JP3441887B2 - Image processing method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image processing for reducing noise without causing image blurring or density unevenness.

[0002]

2. Description of the Related Art As inventions for reducing noise, there are (1) Japanese Patent Application No. 6-305888 and (2) Japanese Patent Application No. 7-13479 filed previously by the present applicant.

In the above-mentioned prior art (1), a brightness change vector in which the magnitude of the brightness change of the image is evaluated for each of a plurality of predetermined directions at each point of the image is created, and the component value of the brightness change vector is minimized. The minimum change direction is obtained for each point of the image, and one-dimensional nonlinear smoothing is performed in the minimum change direction. The above procedure is illustrated in FIG. In the procedure shown in FIG. 2, a brightness change vector map representing the brightness change vector of each point of the image is created in step 101, a minimum change direction map is created from the brightness change vector of each point in step 113, and in step 104. 1 for the smallest change direction
Dimension smoothing was performed.

In the above prior art (2), a blurred image in which the input image is blurred is created, the minimum change direction in which the component value of the brightness change vector of the blurred image is minimized, and the input image is detected in the minimum change direction of the blurred image. The one-dimensional nonlinear smoothing is performed. Since the minimum change direction of the blurred image is detected, in a noisy image, it is often the case that the minimum change direction is found as the original structure of the image when ignoring the part that is contaminated as noise at the gentle edge part of the image. This has the effect of making the edges of the smoothed image look nice.

[0005]

In the prior art (2), the detection accuracy of the minimum change direction is related to how much noise can be reduced. However, in the related art (2), increasing the detection accuracy is not considered so much.

Further, although not only the problems of the prior arts (1) and (2), there is also a problem that moire occurs in the flat portion of the image.

Further, there is a problem that the image processing may excessively smooth the image. If the image is smoothed too much, the difference in shade of the image will not be so great that it will be difficult to distinguish each part.

[0008]

In order to solve the above problems, the following constitution is adopted. In particular, in order to solve the first problem described above, the luminance change vector of the input image is blurred instead of making a blurred image of the input image. That is, vector addition with a predetermined weight is performed between adjacent brightness change vectors to create a blurred brightness change vector, and the direction in which the component value is the minimum is the minimum change direction. Blurring the brightness change vector makes it easier to reflect the original structure of the image than blurring the image, and the maximum weight to be added is effective with a smaller value than when blurring the image, and it is possible to capture even minute structures. Will increase.

In order to solve the above-mentioned second problem, not only one-dimensional smoothing in the direction of minimum change but also one-dimensional smoothing in a direction deviated by 45 degrees from the direction of minimum change were performed. At the edge portion of the image, smoothed data in a direction deviated from the minimum change direction by 45 impairs the edge structure of the image. Therefore, the amount of brightness change in the vicinity at each point of the image is calculated, and that point is the edge of the image. The edge degree is determined by determining whether it is in a flat position or in a flat position, and the data smoothed in the direction of minimum change and the data smoothed in a direction deviated from the direction of minimum change by 45 degrees according to the degree of edge. The weighted average of
When the degree of edge is strong, the proportion of data smoothed in the minimum change direction is large, and when the degree of edge is small, weighted averaging is performed so as to increase the proportion of data smoothed in the direction shifted by 45 degrees. As described above, the edge portion uses the data that gives a smooth impression smoothed in the minimum change direction, and the flat portion uses the value that includes the data smoothed in the direction shifted by 45 degrees without moire, which reduces moire. .

Further, in order to solve the above-mentioned third problem, a weighted average of the output image and the input image after the operation in which the means for solving the second problem of the present invention is performed is performed. By adding the input image, it is possible to correct the output image that has been subjected to the means for solving the second problem of the present invention, when the output image is excessively smoothed.

[0011]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. Although the embodiment has been described as a processing procedure of image processing in an apparatus having a function of performing image processing, it may be configured by either software or hardware as long as this processing is performed.

First, the first embodiment will be described. The first embodiment is particularly for increasing the detection accuracy in the minimum change direction.

This procedure will be described with reference to FIG. First,
In step 101, a brightness change vector map is created. Specifically, at each point of the input image, a value that evaluates the magnitude of the brightness change of the input image is calculated for each of a plurality of predetermined directions, and the brightness change value in each direction obtained by the calculation is a vector component. A brightness change vector as a value is created for each point of the input image to create a brightness change vector map.

Next, in step 102, a predetermined brightness change vector is blurred. Specifically, at each point of the input image, the brightness change vectors in the vicinity at a predetermined distance are vector-added with a predetermined weight to create a blurred brightness change vector to create a blurred brightness change vector map.

Next, in step 103, a minimum direction map is created. At each point of the input image, the direction in which the component value of the blurred brightness change vector becomes the minimum is obtained, and the minimum change direction map in which the minimum change direction at each point of the image is indicated by a predetermined number is created.

Finally, in step 104, one-dimensional nonlinear smoothing is performed on each point of the input image in the minimum change direction obtained in step 103.

A characteristic step in this embodiment is step 102. Further, in FIG. 2, the minimum change direction map is created from the brightness change vector in step 113, but in FIG. 3 of the present invention, the minimum change direction map is created from the blurred brightness change vector in step 103. In the prior art (2), the image is blurred,
The present invention is different in that a blurring operation is performed on the brightness change vector. Due to these differences, the detection accuracy can be further increased in this embodiment.

Next, details of the processing of each step will be described. First, step 101 will be described. There are various calculation formulas for obtaining the brightness change vector, but two typical methods are shown.

[0019]

[Equation 1]

[0020]

[Equation 2]

However, I [j] [i] means the luminance value at the point of the i-th row and the j-th column of the input image, and dir (I [j] [i], d, k) is the i-th row of the input image. It means the brightness value at the point where k sampling points are performed in the d direction from the point in the j-th column, abs means that it takes an absolute value, and Kmax means the length for evaluating the brightness change. S [d] [j] [i] means the value of the luminance change in the d direction at the point on the i-th row and the j-th column, and has shown two formulas, Formula 1 and Formula 2. Whichever calculation is used, there is no great difference in the image quality of the output image finally obtained.

There are various directions for calculating the brightness change, such as 4 directions (up / down, left / right, diagonally right, diagonally left), 8 directions, and 16 directions, and 9 points are taken in each direction in FIGS. 5 and 6. The case (4 points each from the center, kmax = 4) is illustrated. FIG. 5 is a diagram in the case of interpolating a point that does not exist in the pixel of the image, and FIG. 6 shows a case in which the interpolating point is substituted by the points closest to the top, bottom, left, and right.

When the directions for calculating the brightness change are eight directions, the number d indicating the direction is represented by an integer value of 1 to 8, for example. In this case, at the point on the i-th row and the j-th column, S [d]
Since there are eight types of d in [j] [i], it can be regarded as a vector having eight components. In this way, at the point on the i-th row and the j-th column, a vector whose component value is the brightness change value in each direction is i
It is called a luminance change vector in row j column. In addition, the brightness change vector obtained at all points of the image will be referred to as a brightness change vector map. Next, a specific method of step 102 will be described. There are many specific examples of creating a blurred brightness change vector by performing a weighted vector sum of the brightness change vectors. The general form is described in the following Equation 3.

[0024]

[Equation 3]

However, S [d] [j + jj] [i + ii] is i + ii line j + j
The component value in the d direction of the luminance change vector at the point on the j-th column is shown, and S2 [d] [j] [i] indicates the component value in the d direction on the blurred luminance change vector at the point on the i-th row and the j-th column. im in line, jm
Indicates the maximum blurring range of the row. w [jj] [ii] is the maximum blurring weight, and there are various concrete numerical values. For example, im = jm =
When 1, w [0] [0] can be 1/2, and other w [jj] [ii] can be 1/16.

Next, a mathematical expression showing the operation of step 103 will be described. Equations are used to find the direction in which the component of the blurred luminance change vector has the minimum value.

[0027]

[Equation 4]

However, the direction is expressed from 1 from the numerical value.
Assuming there is a dn direction, mindir (S2 [d] [j] [i] | 1 <= d <= dn) is
Each component (S2 [1] of the blurred brightness change vector S2 [d] [j] [i]
[j] [i] to S2 [dn] [j] [i]), which means to find the direction with the smallest value. When there are two or more directions having the minimum value, the direction having a smaller numerical value when the direction is expressed as a number may be used, or conversely, the direction having a larger numerical value may be used. dmin [j] [i] is a numerical expression of the minimum change direction in the i-th row and the j-th column obtained as described above. The minimum change direction map is obtained by obtaining the minimum change direction at all points of the image.

Next, a specific method of step 104 will be described. There are various specific methods for performing one-dimensional nonlinear smoothing in the direction of minimum change. An example of this is the following number 5
Shown in Equation 8.

[0030]

[Equation 5]

[0031]

[Equation 6]

[0032]

[Equation 7]

[0033]

[Equation 8]

However, I [j] [i] means the luminance value at the point of the i-th row and the j-th column of the input image, and dir (I [j] [i], dmin [j] [i],
k) means the luminance value at the point where k sampling points are performed in the direction dmin [j] [i] which is the minimum change direction obtained in step 103 from the point of the i-th row and the j-th column of the input image, and Kmax2 is smoothed. Hmin [j] [i] means the output value after smoothing at the point of row i and column j, and σ0 is an amount proportional to the standard deviation of noise in the input image. Means value,
α is a predetermined value that determines the degree of smoothing. Note that σ0
There are various ways to calculate, for example, the square root of the root mean square of the brightness of a small area at the edge of the image that can be regarded as a noise portion may be used, or the value of the smallest displacement among neighboring points at each image point may be used. You may use an average value.

The following equations 9 to 14 are shown as other examples of performing one-dimensional nonlinear smoothing in the direction of minimum change.

[0036]

[Equation 9]

[0037]   w [k] [j] [i] = u [k] [j] [i] / ut [k] [j] [i] (Equation 10)

[0038]

[Equation 11]

Here, when k = 0 (Equation 12), when k> = 1 (Equation 13), the time blade (Equation 14) when k <=-1 is obtained.

U [0] [j] [i] = 1 ... (Equation 12)

[0041]

[Equation 13]

[0042]

[Equation 14]

However, I [j] [i] means the luminance value at the point of row i, column j of the input image, and dir (I [j] [i], dmin [j] [i],
k) means the luminance value at the point where k sampling points are performed in the direction dmin [j] [i] which is the minimum change direction obtained in step 103 from the point of the i-th row and the j-th column of the input image, and Kmax2 is smoothed. Hmin [j] [i] means the output value after smoothing at the point of row i and column j, and σ0 is an amount proportional to the standard deviation of noise in the input image. Means value,
α is a predetermined value that determines the degree of smoothing, and β is a predetermined parameter value that determines the contribution of smoothing.

In addition to the above, there are various methods for performing one-dimensional nonlinear smoothing in the direction of minimum change. For example, 1 of a predetermined score
You can also select an intermediate value from the dimensional data. Also,
In the case where the input data is a binary image of 0 and 1 and the output also takes 1 as 0, one of the one-dimensional data having a larger number can be selected.

There are various procedures for obtaining the same result as the procedure shown in FIG. For example, in step 102, a blurred brightness change vector is obtained for each point and a map of the blurred brightness change vector is created. In step 102, the operation of step 103 is also performed at the same time to obtain the blurred brightness change vector. It is also possible to obtain the direction in which the component value becomes the minimum value, create a minimum change direction map, and eliminate the need for the blurred brightness change vector map. Similarly, step 10
The non-linear smoothing performed in step 4 may be performed immediately after the minimum change direction is obtained in step 103, and the minimum change direction map may be unnecessary. Further, in step 104, the non-linear smoothing is performed, but the smoothing can be performed by simply taking the one-dimensional sum in the minimum change direction.

Next, a second embodiment of the present invention will be described. The second embodiment is mainly focused on reducing moire generated in a flat portion of an image. The procedure of FIG. 4 is shown. Hereinafter, the numbers in FIG. 4 will be described as step numbers. Since steps 101 to 104 are the same as those in the first embodiment (contents shown in FIG. 3),
The description is omitted.

In step 105, one-dimensional non-linear smoothing is performed at each point of the input image in a direction deviated by a predetermined angle from each minimum change direction obtained in step 103.

Next, in step 106, the amount of change in brightness near each point is calculated at each point of the input image, and the minimum change obtained in step 104 is calculated according to the value of the amount of change in brightness. The weighted average of the values smoothed in the direction, the minimum change direction obtained in step 105, and the values smoothed in the direction shifted by a predetermined angle is changed.

Here, a specific method of step 105 will be described. The calculation for obtaining the direction that is deviated by a predetermined angle from the minimum change direction obtained in step 103 can be performed as follows. However, the direction is from 1 to dn, and the next number means the adjacent direction, and the dn direction and the 1 direction are adjacent to each other. Further, the direction deviated by a predetermined angle is assumed to be the direction in which the dshift number is deviated. At this time, dmin [j] [i], which is the direction in which the dshift number is deviated from dmin [j] [i], can be calculated by the following formula 15.

[0050]   dminshift [j] [i] = (dmin [j] [i] + dshift-1 + dn)% dn + 1 ... (Equation 15) However, A% B means the remainder when A is divided by B.

Various settings can be made for dshift, which is a numerical value indicating the shift angle. For example, when dn is 8 and dshift is smoothed in the direction of 1, the moire in the flat part of the image remains slightly, and ds
When hift is 2 (that is, when the shift is 45 degrees), the moire in the flat portion of the image disappears completely. As the dshift becomes 3 and 4, the blurring of the edge part of the image becomes severe. Therefore, it is considered most appropriate to smooth in the direction shifted by 45 degrees.
There is no extreme angle dependence, and good results can be obtained at angles other than 45 degrees. In addition, smoothing is performed in step 10.
The one used in 4 may be used, or another one may be used.

Next, a specific method of step 106 will be described. There are various specific calculations of the amount of change in brightness near each point of the input image. For example, the deviation value may be obtained by using 9 points of its own point and 8 points in the vicinity thereof, or the sum of the components of the luminance change vector, that is, the value calculated by the following formula 16, or the blurred luminance change vector It may be the sum of ingredients.

[0053]

[Equation 16]

However, S [d] [j] [i] is the component value in the d direction of the luminance change vector at the point of row i and column j, σt [j] [i]
Is an amount that captures a change in luminance in the vicinity of the i-th row and the j-th column of the image.

As described above, there are various concrete methods for calculating the amount of change in brightness near each point of the input image and performing weighted averaging according to the calculated value. For example, the following number 17-
There is a method of calculation as shown in Equation 18.

[0056]

[Equation 17]

[0057]

[Equation 18]

However, Hmin [j] [i] means a value smoothed in the minimum change direction, and Hmv [j] [i] means a value smoothed in a direction deviated from the minimum change direction by a predetermined angle. , Σ0 means a value calculated by an amount proportional to the standard deviation of the image noise,
σt [j] [i] means the amount of luminance change near the point of the i-th row and the j-th column of the image, H [j] [i] is the output value, and γ and f0 are predetermined parameter values. is there.

There are various procedures for obtaining the same result as the procedure shown in FIG. For example, immediately after performing the non-linear smoothing in the minimum change direction in step 104, the non-linear smoothing deviated from the minimum direction by a predetermined angle is performed, and the amount of the brightness change near each point of the image is calculated, and according to the value. It is also possible to carry out weighted averaging by storing the above Hmin [j] [i] and Hmv [j] [i]. In addition, in step 105, the non-linear smoothing is performed, but it is also possible to perform the smoothing by simply taking the one-dimensional sum in the minimum change direction.

Finally, a third embodiment of the present invention will be described. The third embodiment is particularly focused on solving the problem of excessive smoothing.

FIG. 1 is a diagram showing a procedure for solving the third problem of the present invention, and hereinafter, the numbers in FIG. 1 will be described as step numbers. Note that steps 101 to 106 are the same as the description of the second embodiment, so a description thereof will be omitted.

In step 107, the result image obtained in step 106 and the input image are weighted and averaged.

Next, a specific method of step 107 will be described. There are various ways to obtain the weight of the weighted average. For example, the result image of step 106 is set to 80% and the input image is set to 20%, or the result image of step 106 is set to 9%.
The input image can be set to 0% and the input image can be set to 10%, or can be set to an appropriate value according to preference. Further, the weighted average of both may be applied to the entire image with the same weight, or the weighted average may be performed based on an appropriate statistic as in step 106.

In the second embodiment, in step 105, smoothing is performed in the direction that deviates from the minimum change direction by a predetermined angle, but data smoothed in another direction that deviates from the angle may be mixed.
For example, if the data smoothed in the direction shifted by 45 degrees and -45
The average of the data smoothed in the direction shifted by 45 degrees may be used, or the weighted average of the data smoothed in the direction shifted by 45 degrees and the data smoothed in the direction shifted by 5 degrees may be performed. .

Further, a blurred brightness change vector is created in step 102, and the direction in which the component value of the blurred brightness change vector is the minimum is the minimum change direction. Instead, the prior art (2) is used. It is also possible to blur the input image and use the direction in which the component of the brightness change vector of the blurred image has the minimum value as the minimum change direction to perform the processing after step 104.

[0066]

According to the first embodiment, when it becomes difficult to detect the minimum change direction of the original structure of the image due to noise, the accuracy of detecting the minimum change direction of the original structure of the image is improved, and There is an effect that even a fine structure can be detected.

Further, the second embodiment has an effect of reducing the moire of the image flat portion which occurs when the first embodiment of the present invention is carried out.

Further, in the third embodiment, the second embodiment of the present invention is used.
In the case where there is too much smoothing in the case of carrying out the above embodiment, there is an effect of correcting too much smoothing.

As described above, according to the present invention, it is possible to reduce noise and obtain an image with a clean edge portion in which the features of the image are preserved.

[Brief description of drawings]

FIG. 1 is a diagram showing a procedure of a third exemplary embodiment of the present invention.

FIG. 2 is a diagram showing a procedure of a conventional technique.

FIG. 3 is a diagram showing a procedure of the first exemplary embodiment of the present invention.

FIG. 4 is a diagram showing a procedure of a second exemplary embodiment of the present invention.

FIG. 5 is a diagram showing an example of directions in which luminance changes are evaluated and sampling points.

FIG. 6 is a diagram showing another example of a direction for evaluating a luminance change and a sampling point.

[Explanation of symbols]

101 ... Step of creating luminance change vector, 102
... a step of creating a blurred brightness change vector, 103 ... a step of obtaining a minimum change direction from the blurred brightness change vector, 113 ... a step of obtaining a minimum change direction from the brightness change vector, 104 ... a step of smoothing to a minimum change direction, 105 ... step of smoothing in a direction deviated from the minimum change direction by a predetermined angle, 106 ... data smoothed in the direction of the minimum change and data smoothed in a direction deviated from the minimum change direction by a predetermined angle depending on the magnitude of the brightness change Weighted averaging 107, step 107: weighted averaging of the result of step 106 and the input image

─────────────────────────────────────────────────── --- Continuation of the front page (72) Inventor Kunihiko Kido 1099, Ozenji, Aso-ku, Kawasaki-shi, Kanagawa Inside Hitachi Systems Development Laboratory (72) Inventor, Koichi Sano 1099, Ozenji, Aso-ku, Kawasaki-shi, Kanagawa (56) Reference JP-A-7-262368 (JP, A) JP-A-6-348842 (JP, A) JP-A-5-181956 (JP, A) JP-A-8- 161483 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G06T 1/00-7/60 A61B 6/00-6/14

Claims (6)

(57) [Claims] 1. A brightness change vector that calculates a value for evaluating the magnitude of a brightness change of an image in each of a plurality of predetermined directions at each point of an image, and has the brightness change values in the plurality of predetermined directions as component values. Is created for each point of the image, and for the brightness change vector at each point of the image, the brightness change vectors in the vicinity at a predetermined distance are vector-added with a predetermined weight to obtain a blurred brightness change vector. Created at each point of the image, the direction in which the component value of the blurred brightness change vector becomes the minimum at each point of the image is obtained as the minimum change direction, and at each point of the image, 1 is set in each of the minimum change directions. An image processing method characterized by smoothing a dimension. 2. A brightness change vector having a value for evaluating the magnitude of brightness change of the image in each of a plurality of predetermined directions at each point of the image, and a brightness change vector having the brightness change value in the plurality of predetermined directions as a component value. Is created for each point of the image, and for the brightness change vector at each point of the image, the brightness change vectors in the vicinity at a predetermined distance are vector-added with a predetermined weight to obtain a blurred brightness change vector. Created at each point of the image, the direction in which the component value of the blurred brightness change vector becomes the minimum at each point of the image is obtained as the minimum change direction, and at each point of the image, 1 is set in each of the minimum change directions. Dimensional smoothing is performed, and at each point of the image, one-dimensional smoothing is performed in a direction deviated from the minimum change direction by a predetermined angle. Of the luminance change, and a value smoothed in the minimum change direction and a value smoothed in a direction deviated from the minimum change direction by a predetermined angle in accordance with the captured value of the brightness change. An image processing method characterized by performing weighted averaging with different weights. 3. The image processing method according to claim 2, wherein the predetermined angle is 45 degrees. 4. The image processing method according to claim 2, wherein the sum of components of the brightness change vector is adopted as the amount of the brightness change near each point. 5. A brightness change vector at each point of the image, which calculates a value for evaluating the magnitude of the brightness change of the image in each of a plurality of predetermined directions, and which has the brightness change values in the plurality of predetermined directions as component values. Is created for each point of the image, and for the brightness change vector at each point of the image, the brightness change vectors in the vicinity at a predetermined distance are vector-added with a predetermined weight to obtain a blurred brightness change vector. Created at each point of the image, the direction in which the component value of the blurred brightness change vector becomes the minimum at each point of the image is obtained as the minimum change direction, and at each point of the image, 1 is set in each of the minimum change directions. Dimensional smoothing is performed, and at each point of the image, one-dimensional smoothing is performed in a direction deviated from the minimum change direction by a predetermined angle. Of the luminance change, and a value smoothed in the minimum change direction and a value smoothed in a direction deviated from the minimum change direction by a predetermined angle in accordance with the captured value of the brightness change. The image processing method is characterized by performing weighted averaging with different weights, and performing weighted averaging of the input image and the weighted average. 6. At each point of the image, the brightness values of neighboring points are added with a predetermined weight to create a blurred image, and at each point of the blurred image, the brightness of the blurred image is obtained in a plurality of predetermined directions. A value for evaluating the magnitude of change is calculated, a brightness change vector having the brightness change values in the plurality of predetermined directions as component values is created for each point of each of the blurred images, and at each point of the blurred image, The direction in which the component value of the brightness change vector is the minimum is obtained as the minimum change direction, and at each point of the image corresponding to each point of the blurred image, the minimum change direction of the blurred image is set to 1 Dimensional smoothing is performed, and at each point of the image, one-dimensional smoothing is performed in a direction deviated from the minimum change direction by a predetermined angle, and at each point of the image, a change in luminance near each point is captured. Calculate the amount of the brightness change An image characterized by performing weighted averaging by changing the weight of the value smoothed in the minimum change direction and the value smoothed in the direction deviated from the minimum change direction by a predetermined angle according to the value Processing method.