JP4412435B2 - Field interpolation method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a field interpolation method for converting an interlaced image to a non-interlaced image, and more particularly to an improvement for suppressing erroneous detection and detection omission without increasing the amount of calculation.
[0002]
[Prior art]
The following three methods (1) to (3) are conventionally known as methods for converting an interlaced image into a non-interlaced image by interpolating lines (scanning lines) in each field of the interlaced image. It has been. That is,
(1) Interpolation method using adjacent field pixels;
(2) a method of interpolating using adjacent pixels above or below; and
(3) A method of performing interpolation using an average value of values of pixels adjacent above and below.
[0003]
Method (1) is called inter-field interpolation, and interpolation is performed using pixels in adjacent fields as shown in FIG. Methods (2) and (3) are called intra-field interpolation, and interpolation is performed within the same field as shown in FIG. 4 (b). 4 (a) and 4 (b), the vertical axis V represents the vertical direction on the screen (direction perpendicular to the line), and symbols (i−1), (i), and (i + 1) represent time. Represents the identification number assigned to the field along the flow of. Field (i) is a field to be interpolated, field (i-1) corresponds to the previous field, and field (i + 1) corresponds to the next field. To do. 4A and 4B, black circles represent pixels to be interpolated (interpolated pixels), and white circles represent pixels included in the interlaced image before interpolation.
[0004]
Inter-field interpolation is suitable for still images, while intra-field interpolation is suitable for moving images. That is, each has suitability / unsuitability depending on the movement of the image. For this reason, in order to obtain a high-quality non-interlaced image, a method called “motion adaptive interpolation” is employed in which the motion amount of the image is obtained and the intra-field interpolation and inter-field interpolation are gradually switched. In this method, a motion coefficient k (0 ≦ k ≦ 1) that expresses the amount of motion of an image is calculated, and a value obtained by weighted averaging with a motion coefficient k is obtained for a value obtained by intra-field interpolation and a value obtained by inter-field interpolation. Is adopted as the value of the interpolated pixel. The motion coefficient k is k = 0 for a still image having no image motion, and k = 1 for a moving image having a large motion.
[0005]
[Problems to be solved by the invention]
However, in the conventional motion adaptive interpolation, a detection omission that detects a moving image region as a still image region is one cause of image quality degradation. If the detection sensitivity is increased to avoid this, the still image region is reversed. Erroneously detected as a moving image region, and even if the amount of calculation is increased to increase the detection accuracy, there is a problem that an effect corresponding to that cannot be obtained.
[0006]
The present invention has been made to solve the above-described problems in the prior art, and is a field that reduces false detection and detection omission through a simple calculation, and realizes field interpolation adapted to the motion and contour of an image. An object is to provide an interpolation method.
[0007]
[Means for Solving the Problems]
A method according to a first aspect of the present invention is a field interpolation method for converting an interlaced image into a non-interlaced image by interpolating a line in each field, and the value of an interpolated pixel that is a pixel on the line to be interpolated is The step of determining is: (a) in the same field as the interpolation pixel, on the line adjacent above the interpolation pixel, the interpolation pixel based on the value of M (≧ 1) pixels closest to the interpolation pixel A first candidate value that is a candidate for a value, and a second candidate value that is a candidate for the value of the interpolated pixel based on the values of the M pixels closest to the interpolated pixel on a line adjacent below the interpolated pixel (B) calculating a maximum value of the first candidate value and the second candidate value; and (c) calculating a minimum value of the first candidate value and the second candidate value. (D) adjacent to the same field before and after A step of calculating an average value of pixel values at positions corresponding to the interpolation pixels in two fields, and (e) the interpolation in N (≧ 1) fields selected within a range before the same field. A step of calculating a variable that varies in a range from zero to an absolute value of a pixel value variation range, depending on a motion coefficient at a position corresponding to a pixel, so that each of these motion coefficients becomes smaller, ( f) calculating an intermediate value between a value obtained by adding the variable to the maximum value, a value obtained by subtracting the variable from the minimum value, and the average value; and (g) calculating in the step (f). Determining a value of the interpolated pixel based on the intermediate value .
[0008]
In the method of the second invention, in the field interpolation method of the first invention, the step (a) sets the M as 1, and (a-1) sets the first candidate value above the interpolated pixel recently. A step of providing a value of a pixel that is in contact; and (a-2) a step of providing a value of a pixel that is closest to the lower side of the interpolation pixel as the second candidate value.
[0009]
In the method of the third invention, in the field interpolation method of the first invention, the step (a) sets the M to 3, and (a-1) the two lines adjacent above and below the interpolated pixel. A step of providing, as the first and second candidate values, a set of values that are closest to each other among a set of pixels that are symmetrically positioned with respect to the interpolation pixel as a center.
[0010]
In the method of the fourth invention, in the field interpolation method of the first invention, the step (a) sets the M to 3, and (a-1) the two lines adjacent above and below the interpolation pixel. In each of the steps, a value of a point located between the M pixels is calculated based on the value of the M pixels, and (a-2) the interpolated pixel on the two lines. And a step of giving the value of the set whose values are closest to each other as the first and second candidate values among the set of pixels and the set of points that are symmetric with respect to each other.
[0011]
In the method of the fifth invention, in the field interpolation method of any one of the first to fourth inventions, the step (e) comprises (e-1) the interpolation pixel for each of the N fields. Comparing the motion coefficient at the position corresponding to the reference value with (e-2) based on the result of the step (e-1), if all the motion coefficients are lower than the reference value, the variable Is set to the absolute value of the pixel value fluctuation range, the variable is set to zero if all the motion coefficients are higher than the reference value, and when only some of the motion coefficients are lower than the reference value, And a step of setting the variable higher as the lower motion coefficient increases.
[0012]
According to a sixth invention, in the field interpolation method according to any one of the first to fifth inventions, the step (e) includes (e-3) calculating a motion coefficient at the interpolation pixel in the same field. And (e-4) obtaining the motion coefficient at the position corresponding to the interpolated pixel in the stored N fields, and (e-5) the step (e- And a step of calculating the variable based on the motion coefficient obtained in 4).
[0013]
According to a seventh aspect of the present invention, in the field interpolation method of any one of the first to fourth aspects, the step (e) includes (e-1) calculating a motion coefficient at the interpolated pixel in the same field. (E-2) comparing the motion coefficient with a reference value and storing the result, and (e-3) storing the result corresponding to the interpolation pixels in the N fields stored. Based on the comparison result between the motion coefficient at the position and the reference value, if all the motion coefficients are lower than the reference value, the variable is set to the absolute value of the pixel value fluctuation range, and all the motion coefficients are Setting the variable to zero if it is higher than a reference value, and setting the variable higher as there are more motion coefficients lower than the reference value when only some of the motion coefficients are lower than the reference value.
In the method of the eighth invention, in the field interpolation method of any one of the first to seventh inventions, the step (g) includes (g-1) the intermediate value calculated in the step (f), Determining the value of the interpolated pixel.
According to a ninth aspect of the present invention, in the field interpolation method of any one of the first to seventh aspects, the step (g) includes (g-1) inter-field interpolation, intra-field interpolation, and motion adaptive interpolation. And (g-2) the interpolation value calculated in the step (g-1) and the intermediate value calculated in the step (f). And calculating the mixed value as the value of the interpolated pixel.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a flowchart showing a processing procedure of a field interpolation method according to an embodiment of the present invention. 2 and 3 are explanatory diagrams showing the principle of processing along the procedure of FIG. In these drawings, the meanings of the symbol V and the symbols (i-1), (i), (i + 1) are the same as the meanings of the same symbols in FIGS. 4 (a) and 4 (b). A symbol H represents a horizontal direction on the screen (a direction parallel to the line).
[0015]
Reference symbol P represents an interpolation pixel, and reference symbols A, B, and C indicate the three pixels closest to the interpolation pixel P on the line adjacent to the interpolation pixel P in the field (i) to which the interpolation pixel P belongs. Symbols A ′, B ′, and C ′ represent three pixels that are closest to the interpolation pixel P on a line adjacent to the lower side of the interpolation pixel P in the field (i). The pixels A and A ′ are located symmetrically with respect to the interpolation pixel P. Similarly, the pixels B and B ′ and the pixels C and C ′ are in symmetrical positions. Pixels B and B ′ are adjacent to the interpolation pixel P in the vertical direction.
[0016]
Further, the symbol X represents a pixel at a position corresponding to the interpolation pixel P in the field (i−1) adjacent to the field (i), and the symbol Y represents the field (i) adjacent to the field (i). +1) represents a pixel at a position corresponding to the interpolation pixel P. Further, in FIG. 3, parallel bands extending horizontally represent lines (scanning lines) existing in each field before interpolation.
[0017]
In a color image, the Y component (luminance component) is preferably selected as the pixel value to be calculated, but other components can also be used, and each of the three components can be used individually. It is also possible to perform the same calculation. Hereinafter, in order to simplify the description, a pixel and its value (pixel value) are represented by the same symbol.
[0018]
FIG. 1 shows a processing procedure for one interpolation pixel P, and the processing in FIG. 1 is repeated for each interpolation pixel P. When processing for one interpolation pixel P is started, a value D as a candidate value of the interpolation pixel P based on the values of the pixels A, B, and C and an interpolation based on the values of the pixels A ′, B ′, and C ′. A value D ′ as a candidate for the value of the pixel P is calculated (step S1).
[0019]
In order to determine the candidate values D and D ′, among the set of pixels (A, A ′), (B, B ′), (C, C ′) at symmetrical positions with the interpolation pixel P as the center, The group with the closest value may be a candidate value group (D, D ′). This means that a set of candidate values (D, D ′) is set along the direction of the contour of the image. For example, when the direction of the outline extends obliquely from the upper right to the lower left, the values of the pixels C and C ′ are closest to each other, and thus a set of candidate values (D, D ′) = (C, C ′) is determined. .
[0020]
Alternatively, on the two lines adjacent to the top and bottom of the interpolated pixel P, the values of the points located between the pixel groups (A, A ′), (B, B ′), (C, C ′) are weighted averages. The set of points is added to the set of three pixels, and the set whose values are closest to each other is set as a set of candidate values (D, D ′). Also good.
[0021]
That is, the value (rA− {1−r} B, rA ′ − {1−r} B ′) is set as a set of points that fill between the pixel sets (A, A ′) and (B, B ′). The value (rC− {1-r} B, rC ′ − {) is used as a set of points that fill the space between the pixel pairs (C, C ′) and (B, B ′). A set with 1-r} C ′) can be supplemented. Here, r is a real number in the range of 0 ≦ r ≦ 1. If r = 0 or 1, the set of points coincides with any of the three sets of pixels. For example, when r = 0.5 is given as the value of r, a candidate value pair (D, D ′) can be selected from five pairs including pixels and points. Thereby, a set of candidate values (D, D ′) closer to the original value of the interpolation pixel P can be determined.
[0022]
r is treated as a variable, and among the pairs of all points corresponding to all r in the range of 0 ≦ r ≦ 1, the pair closest to each other can be used as the candidate value pair (D, D ′). Good. As a result, a set of candidate values (D, D ′) closer to the original value of the interpolated pixel P can be determined.
[0023]
On the other hand, as the simplest method, the pixel sets (A, A ′), (B, B ′), (C, C ′) are adjacent (A, A ′) above and below the interpolation pixel P. It is also possible to always determine a set of candidate values as (D, D ′) = (A, A ′), considering only In this case, the direction of the contour of the image is not considered.
[0024]
When the set of candidate values (D, D ′) is determined, the maximum value G = max (D, D ′) and the minimum value L = min (D, D ′) are calculated. Further, an average value Z = (X + Y) / 2 of the values of the pixels X and Y in the adjacent field is calculated (step S2).
[0025]
Further, in parallel with steps S1 and S2, a motion coefficient k (0 ≦ k ≦ 1) at the interpolation pixel P is calculated (step S3). Since the calculation of the motion coefficient k itself is well known in the art, detailed description thereof is omitted.
[0026]
Next, the motion coefficient k newly calculated in step S3 and the motion coefficient k already calculated for the position corresponding to the interpolation pixel P in the field before the field (i) are obtained. A variable s depending on the history is calculated (step S4). For this purpose, for example, the newly calculated motion coefficient k may be stored in some storage medium, and the motion coefficient k related to the past field may be read from the storage medium. The motion coefficient k itself may be stored, or only the magnitude relationship between the motion coefficient k and the reference value T may be stored. The variable s is calculated so as to change in the range from zero to the absolute value of the pixel value fluctuation range so that the value of each motion coefficient k in the history becomes smaller as the value increases. An example of calculation will be described later.
[0027]
When both steps S2 and S4 are completed, an intermediate value of G + s obtained by adding the variable s to the maximum value G, Ls obtained by subtracting the variable s from the minimum value L, and an average value Z is calculated (step S5). Next, interpolation of the pixel P is performed by determining the calculated intermediate value as the value of the interpolation pixel P (step S6). In step S6, instead of using the intermediate value directly, a mixed value with the interpolation value calculated by the conventional method may be adopted as the value of the interpolation pixel P. As the mixed value, for example, a weighted average value can be used. When the process of step S6 ends, the interpolation for the interpolation pixel P is completed.
[0028]
Next, the calculation (step S4) for determining the variable s will be described in detail. As an example, in step S4, motion coefficients k (i-2), k (at positions corresponding to the interpolation pixel P in three consecutive fields (i-2), (i-1), (i). i-1) and k (i) are obtained. These motion coefficients k are first compared with a reference value T. The reference value T is a constant set in advance in a range of 0 <T <1, and is, for example, 1/4 that is close to 0.
[0029]
As a result of the comparison, if k (i−2) <T, k (i−1) <T, and k (i) <T, the variable s is set to s = (absolute value of the pixel value fluctuation range). Is done. Accordingly, since the maximum value G> the maximum pixel value and the minimum value L <the minimum pixel value, the intermediate value calculated in step S5 is median () = Z. Adopting this intermediate value as the value of the interpolated pixel P means performing inter-field interpolation using the interpolated pixel P. That is, when the image is a still image or close to it in a certain range of history, inter-field interpolation is selected.
[0030]
As a result of the comparison, if k (i−2) ≧ T, k (i−1) ≧ T, and k (i) ≧ T, the variable s is set to s = 0. Thereby, the intermediate value calculated in step S5 is median () = median (G, L, Z). If this intermediate value is adopted as the value of the interpolated pixel P, interpolation is performed with a value less than or equal to G and greater than or equal to L, so that the generation of noise due to detection omission can be suppressed.
[0031]
As a result of the comparison, when only a part of k (i-2), k (i-1), and k (i) is lower than the reference value T, the variable s is 0 <s <(pixel value The absolute value of the fluctuation range is set so as to increase as the motion coefficient lower than the reference value T increases.
[0032]
In the above description, an example in which a history over three fields is considered as the history of the motion coefficient k is shown, but a history over more or less fields may be considered. For example, it is possible to consider only the motion coefficient k (i) of the field (i) to which the interpolation pixel P belongs. In this case, if k (i) <T, s = (absolute value of the pixel value fluctuation range) is set, and if k (i) ≧ T, s = 0 is set.
[0033]
It is also possible to determine the variable s in consideration of only the motion coefficient of the field before the field (i), for example, only the history of the motion coefficients k (i-2) and k (i-1). That is, in general, it is possible to determine the variable s based on the history of motion coefficients of N (≧ 1) fields selected within the range before the field (i). When N ≧ 2, it is desirable that the N fields are continuous, but it is possible to take a non-contiguous form.
[0034]
Further, in FIG. 1, steps S3 and S4 are illustrated as a highly efficient procedure executed in parallel with steps S1 and S2, but can be executed before and after each other.
[0035]
【The invention's effect】
In the method of the first invention, a variable that varies in the range from zero to the absolute value of the pixel value fluctuation range is calculated based on the motion coefficient history, and a value obtained by adding the variable to the maximum value and a variable for the minimum value are calculated. Since the intermediate value between the value obtained by subtracting and the average value is calculated, field interpolation adapted to the motion of the image reduces the false detection and detection omission, and also calculates the intermediate value through a single operation. Realize.
[0036]
In the method of the second invention, the values of the pixels closest to the upper and lower sides of the interpolation pixel are given as the first and second candidate values, so that the calculation is further simplified.
[0037]
In the method according to the third aspect of the invention, the first and second candidate values are selected from the values of the three sets of closest pixels, so that an appropriate candidate value corresponding to the direction of the contour line of the image is selected.
[0038]
In the method of the fourth invention, the first and second candidate values are selected from the total of not only the values of the three nearest neighbor pixels but also the values supplemented based on them. More suitable candidate values can be obtained.
[0039]
In the method of the fifth aspect of the invention, the motion coefficient is compared with the reference value, and the variable is determined based on the result, so the calculation required for calculating the variable is simplified.
[0040]
In the method of the sixth aspect of the invention, the motion coefficients at the interpolated pixels in the same field are calculated and stored, and the already calculated and stored values are used as the N motion coefficients necessary for calculating the variable. Therefore, the calculation efficiency is improved.
[0041]
In the method of the seventh invention, the motion coefficient is compared with the reference value, and the variable is determined based on the result, so that the calculation required for the calculation of the variable is simplified. Moreover, after the motion coefficient at the interpolated pixel in the same field is calculated and compared with the reference value, the result is stored, and the already stored result is used as N comparison results necessary for calculating the variable. Therefore, the storage capacity can be reduced when a storage medium or the like is used. In addition, the calculation efficiency is further improved.
[Brief description of the drawings]
FIG. 1 is a flowchart illustrating a processing procedure of a method according to an embodiment.
FIG. 2 is an explanatory diagram illustrating a principle of processing according to FIG. 1;
FIG. 3 is an explanatory diagram illustrating a principle of processing according to FIG. 1;
FIG. 4 is an explanatory diagram showing the principle of a conventional method.
[Explanation of symbols]
A, B, C Nearest pixel A ′, B ′, C ′ Nearest pixel D First candidate value D ′ Second candidate value G Maximum value k Motion coefficient L Minimum value P Interpolated pixel s Variable T Reference value X, Y Pixel Z average value

Claims (9)

A field interpolation method for converting an interlaced image into a non-interlaced image by interpolating lines in each field,
Determining the value of an interpolated pixel that is a pixel on the line to be interpolated,
(a) In the same field as the interpolated pixel, the interpolated pixel value is based on the value of M (≧ 1) pixels closest to the interpolated pixel on the line adjacent above the interpolated pixel. Calculating a first candidate value and a second candidate value that is a candidate for the value of the interpolated pixel based on the values of the M pixels closest to the interpolated pixel on a line adjacent below the interpolated pixel; When,
(b) calculating a maximum value of the first candidate value and the second candidate value;
(c) calculating a minimum value of the first candidate value and the second candidate value;
(d) calculating an average value of pixel values at positions corresponding to the interpolated pixels in two fields adjacent before and after the same field;
(e) Depends on a motion coefficient at a position corresponding to the interpolation pixel in N (≧ 1) fields selected within the range before the same field, and the smaller each of these motion coefficients, the smaller the motion coefficient. A step of calculating a variable that changes in a range from zero to an absolute value of the pixel value fluctuation range,
(f) calculating an intermediate value between a value obtained by adding the variable to the maximum value, a value obtained by subtracting the variable from the minimum value, and the average value;
(g) determining the value of the interpolated pixel based on the intermediate value calculated in the step (f);
A field interpolation method comprising: In the step (a), M is 1,
(a-1) providing a value of a pixel closest to the interpolation pixel as the first candidate value;
The field interpolation method according to claim 1, further comprising: (a-2) providing a value of a pixel nearest to the lower side of the interpolation pixel as the second candidate value. In the step (a), M is 3;
(a-1) Among the sets of pixels that are on the two lines adjacent above and below the interpolated pixel and are symmetrical with respect to the interpolated pixel, the group that is closest in value to each other The field interpolation method according to claim 1, further comprising the step of providing a value of as a first candidate value and a second candidate value. In the step (a), M is 3;
(a-1) calculating a value of a point located between the M pixels in each of the two lines adjacent above and below the interpolation pixel based on the value of the M pixels When,
(a-2) A set of values closest to each other among a set of pixels and a set of points that are on the two lines and are symmetric with respect to the interpolation pixel, The field interpolation method according to claim 1, further comprising the steps of giving as first and second candidate values. Step (e)
(e-1) for each of the N fields, comparing the motion coefficient at the position corresponding to the interpolated pixel with a reference value;
(e-2) Based on the result of the step (e-1), if all the motion coefficients are lower than the reference value, the variable is set to the absolute value of the pixel value fluctuation range, and all the motion coefficients are If the variable is higher than a reference value, the variable is set to zero, and when only some of the motion coefficients are lower than the reference value, the more the motion coefficient lower than the reference value, the higher the variable is set;
The field interpolation method according to claim 1, further comprising: Step (e)
(e-3) calculating and storing a motion coefficient at the interpolation pixel in the same field;
(e-4) obtaining the motion coefficient at the position corresponding to the interpolated pixel in the N fields stored;
The field interpolation method according to claim 1, further comprising: (e-5) calculating the variable based on the motion coefficient obtained in the step (e-4). . Step (e)
(e-1) calculating a motion coefficient at the interpolation pixel in the same field;
(e-2) comparing the motion coefficient with a reference value and storing the result;
(e-3) Based on the comparison result between the motion coefficient at the position corresponding to the interpolation pixel in the N fields stored and the reference value, all the motion coefficients are based on the reference value. If it is low, the variable is set to the absolute value of the pixel value fluctuation range, if all the motion coefficients are higher than the reference value, the variable is set to zero, and when only some of the motion coefficients are lower than the reference value, The field interpolation method according to claim 1, further comprising a step of setting the variable higher as the number of motion coefficients lower than the reference value increases. Step (g)
  (g-1) A step of determining the interpolated pixel value as the intermediate value calculated in the step (f)
The field interpolation method according to claim 1, further comprising: Step (g)
  (g-1) calculating an interpolated value of the interpolated pixel by any one of inter-field interpolation, intra-field interpolation, and motion adaptive interpolation;
  (g-2) A mixed value of the interpolated value calculated in the step (g-1) and the intermediate value calculated in the step (f) is calculated, and the mixed value is determined as the value of the interpolated pixel. And a process of
The field interpolation method according to claim 1, further comprising:

Images