal direction. In this image, the image level of pixels (3, D) and (4, D) has decreased due to noise, and no matter how you choose the threshold value, normal processing cannot prevent line breaks. . This is because in order to avoid interruptions, the binarization threshold must be set to 11.
7 or less, but in that case, pixels (2,B), (2,C), (2,E
) and (2°F), and the two lines will be attached to each other.

Applying the present invention and taking the average value using the mask in the direction O shown in FIG.
The average of the 5 pixels C), (3,D), (3°E), (3,F,) is (132+130+118+128+13415) = 128.4 and 1 pixel (4
゜D) is (4,B), (4,C), (4,0), (4
,E), (4,F), (133+1
30+118+132+13615=129.8, and even if the threshold value is set to 128, the line will be binarized without interruption.

Conversely, even if there are pixels with high image levels due to noise in the second and fifth stages, by applying the present invention, it is possible to binarize them without connecting two lines.

Embodiment 3 An example in which the line image background erasure and binarization method is applied to a fingerprint image will be shown.

FIG. 11 shows a fingerprint attached to a photograph detected by the laser fluorescence method. When a fingerprint is irradiated with laser light, the fingerprint components emit fluorescence and a fingerprint image can be photographed, but the photograph itself is superimposed as a background image.

When the laser beam irradiation is stopped, only the background image is obtained as shown in FIG. When background removal according to the present invention is performed from FIGS. 11 and 12, only line figures (fingerprint images) can be extracted as shown in FIG. 13. When FIG. 13 is subjected to the binarization process according to the present invention, a good binarized image as shown in FIG. 14 is obtained.

On the other hand, if we binarize Fig. 13 with a uniform threshold, we will get something like Fig. 15, where the central part is collapsed,
The edges are blurred and a good binary image cannot be obtained. Furthermore, even if the present invention is applied immediately to Fig. 11, the direction of the line figure cannot be calculated correctly due to the superimposed background image, resulting in a result as shown in Fig. 16, and a good binary image cannot be obtained. I can't. From these results, the excellent effects of the present invention can be seen.

Embodiment 4 FIG. 17 is a block diagram showing an example of the configuration of a line figure background erasing and binarization processing apparatus suitable for carrying out the fifth aspect of the present invention.

In FIG. 7, 21 is an image input device, 41 is an input switching circuit, and first, an image (image A) in which a background is superimposed on a graphic image is recorded in the first image memory 22. Next, an image containing only the background (image B) is recorded in the second image memory 29. The address generation circuit 26 functions as an incremental counter when performing sequential writing or reading.6 The threshold calculation circuit 37 calculates a threshold for separating a partial image of the background image from the image A. As the threshold value, for example, one calculated by Otsu's discrimination threshold tilt method, which is a well-known threshold setting method when binarizing a grayscale image, can be used.

The calculated threshold value h is sent from the output terminal T to the comparator circuit 38.

The random numbers generated by the random number generation circuit 34 are sent to the address generation circuit 26, which randomly generates addresses for the first image memory 22, and randomly extracts pixels of the image A.

The extracted pixel gradation ga is sent to the comparison circuit 38, and g
If a is smaller than th, it is regarded as a background part and stored in the regression analysis data buffer 39, and the gradation gb data buffer 39 of the corresponding pixel of image B is also sent to 19 and stored.

Such data extraction is performed until the buffer 39 is filled. After the data extraction is completed, the regression analysis circuit 40
Regression analysis is performed.

In this example, a function of the form 3'=p It is sent to the container 36. Then the image memories 22i3 and 2
9, and the values from the memory 29 are corrected by a multiplier 35 and an adder 36.
A subtractor 33 performs subtraction processing to erase the background.

The processed value is recorded in the first image memory 22 via the switching circuit 41.

The background has now been erased. Subsequently, the six direction calculation circuit 27 that performs the binarization process divides the background-erased image stored in the first image memory 22 into local regions, and calculates the direction of the line figure for each region. The calculated direction is recorded in the direction recording memory 30.

The binarization circuit 32 binarizes the grayscale image in the first image memory 22 while reading the calculated direction from the direction recording memory 30, and stores the result in the second image memory 29.

The output switching circuit 23 selects signals from the image input device 21, the first image memory 22, and the second image memory 29,
This signal passes through the D/A converter 24 and is displayed on the monitor 25.

The address generation circuit 26 sends the data of the first image memory 22 to each circuit in order (with appropriate timing) for each process of each circuit described above, and also sends the processed data to the second image memory 29. stored in the first image memory 22
Alternatively, it is for sending data stored in the second image memory 29 to the switching circuit 23.

The timing controller 28 is for sending control signals to each circuit in order to perform the processing of each circuit in an orderly manner, and can be operated in response to a command from the operation switch 31. Note that the results of these operations can be viewed on the monitor 25.

In the background removal method described above, a threshold is set at the image level as a method of separating a partial image of the background image from an image in which the background is superimposed on a line figure image. Of course, it is also possible to take advantage of the difference in texture between the two.

[Effects of the Invention] As explained above, in binarizing a grayscale image of a line figure, after dividing the image into local regions in which the line figure can be regarded as a group of parallel straight lines within that region, each local region is The direction of the line figure is calculated for each direction, the weighted average value centered on each pixel is determined for each local region using the mask pattern for each direction, and the weighted average value is compared with a threshold value to perform binarization. In other words, when performing binarization, we do not use the image level information of just one pixel, but also the image level information of neighboring pixels to determine whether the line figure continues in that direction. By performing the conversion, it is possible to perform binarization processing that is strong against noise, and it is possible to obtain a good binary image even for a low-quality grayscale image.

Further, in the binarization process of the line figure, the background can be removed and the binarization process can be performed, and, for example, fingerprint detection can be properly processed.

The apparatus of the present invention can suitably implement the method having the above effects.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing the procedure of the method of the present invention;
The figure shows an example of dividing an image containing line figures into local regions, Figure 3 shows the direction of quantization in four directions, Figure 4 shows the direction of quantization in eight directions, and Figure 5 shows the direction of quantization in eight directions. The figure shows an example of a mask pattern for calculating the direction of a line figure.
Figure 6 is a diagram showing an example of calculating the direction of the line figure in Figure 2 for each local region, Figure 7 is an explanatory diagram of an image including a region with a large curve, and Figure 8 is centered on each pixel for each direction. FIG. 9 is a block diagram showing an example of a line figure binarization processing device suitable for implementing the present invention, and FIG. 1O is a diagram showing an example of a mask pattern for calculating a weighted average value. Figures 11 to 16 are photographs of fingerprints, Figure 17 is a block diagram of the fifth invention of the present invention, and Figure 18 is a diagram of the third invention of the present invention. Flowchart FIG. 19 is a flowchart of the fourth invention of the present invention. l... Image input device 2... First image memory 3... Output switching circuit 4... D/A converter 5... Monitor 6... Address generation circuit 7... Direction calculation Circuit 8...timing controller 9...second image memory 10...direction recording memory 11...operation switch 12...binarization circuit 21...image input device 22... First image memory 23...Output switching circuit 24...D/A converter 25...Monitor 26...Address generation circuit 27...Direction calculation circuit 28...Timing controller 29... Second image memory 30 -- Direction recording memory 31 -- Operation switch 32 -- Binarization circuit 33 -- Subtractor 34 -- Random number generation circuit 35 -- Multiplier 36 -- Adder 37... Threshold calculation circuit 38 - Comparison circuit 39... Data buffer for regression analysis 40... Regression analysis circuit

Claims (1)

[Scope of Claims] 1. When obtaining a binary image of a line figure image from a digitized grayscale image containing a line figure, (a) the line figure is entirely covered by a local area that can be regarded as a group of parallel straight lines within that area; Divide the image, (b) then calculate the direction of the line figure considered as a straight line group for each local area, and (c) calculate the weight centered on each pixel using a mask pattern determined for each direction. (d) Depending on whether these weighted average values exceed a preset threshold, one of the two values is given to the central pixel of the mask pattern; A method for binarizing line figure pixels, characterized in that this process is performed for all pixels in each local area. 2. The method of calculating the direction of the line figure, which is regarded as a group of straight lines for each local area, is as follows: (a) Using a mask pattern determined for each direction sequentially, and focusing on each pixel, the direction of the line figure is calculated by (b) Calculate the direction of the line figure in the vicinity of the pixel based on the direction of the mask pattern that maximizes the result of the product-sum operation; (c) Calculate the direction of the line figure in the vicinity of the pixel for each local area. 2. The line figure image according to claim 1, wherein the direction of the figure is calculated, the results are totaled for each local area, and the direction with the highest number is determined as the direction of the line figure in the local area. Value processing method. 3. The method of giving binary values based on the threshold values for all pixels for each local area is the highest value when the results of calculating the direction of the line figure in the vicinity are aggregated for each local area. The number of directions is compared with the average value of the number of each direction, and if the difference is not greater than a preset threshold, the local area is determined to be directionless and does not contain a line figure. 3. The method of binarizing a line graphic image according to claim 1, wherein one of the binary values is uniformly given to all pixels in the local area. 4. In claim 2, the results of calculating the direction of the line figure for each pixel are aggregated for each local area, and the maximum number of directions is compared with the average value of the number of each direction. If the difference is not greater than a preset threshold, the local area is further divided into smaller areas, and the calculation results of the line figures for each pixel are aggregated for each of the smaller areas to find the largest number. The number of directions is compared with the average value of the number of each direction, and if the difference is greater than or equal to a preset threshold, the direction with the largest number is set as the direction of the line figure in that small area, and the direction A weighted average value centered on each pixel is calculated using a mask pattern, and one of the two values is determined depending on whether or not these weighted average values exceed a preset threshold. If the difference in the average value of the number of directions applied to the center pixel of the mask pattern becomes less than the threshold value again, that small area is further divided and the same operation is performed on the divided area. A binarization processing method for a line figure image, characterized in that the steps described above are repeatedly performed until the divided regions have a direction or the size of the divided regions becomes equal to or less than a preset threshold value. . 5 an image input device for capturing an image of a line figure; a first image memory for storing a gray scale image of the line figure input to the image input device; A direction calculation circuit that calculates the direction, a direction recording memory that records the calculated direction, and a binary value that reads out the recorded direction and binarizes the grayscale image stored in the first image memory. a second image memory that stores the binarized signal, an output switching circuit that selects the respective output signals of the image input device, the first image memory, and the second image memory; A binarization processing device for a line graphic image, comprising: a D/A converter that converts a selected output signal into an analog signal; and a monitor that displays the analog signal. 6 A grayscale image in which the background is superimposed on the line figure to be extracted (
The background is erased and only the line figures to be extracted are extracted from the grayscale image (hereinafter referred to as image A) and the background only (hereinafter referred to as image B), and the line figures are extracted from the grayscale image containing the extracted line figures. To obtain a binarized image, (a) separate a partial image of the background image from image A, (b) extract a set C of an appropriate number of pixels from the partial image, and (c) extract a set C of pixels from image B. (d) Perform regression analysis on the gradation of corresponding pixels in set C and set D, and from the results, extract the background part of image A and image B.
(e) While correcting image B using the conversion coefficient, image A
By subtracting image B from , an image containing only the line figure to be extracted by erasing the background of the line figure is obtained, and (f) Claims 1 and 2 are added to the grayscale image containing the extracted line figure.
A method for background erasure and binarization of a line graphic image, characterized in that the method for binarizing a line graphic image described in , 3 or 4 is applied. 7. Background erasing of a graphic image according to claim 6, wherein the means for separating a partial image of the background image from the image A is to set a threshold at the image level and set a portion below the threshold as the background. Binarization processing method. 8. The background erasing and binarization processing method for a graphic image according to claim 6, wherein the means for separating the partial image of the background image from the image A is to utilize a difference in texture between the graphic part and the background part. 9 an image input device that captures an image; a first image memory that stores an image (image A) in which a background is superimposed on a line figure image that is input to the image input device; and a background that is then input to the image input device; means for separating a partial image of the background image of image A; means for extracting a set C of an appropriate number of pixels from the partial image; A means for extracting a set D of pixels corresponding to set C, and a regression analysis of the gradations of corresponding pixels in sets C and D, and from the results, the background part of image A and the corresponding part of image B are extracted. A regression analysis circuit that calculates a conversion coefficient between the gradations of a background image subtraction circuit that stores the background image in the image memory; a direction calculation circuit that divides the background-erased grayscale image into local regions and calculates the direction of the line figure for each region; and a direction calculation circuit that records the calculated direction. and a binarization circuit that reads out the recorded direction, binarizes the gray scale image stored in the first image memory or the second image memory, and stores it in the other image memory. and an output switching circuit that selects each output signal of the image input device, the first image memory, and the second image memory, and a D/A converter that converts the selected output signal into an analog signal. , a monitor for displaying the analog signal; and a line figure image background erasing and binarization processing device. 10. A device comprising the line graphic image binarization processing device according to claim 5 and its peripheral equipment, wherein the direction calculation circuit is a circuit that divides into local areas so that the area has a size suitable for a fingerprint image. An image processing device for fingerprint images, characterized by: 11. A circuit comprising the line figure image background erasing and binarization processing device according to claim 9 and its peripheral equipment, and in which the direction calculation circuit divides into local regions so that the region has a size suitable for a fingerprint image. An image processing device for fingerprint images, characterized in that: