JP2000276591A - Method and device for image processing - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method and an image processor which are high in quality and short in processing time, about an image processing method and an image processor. SOLUTION: This processor contains a memory 12 to which an input image is inputted and stored and a CPU 11 which reads image data stored in the memory 12 and performs edge enhancement processing and/or smoothing processing according to the shape of a subject. Thus, this image processing method makes edge enhancement have a directivity and performs a method making the directivity depend on the shape of the subject. It is possible to obtain a high quality image without increasing noise in comparison with the case in which an edge is enhanced without considering the shape.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and an image processing apparatus, and more particularly, to an image processing method and an image processing apparatus in which image processing is changed depending on the shape of a subject.

[0002]

2. Description of the Related Art As a method of giving an image-dependent direction to edge enhancement or smoothing of an image, there is a method of enhancing an edge of a corona. FIG. 6 is an explanatory diagram of a conventional corona edge enhancement process. In (a), 1 is the moon (or sun) in the total solar eclipse, and 2 is the corona. Here, the moon part 1 is dark in the dark part. Here, as shown in (b), the image of (a) is shifted θ from the original position with respect to the center O of the moon.
Only the rotation process is performed. As a result, a difference between components in the tangential direction with respect to the center O of the moon is extracted. By adding the difference of the components in the tangential direction to the original image shown in (a), it is possible to obtain an image in which edges are enhanced as shown in (c).

A sobel filter is a general filter that changes the direction of a mask depending on the edge direction of an image.

[0004]

In the method shown in FIG.
There are many manual processes such as image rotation and reference point setting,
The burden on the operator and the work time were also required.
Further, since the processing is not performed at once, the quantization error also increases.

Further, in the case of the above-mentioned Sobel filter, since a process for giving the same directionality to elements (for example, noise) other than the subject is performed, not only the edge of the subject but also the noise is greatly emphasized, so that image deterioration is caused. Becomes large.

SUMMARY OF THE INVENTION The present invention has been made in view of such problems, and has as its object to provide a high-quality image processing method and an image processing apparatus having a short processing time.

[0007]

A first aspect of the present invention for solving the above-mentioned problems is an image processing method for giving directionality to edge enhancement, wherein the direction of edge enhancement or the direction of edge enhancement depends on the predicted shape of the subject. It is characterized in that the processing method of the strength is changed.

With this configuration, by giving directionality to the edge enhancement processing, it is possible to obtain a high-quality image without increasing noise as compared with a case where the edges are enhanced without considering the shape.

(2) A second aspect of the present invention for solving the above-mentioned problem is an image processing method for giving directionality to smoothing processing,
It is characterized in that the processing method of the direction or intensity of the smoothing process is changed according to the predicted shape of the subject.

With this configuration, by giving directionality to the smoothing process, noise can be efficiently reduced and a high-quality image can be obtained. (3) A third invention for solving the above-mentioned problem is an image processing method in which edge enhancement and smoothing processing are provided with directionality, and a method of processing the direction or intensity of edge enhancement and smoothing processing depending on the shape of a subject. Is changed.

With such a configuration, the quantization error can be reduced by performing the edge enhancement processing and the smoothing processing at the same time, so that a higher quality image can be obtained. (4) In the invention described in any one of claims 1 to 3, the shape of the subject is concentric and is set in a radiation direction from the center of the circle.

According to this structure, since the directionality is concentric and radial from the center of the circle, the quantization error can be reduced for an image whose processing direction is concentric and close to radial. The processing can be speeded up.

(5) The invention according to claim 4, characterized by including a method of extracting the center of the concentric circle.
With such a configuration, by including the method of extracting the center, the processing can be automated and the load on the operator can be reduced.

(6) The invention according to claim 5, wherein the extraction method includes a discrimination based on an edge. With this configuration, when the reference target has a feature in the edge, the extraction rate can be increased.

(7) In the invention described in claim 5, the extraction method includes a distinction based on brightness. With such a configuration, the extraction rate can be increased for a reference object having a characteristic of light and dark.

(8) In the invention described in any one of claims 1 to 7, a portion not requiring image processing is extracted from the shape, and the image processing is not performed on the portion.

According to this structure, a portion that does not require image processing is extracted, and image processing is not performed on the portion. Can be achieved.

(9) In the invention described in claim 8, the extraction method includes a discrimination based on an edge. With such a configuration, a portion that does not require image processing is extracted by using the discrimination based on the edge, so that the calculation time can be reduced and the processing can be speeded up.

(10) In the invention described in claim 8, the extraction method includes a distinction based on brightness. With this configuration, a portion that does not require image processing is extracted by using the distinction based on brightness, so that the calculation time can be reduced and the processing can be speeded up.

(11) The invention according to claims 1 to 3, wherein the shape is a shape peculiar to the corona of a solar eclipse. With such a configuration, the corona of the linear eclipse can be particularly efficiently emphasized by setting the edge of the subject as the corona of a solar eclipse.

(12) In the invention according to claim 11,
In the specific shape of the total solar eclipse, the center of the moon is extracted, and edge enhancement is performed only in the tangential direction of a circle centered on the center.

With this configuration, the center of the moon is extracted, and the edge is emphasized only in the tangential direction of a circle centered on the center of the moon, so that the corona of the sun can be enhanced and a high quality image can be obtained. Can be.

(13) In the eleventh aspect,
In the shape unique to the total solar eclipse, extract the center of the moon,
It is characterized in that the smoothing process is performed only in the normal direction of a circle centered on the center.

With this configuration, the noise contained in the corona can be efficiently reduced by extracting the center of the moon and performing the smoothing process only in the normal direction of the circle centered on the center. .

(14) In the invention according to claim 11,
In the shape unique to the total solar eclipse, extract the center of the moon,
It is characterized in that edge enhancement processing is performed in the tangential direction of a circle centered on the center, and smoothing processing is performed in the normal direction of the circle.

According to this structure, the quantization error can be reduced by performing edge enhancement in the tangential direction of the circle and smoothing processing in the normal direction of the circle to obtain a higher quality image. Can be.

(15) The invention according to claims 12 to 14, wherein the moon portion is determined by a method including edge extraction or determination based on brightness, and a reference point for processing is calculated.

With this configuration, it is possible to calculate a reference point for processing by edge extraction or discrimination based on brightness. (16) The invention according to claim 15, wherein the discriminating method includes a process of performing nothing on the lunar surface portion.

With this configuration, no processing is performed on the lunar surface portion, so that the processing speed can be increased. (17) A fourth invention for solving the above-mentioned problem is a memory for inputting and storing an input image, and reading out image data stored in the memory, and performing edge enhancement processing and / or smoothing processing according to the shape of a subject. And a CPU for performing the following.

With this configuration, all of the methods described in claims 1 to 16 can be realized, input image data can be processed as it is, and high quality processing can be performed.

[0031]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the apparatus of the present invention. In the figure, 1 is an image input unit for inputting image data, 2 is the image input unit 1
C that performs various image processing by receiving image data from 0
PU and 12 are memories for temporarily storing an image input from the image input unit 10 and for storing various information.

An output unit 13 outputs image data processed by the CPU. For example, a CRT, a plasma display, a liquid crystal display, a color printer, or the like is used. A user operation unit 14 for inputting various operation commands and the like to the CPU 11 uses, for example, a mouse, a joystick, and a keyboard. Reference numeral 15 denotes a display for user operation, for example, a CRT, a liquid crystal display, a plasma display, or the like. The operator inputs various control commands according to the display on the user operation display unit 15. When a CRT or the like is used as the image output unit 13, the user operation display unit 15 can be shared with the image output unit 13.

The method of the present invention will be described below by using the thus configured apparatus. Image data having a certain shape input from the image input unit 10 is temporarily stored in the memory 12 via the CPU 11. CP from user operation unit 14
When a processing request instruction is issued to U11, the CPU 11
Reads out the stored image data from the memory 12 as necessary, and changes the processing method of the edge emphasis direction or the intensity according to, for example, the shape of the subject. Similarly, the direction or intensity of the smoothing process is changed according to the shape of the subject. Furthermore, the processing method of the direction or intensity of the edge enhancement and smoothing processing is changed according to the shape of the subject. The image processed in this way is output from the image output unit 13.
The output mode is display on a CRT, which is the image output unit 13, or printing on a printer.

According to the present invention, by giving directionality to the edge enhancement processing, it is possible to obtain a high-quality image without increasing noise as compared with the case where the edges are enhanced without considering the shape. Further, according to the present invention, by giving directionality to the smoothing processing, noise can be efficiently reduced and a high-quality image can be obtained.

Further, by simultaneously performing the edge emphasis processing and the smoothing processing, the quantization error can be reduced, and a higher quality image can be obtained. In addition, by using an object that is concentric and has a radial direction from the center of the circle, the quantization error can be reduced for an image in which the processing direction is concentric and almost radial, and the processing speed is increased. Can be achieved.

In this case, the CPU 11 can extract the center of the concentric circle. According to this, by extracting the center, the processing can be automated and the load on the operator can be reduced.

The method of extracting the center of the concentric circle can be realized by performing discrimination based on edges. According to this, when the reference target has an edge, the extraction rate can be increased. Similarly, a method of extracting the center of the concentric circle can be realized by performing determination based on brightness.
According to this, it is possible to increase the extraction rate for a reference object having a characteristic of light and dark.

In the above-described image processing, the CPU 11 can extract a portion that does not require image processing. According to this, by not performing the image processing on the portion that does not require the image processing, unnecessary processing need not be performed, and the processing speed can be increased.
In this case, the edge-based determination can be used for the extraction processing of the part that does not require the processing. According to this, a portion that does not require image processing is extracted by using the discrimination based on the edge, so that the calculation time can be saved and the processing can be speeded up. The distinction based on brightness can be used for the extraction processing of the part that does not require image processing. According to this, a portion that does not require image processing is extracted by using the distinction based on brightness, so that the calculation time can be saved and the processing can be speeded up.

Here, a subject image of a total solar eclipse can be used as the subject image input from the image input unit 10. According to this, the corona of the linear eclipse can be particularly efficiently emphasized by using the corona of the solar eclipse as the edge of the subject.

FIG. 2 is a diagram showing an embodiment of the method of the present invention. Here, a case where the subject is a corona of a solar eclipse is taken as an example. The center ● is the moon (sun),
Coronas 20 and 21 are spread around it. Corona 20 is a particularly strong corona that occurs around month 1,
The corona 21 is a weak corona generated on the outer periphery of the corona 20. An example in which this corona is emphasized will be described.

Since it is not necessary to emphasize the edge of the reference month portion, the same value as the input value is used as the output value for this portion. In this embodiment, first, a month portion is extracted to obtain a center, and this is used as a reference. Next, edge enhancement is performed only on a tangential direction of a circle centered on a reference point with respect to a region other than the moon, Smoothing processing is performed only in the normal direction. In the following processing, processing by the CPU 11 is all performed.

(A) Extraction of month and its center This will be described with reference to FIG. In general, the edge of the moon during a total solar eclipse shows a corona, and the brightness changes more sharply than the other parts. In an image whose gradation is not inverted, month 1 appears dark. Month 1 can be considered as a yen. These are used to extract the month.

A horizontal line of an image having two changes in brightness exceeding a threshold value is extracted. FIG. 2 shows a change in brightness for each line in the horizontal direction and a change in brightness for each line in the vertical direction. As is clear from the figure, the area corresponding to the lunar surface is darkened with reduced brightness. Points A and B in the figure are two points where there is a sharp change in brightness. The above processing is also performed in the vertical direction.

If the portion on the left side of the image is suddenly dark and the right side is suddenly bright, the two points A and B are regarded as candidates for a lunar edge, and the coordinates of the coordinates are averaged. Let the value be a candidate for the x coordinate of the month.

X-coordinate candidates are extracted from all the horizontal lines, and the mode of the candidates is regarded as the x-coordinate. The same calculation is performed on the vertical line of the image to obtain the y coordinate. The x-coordinate and the y-coordinate obtained by the above procedure are used as the center coordinates of the moon, and the diameter of the moon is obtained from the difference between the coordinates of the candidate for the edge of the moon obtained by the vertical and horizontal lines on the coordinates.

(B) Edge enhancement and smoothing processing The edge enhancement processing and smoothing processing will be described with reference to FIG. Edge enhancement processing is performed only in the tangential direction of the circle centered on the reference point (the center point of the moon), and smoothing processing is performed only in the normal direction. Regarding the coordinates of the image, the point Z at the upper left is set as the origin (0, 0), and the coordinates of the center point of the month obtained in (a) are set as (a, b). And the lower right is the number of pixels of the horizontal line,
This is the number of pixels in the vertical line. The parameters determined by the user in advance are i and j, and the coordinates of the pixel of interest are (x, y)
And

Assuming that the slope of a straight line passing through the reference point (a, b) and the pixel of interest is t, t is expressed by the following equation. t = (by) / (ax) (1) The slope is on the straight line of t passing through the point (x, y), and the absolute value of the distance from the point (x, y) is i. Pixels (or pixel interpolation) are c1 and c2.

Similarly, a pixel (or a pixel interpolation) passing through a point (x, y), on a straight line having a slope of 1 / t, and having an absolute value of a distance from (x, y) being i is defined as d1 and d2. The output value of the target pixel is obtained by the following equation.

The output value of the target pixel in the case of only edge enhancement is expressed by the following equation. Output value of target pixel = input value of target pixel + j (input value of target pixel × pixel value of 2-d1−pixel value of d2) (2) where j is an enhancement coefficient. According to this equation, (−d
The pixel value of 1−the pixel value of d2) is a portion indicating the edge enhancement operation. According to this, by extracting the center of the moon and performing edge enhancement processing only in the tangential direction of a circle centered on the center of the moon, the corona of the sun can be enhanced, and a visually high-quality image can be obtained.

Next, the output value of the target pixel in the case of only the smoothing process is represented by the following equation. Output value of target pixel = (1−j) × input value of target pixel + j (pixel value of c1 + pixel value of c2 + input value of target pixel) / 3 (3) Here, it is assumed that 0 ≦ j ≦ 1. In this equation, (pixel value of c1 +
The portion of (pixel value of c2 + input value of target pixel) is a portion indicating the smoothing process. According to this, the center of the moon is extracted and the smoothing process is performed only in the normal direction of the circle centered on the center of the moon, so that the noise included in the corona can be efficiently reduced.

The output value of the target pixel when the edge enhancement processing and the smoothing processing are combined is expressed by the following equation. Output value of target pixel = Input value of target pixel + j (pixel value of c1 + pixel value of c2−pixel value of d1−pixel value of d2) (4) Here, j is used as an enhancement coefficient. (C1 +
c2) indicates a smoothing process, and (-d1-d2) indicates an edge enhancement process.

However, in the black circle indicating the moon, the output value of the target pixel = the input value of the target pixel.

Further, by performing edge enhancement in the tangential direction of the circle and smoothing processing in the normal direction of the circle, the quantization error can be reduced and a higher quality image can be obtained.
Further, a reference point for processing can be calculated by edge extraction or discrimination based on brightness. In addition, no processing is performed on the moon portion, so that the processing can be speeded up.

Further, according to the configuration of the invention shown in FIG. 1, all of the above-described processing methods can be realized, input image data can be processed as it is, and high-quality processing can be performed.

FIG. 4 is an explanatory diagram of the effect of the present invention.
(A) shows the state of the corona before the image processing, and (b) shows the state of the corona after the image processing after the edge enhancement and the smoothing processing. It can be seen that the corona flare (flame) can be observed well after the image processing.

FIG. 5 is a diagram showing an example of image processing in which image processing is applied to an eye as a subject. (A) shows the state of the eye, and (b) shows the state of the eye after image processing. 20 is a pupil, 21 is an iris, 22 is a white eye, and 23 is skin. ADVANTAGE OF THE INVENTION According to this invention, the pattern of an iris, the edge of each boundary (pupil-iris, iris-white-eye, white-eye-skin) are emphasized, and the smoothing process of each area (pupil-white-eye, skin) is efficiently performed. Can be.

In the above-described embodiment, the case where the reference point is automatically calculated has been described. However, an interface that can be manually adjusted by an operator may be provided. Further, in the embodiment, the parameters of the edge enhancement and the smoothing process are shared as i, but each may have a parameter independently.

In the above-described embodiment, the mask range (parameter i) for edge enhancement and smoothing processing is constant irrespective of the distance between the target pixel and the reference point, but may be a variable. Similarly, the filter strength (parameter j) is fixed, but may be a variable. Further, in the embodiment, the number of peripheral pixels of the target pixel is set to four, but more peripheral pixels may be used for higher image quality.

In the above-described embodiment, the filter is concentrically formed around the reference point. However, the filter may be changed in accordance with the characteristics of the shape of the subject. For example, it is conceivable that in a corona, the intensity is increased with a small mask near the pole and weakened with a large mask near the equator.

Further, in the above-described embodiment, the case where image processing is performed on a corona photograph is taken as an example. However, by changing the extraction method or providing an interface that can be manually operated, the comet A filter which is applied to a subject other than the corona, such as a frame portion and a tail portion, may be used.

[0061]

According to the present invention described above, the following effects can be obtained. (1) According to the first aspect, there is provided an image processing method for giving directionality to edge enhancement, wherein a method of making the directionality dependent on the shape of a subject is performed. This makes it possible to obtain a high-quality image without increasing noise as compared with the case where the edge is emphasized without considering the shape.

(2) According to the second aspect, similarly, by giving directionality to the smoothing processing, noise can be efficiently reduced and a high quality image can be obtained. (3) According to the third aspect, by simultaneously performing the edge enhancement processing and the smoothing processing, a quantization error can be reduced, and a higher quality image can be obtained.

(4) According to the fourth aspect, the directionality is concentric and is set in the radial direction from the center of the circle, so that the processing directionality is concentric and the image is closer to the radial direction from the center. Errors can be reduced and the processing speed can be increased.

(5) According to the fifth aspect, by extracting the center of the object shape, automation can be achieved, and the load on the operator can be reduced. (6) According to the sixth aspect, by using the discrimination based on the edge as the extraction method, the extraction rate is increased when the reference target has an edge.

(7) According to the seventh aspect, by using the distinction between light and dark as the extraction method, the extraction rate is increased when the reference object has light and dark characteristics. (8) According to the eighth aspect, a portion that does not require image processing is extracted, and the image processing is not performed on the portion.
This eliminates the need for unnecessary processing,
The processing can be speeded up.

(9) According to the ninth aspect, a portion that does not require image processing is extracted by using discrimination based on edges, so that calculation time can be saved and processing can be speeded up. (10) According to the tenth aspect, since a portion that does not require image processing is extracted using the distinction based on brightness, calculation time can be reduced and processing can be speeded up.

(11) According to the eleventh aspect, the corona of the linear shape is particularly efficiently emphasized by setting the edge of the subject as a corona of a solar eclipse. (12) According to the twelfth aspect, by extracting the center of the moon (sun) and applying edge enhancement only to the tangential direction of a circle centered on the center, the corona of the sun can be enhanced and the image quality is high in appearance. An image is obtained.

(13) According to the thirteenth aspect, by performing the smoothing process only in the normal direction of the circle, the noise included in the corona can be reduced efficiently. (14) According to the fourteenth aspect, edge enhancement in the tangential direction of the circle and smoothing processing in the normal direction of the circle are performed. As a result, the quantization error can be reduced, and a higher quality image can be obtained.

(15) According to claim 15, the moon portion is determined by a method including edge extraction or determination based on brightness,
Extract the processing reference point. Thereby, the reference point of the processing can be calculated by edge extraction or discrimination based on brightness.

(16) According to the sixteenth aspect, since no processing is performed on the lunar surface portion, the processing can be speeded up. (17) According to the seventeenth aspect, all of the image processing methods according to the present invention can be realized, and input image data can be processed as it is, and high-quality processing can be performed.

As described above, according to the present invention, it is possible to provide an image processing method and an image processing apparatus with high quality and short processing time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the apparatus of the present invention.

FIG. 2 is a diagram showing an embodiment of the method of the present invention.

FIG. 3 is an explanatory diagram of an edge enhancement process and a smoothing process.

FIG. 4 is an explanatory diagram of an effect of the present invention.

FIG. 5 is a diagram showing an example in which the present invention is applied to an eye.

FIG. 6 is an explanatory diagram of a conventional edge enhancement process.

[Explanation of symbols]

 Reference Signs List 10 Image input unit 11 CPU 12 Memory 13 Image output unit 14 User operation unit 15 User operation display unit

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Shigeto Hirabayashi 1 Sakuracho, Hino-shi, Tokyo Konica Corporation F-term (reference) 5B057 AA20 CA02 CA08 CA12 CA16 CB02 CB08 CB12 CB16 CC03 CE03 CE05 DC09 5C077 LL19 PP02 PP03

Claims (17)

[Claims] 1. An image processing method for giving directionality to edge enhancement, wherein the processing method of the direction or intensity of edge enhancement is changed according to the predicted shape of a subject. 2. An image processing method for providing directionality to a smoothing process, wherein the direction or intensity of the smoothing process is changed according to the predicted shape of a subject. 3. An image processing method having directionality in edge enhancement and smoothing processing, wherein the processing method of the direction or intensity of edge enhancement and smoothing processing is changed according to the shape of a subject. Method. 4. The image processing method according to claim 1, wherein the shape of the subject is concentric and is set in a radial direction from the center of the circle. 5. The image processing method according to claim 4, further comprising a method of extracting a center of the concentric circle. 6. The image processing method according to claim 5, wherein said extraction method includes a discrimination based on an edge. 7. The image processing method according to claim 5, wherein the extraction method includes a distinction based on brightness. 8. The image processing method according to claim 1, wherein a portion that does not require image processing is extracted from the shape, and the image processing is not performed on the portion. 9. The image processing method according to claim 8, wherein the extraction method includes a discrimination based on an edge. 10. The image processing method according to claim 8, wherein said extraction method includes a distinction based on brightness. 11. The image processing method according to claim 1, wherein the shape is a shape unique to a corona of a total solar eclipse. 12. In the specific shape of the total solar eclipse,
12. The method according to claim 11, wherein a center of the moon is extracted, and edge enhancement processing is performed only in a tangential direction of a circle centered on the center.
The image processing method described in the above. 13. The image processing method according to claim 11, wherein a center of the moon is extracted in a shape unique to the total solar eclipse, and a smoothing process is performed only in a normal direction of a circle centered on the center. Method. 14. Extracting the center of the moon in the shape unique to the total solar eclipse, performing edge enhancement processing in a tangential direction of a circle centered on the center, and performing smoothing processing in a normal direction of the circle. The image processing method according to claim 11, wherein: 15. The image processing method according to claim 12, wherein the moon portion is determined by a method including edge extraction or determination based on brightness, and a reference point for processing is calculated. . 16. The image processing method according to claim 15, further comprising a process of performing no operation on the lunar surface portion in said determination method. 17. A memory for inputting and storing an input image, and a CPU for reading image data stored in the memory and performing edge enhancement processing and / or smoothing processing according to the shape of a subject. Image processing apparatus.