JP3746441B2 - Video signal processing device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a video signal processing apparatus such as a television receiver, and is particularly suitable for obtaining a high-quality progressive image when converting an interlaced image to a progressive image (hereinafter referred to as IP conversion). The present invention relates to a processing apparatus.
[0002]
[Prior art]
In many video signals, interlaced scanning is used as a scanning form. However, when this video signal is displayed on the interlaced scanning image display unit, interlace interference such as line flickering occurs and the image quality deteriorates.
[0003]
This interlace interference can be eliminated by converting the interlaced image into a progressive image and displaying it in the form of progressive scanning. A television receiver having this function is also commercialized.
[0004]
In the IP conversion, signal processing for generating a signal of a scanning line missing in interlace scanning by interpolation processing is performed. This interpolation processing includes motion adaptive type and motion compensation type signal processing.
[0005]
In the motion adaptive type, inter-frame interpolation suitable for a still image region and intra-field interpolation suitable for a moving image region are interpolated by changing the mixing ratio according to the motion information of the image, and the inter-frame difference signal is calculated. Since motion information is detected in large and small sizes, this motion information does not necessarily correspond to an accurate motion of an image. For this reason, in the still image area, the interlace interference can be completely removed and a significant image quality improvement effect can be obtained. There are few issues.
[0006]
Here, as a method of intra-field interpolation, there is conventionally known a method in which interpolation pixels can be averaged using upper and lower pixels that actually exist. In this method, since one field of one frame is generated, sampling is performed at (1/2) × f. According to Shannon's theorem, the frequency band that can be reproduced by this method As shown in FIG. 3, the maximum is (1/4) × f. However, in operation in an actual circuit, frequency components in the vicinity of (1/4) × f are lost, and frequency components up to (1/4) × f are not completely compensated. ), And has a problem that the image tends to have a reduced resolution. Note that f is an acronym representing frequency, and f described when quoting Shannon's theorem naturally means sampling frequency, and f in the following description is also sampling frequency. say.
[0007]
On the other hand, the motion compensation type detects the motion of an image as motion vector information, performs signal processing for motion compensation on the preceding and succeeding field signals with this motion vector information, and generates an interpolated scanning line signal. With this method, interpolation processing that substantially matches the motion of the image can be performed.
[0008]
Regarding an image processing apparatus for IP conversion, Japanese Patent Laid-Open No. 10-112845 as a conventional technique describes a technique for detecting a motion vector between frames of an image and performing IP conversion.
[0009]
[Problems to be solved by the invention]
However, if the motion vector between frames is detected and IP conversion is performed as in the technique described in the above publication, it is essential to detect a motion vector with high accuracy in order to perform accurate motion compensation signal processing. For this reason, there is a problem that enormous signal processing is required.
[0010]
In addition, in the intra-field interpolation, the frequency band that can be reproduced by the conventional method is (1/4) × f at the maximum as shown in FIG. 3, but actually, it is in the vicinity of (1/4) × f. There is a problem that the frequency component is lost, and the image is likely to have a reduced resolution.
[0011]
The object of the present invention is made in view of such a situation, and when a conventional motion adaptive interpolation reduces an annoying image degradation and performs a video portion interpolation with a small amount of signal processing. It is to be able to prevent a decrease in resolution.
[0012]
Furthermore, when compensating for lost frequency components to improve the degradation of resolution, overshoot and undershoot may occur, resulting in an unnatural image. Suppresses shoots and undershoots to obtain a natural image.
[0013]
[Means for Solving the Problems]
In order to solve the above problems, the present invention mainly adopts the following configuration.
In a video signal processing apparatus for converting an interlaced image into a progressive image,
Generating means for generating interpolation reference data based on an averaging process of upper and lower pixels existing in the current field;
An extraction means that is a band-pass filter process that extracts a frequency component that can be lost in the averaging process of the upper and lower pixels by using an adjacent pixel located in a direction perpendicular to the interpolation target pixel;
Adding means for adding the interpolation reference data output from the generating means and the correction value output from the extracting means;
Suppressing means that is provided on the output side of the adding means and suppresses overshoot and undershoot occurring in the interpolation data output from the adding means;
Maximum value detecting means for detecting a maximum value from two line signals existing above and below the interpolation line;
Minimum value detecting means for detecting a minimum value from two line signals existing above and below the interpolation line;
Maximum value coefficient multiplication means for multiplying the maximum value obtained from the maximum value detection means by a coefficient (1 + K);
Minimum value coefficient multiplication means for multiplying the minimum value obtained from the minimum value detection means by a coefficient (1-K),
The suppression means suppresses the interpolation data output from the addition means to be within a value obtained from the minimum value coefficient multiplication means and the maximum value coefficient multiplication means,
The interpolation is performed using only the signal in the current field to improve the reduction in resolution.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
A video signal processing apparatus according to an embodiment of the present invention will be described below with reference to the drawings. Here, FIG. 1 is a diagram illustrating an example of input image data, and FIG. 2 is a configuration example of an image processing apparatus according to an embodiment of the present invention. 3 is a diagram showing frequency bands that can be reproduced by intra-field LPF interpolation, FIG. 4 is a diagram showing frequency characteristics of the BPF 2 employed in the present embodiment shown in FIG. 2, and FIG. FIG. 3 is a diagram illustrating frequency characteristics of an LPF 9 used in the present embodiment shown in FIG. FIG. 6 is a diagram illustrating an interpolated image of an interlaced image having a diagonal line after the addition of the adder 3 according to the present embodiment, and FIG. 7 is an upper and lower average value of the interlaced image having a diagonal line. It is a figure which shows the example of an image interpolated by. FIG. 8 is a diagram illustrating an example of an image obtained by interpolating an interlaced image having a horizontal edge. FIG. 9 shows an interlaced image (a) having an oblique line before passing through the interpolation value suppression circuit 8 and an interlaced image having an oblique line after passing through the interpolation value suppression circuit 8 according to this embodiment. It is a figure which shows the interpolated image (b).
[0018]
First, the relationship between FIG. 1 and FIG. 2 will be described. The white circles in FIG. 1 indicate pixels in the field, and the black circles indicate pixels not in the field. In the present invention, the interpolation data between the A line and the B line shown in FIG. 1 and intersecting with the C line of the vertical line is to be obtained by elaborating. The signals shown as A, B, and C in FIGS. 1 and 2 correspond to each other.
[0019]
The configuration example of this embodiment shown in FIG. 2 is an LPF (low-pass filter) 1 for obtaining an interpolation reference pixel (reference value of an interpolation pixel) from upper and lower pixels that exist in the current field, and upper and lower pixels that exist in the current field. BPF (bandpass filter) 2 for obtaining a frequency component that can be lost in the averaging process, an adder 3 for adding the frequency component of the current field obtained by BPF2 and the interpolation reference pixel obtained by LPF1, an interpolation line The maximum value obtained from the detector 4 that detects the maximum value from the upper and lower two-line signals (A line and B line signals), the detector 5 that detects the minimum value from the upper and lower two-line signals of the interpolation line, and the maximum value detector 4 A coefficient multiplication circuit 6 that multiplies the value by a coefficient (1 + K), a coefficient multiplication circuit 7 that multiplies the minimum value obtained from the minimum value detector 5 by a coefficient (1-K), and interpolation data obtained by the adder 3 is multiplied by a coefficient. Circuit 6 and 7 Interpolation value suppression circuit 8 that falls within the obtained value, LPF 9 that reduces noise generated in the signal by interpolation value suppression circuit 8 that is nonlinear processing, intra-field interpolation data or previous field interpolation position data from the motion detection result SEL10 to be selected.
[0020]
The operation of the configuration shown in FIG. 2 will be described below. The upper and lower 3 pixels (total 6 pixels existing in the field) of the interpolation data at the center of the range indicated by C in FIG. 1 are input to LPF1 and BPF2. The LPF1 outputs interpolation reference data for obtaining interpolation data from the input data. This interpolation reference data is output in a state in which there is a possibility that the frequency component in the area filled with a broken line shown in FIG. 3 is lost (in the process of signal processing in an actual circuit), and is input to the adder 3 Is done.
[0021]
Therefore, the BPF 2 has a frequency characteristic including a frequency region that may be lost as shown in FIG. 4, and outputs a frequency component around the interpolation data from the input current field data, which is added as a correction value. Input to the device 3.
[0022]
Therefore, the adder 3 adds the interpolation reference data output from the LPF1 in advance and the correction value output from the BPF 2 and outputs the interpolation data reproducing the frequency up to (1/4) × f.
[0023]
FIG. 6 simulates an interlaced image interpolated with a slant line that rises to the left, and shows the change in the horizontal direction of the interpolation line and its upper and lower line data. When attention is paid to the n-th pixel in the horizontal direction and its neighboring pixels, the interpolation data strengthens the connection between the line A and the line B.
[0024]
As shown in FIG. 7, when the vertical average value is used as the interpolation data as in the conventional method, the horizontal n-th pixel that strengthens the connection between the line A and the line B is not obtained, and the resolution is improved. It becomes a fallen image. In other words, when the luminance Y in FIG. 7 is viewed, the level of the interpolation line C is about half that of the line A or B and does not become a smooth white line as an oblique line. On the other hand, when the interpolation line C in FIG. 6 is viewed, it has about 80% of the luminance of the line A or the line B and can be a smooth diagonal line including the interpolation line C.
[0025]
As described above, in this embodiment, by adding the interpolation reference data and the correction value output from the BPF 2, a smooth tan line can be expressed while maintaining a sense of resolution.
[0026]
On the other hand, in order to detect the upper limit and lower limit values of the interpolation data, the data on the upper and lower lines of the interpolation data shown in FIGS. 1A and 1B (14 pixels in total 7 pixels in the horizontal direction) are used as the maximum value detector 4 and the minimum value. Each is input to the detector 5. The detector 4 detects the maximum value of the upper and lower line data, and this value is multiplied by a coefficient (1 + K) in the coefficient multiplication circuit 6. The detector 5 detects the minimum value of the upper and lower line data, and the coefficient multiplication circuit 7 multiplies the value by a coefficient (1-K).
[0027]
The data output from the adder 3 reproduces the frequency band up to (1/4) × f. In the signal processing for obtaining the interpolation signal based on the existing pixel signal, overshoot and under Shoots may be seen. As shown in FIG. 8, when interlaced images having horizontal edges are interpolated, overshoot in which the interpolation data becomes larger than necessary and undershoot in which the interpolation data becomes smaller than necessary are observed in the upper and lower portions of the edges. There is a case. In order to suppress this, the interpolated data output from the adder 3 is input to the interpolated value suppression circuit 8 that falls within the maximum and minimum values of the upper and lower lines output from the coefficient multiplication circuits 6 and 7.
[0028]
The effect of the interpolation value suppression circuit 8 will be described with reference to FIG. FIG. 9 simulates the case of interpolating an interlaced image having a tanning line that rises to the left. The nth interpolation data in the horizontal direction of the interpolation line C output from the adder 3 has a value protruding from the signals of the pixels (line A and line B existing in the field) of the upper and lower lines (FIG. 9 ( a)). In order to alleviate such prominent values, that is, overshoot, overshoot and undershoot as shown in FIG. 8, the interpolation data is averaged between the upper and lower pixels, or the interpolation data is stored within the upper and lower pixel values. However, in such a method, the effect of compensating for the lost frequency component by BPF 2 described with reference to FIG. 6 is lost, and the resolution is lowered.
[0029]
Therefore, in the present invention, the maximum value and the minimum value of the upper and lower lines of the interpolation line are detected and set as the upper and lower limits of the interpolation data, and the overshoot and undershoot of the interpolation data output from the adder 3 are suppressed. (FIG. 9B).
[0030]
In the present invention, the upper and lower limits of the interpolation data are not constant values but values around the interpolation pixel, so that the values are flexible in the screen and a natural image can be obtained. Here, the total number of pixels in the horizontal direction of the upper and lower lines for finding the maximum value and the minimum value is 14 pixels in FIG. 1, but the number of pixels may have to be further reduced depending on the circuit scale. In this case, the upper and lower limits of the interpolation data can be varied using the coefficient multiplication circuits 6 and 7, and the same effect can be obtained even if the number of pixels in the horizontal direction is small. As a result, an image in which the tan lines are smoothly connected as shown in FIG. 9 is obtained.
[0031]
The circuit 8 for removing the overshoot and undershoot of the signal output from the adder 3 may cause noise due to non-linear processing, which may reduce the image quality. Therefore, it is possible to input the signal obtained from the circuit 8 to the LPF 9 having the frequency characteristics as shown in FIG. 5 to reduce noise due to nonlinear processing and obtain interpolation data in which the lost frequency component is compensated.
[0032]
Referring to the circuit configuration shown in FIG. 2, the result of motion detection as a conventional technique, the interpolation data output from the LPF 9, and the previous field interpolation position data are input to SEL10, and the still image portion includes The final interpolation data is obtained by outputting the previous field interpolation position data and the interpolation data in which the frequency component according to the present invention is supplemented for the moving image portion.
[0033]
Here, further, image processing such as general contour enhancement for an image will be described. Overshoot and undershoot may occur after the image enhancement processing, and unnaturalness may appear in the image. When it is desired to obtain a natural image by suppressing the overshoot and undershoot that are the cause of this unnaturalness, the image signal input, the maximum value detection circuit, and the minimum value detection that constitute the main part of the circuit shown in FIG. A circuit, a coefficient multiplying circuit, and an image suppression circuit can be applied and exhibit an effect. An example of the configuration is shown in FIG.
[0034]
A delay circuit 11 that obtains 1H delayed data by delaying the input signal by one horizontal scanning period; a delay circuit 12 that obtains data delayed by 2H by further delaying the signal output from the delay circuit 11; A maximum value detector 13 for detecting the maximum value of a total of 14 pixels of 7 pixels in the horizontal direction of 7 pixels in the horizontal direction and 2H delayed data, and 7 pixels in the horizontal direction of the data delayed by 2H in the horizontal direction of the input signal A minimum value detector 14 for detecting a minimum value of a total of 14 pixels, a coefficient multiplier circuit 15 for multiplying the aforementioned maximum value by (1 + K), and a coefficient multiplier circuit 16 for multiplying the aforementioned minimum value by (1−K). The suppression circuit 17 includes a maximum value and a minimum value multiplied by a coefficient as upper and lower limits, and a signal delayed by 1H within the range.
[0035]
The operation of each component shown in FIG. 10 will be described below. A video signal is input to the delay circuit 11 and delayed by one horizontal scanning period to obtain 1H delayed data. Next, the signal output from the delay circuit 11 is further delayed by 1H, and data delayed by 2H in the delay circuit 12 is obtained. The input signal and the data delayed by 2H are input to the maximum value detector 13, and the maximum value of a total of 14 pixels of 7 pixels in the horizontal direction of the input signal and 7 pixels in the horizontal direction of the data delayed by 2H is detected.
[0036]
The minimum value detector 14 receives the input signal and data delayed by 2H, and detects the minimum value of a total of 14 pixels of 7 pixels in the horizontal direction of the input signal and 7 pixels in the horizontal direction of the data delayed by 2H. The maximum value output from the maximum value detector 13 is multiplied by (1 + K) by the coefficient multiplier circuit 15, and the maximum value output from the maximum value detector 13 output from the minimum value detector 14 is output by the coefficient multiplier circuit 15 ( 1 + K) times, and the minimum value is multiplied by (1−K) by the coefficient multiplication circuit 16.
[0037]
The signal delayed by 1H by the delay circuit 11 and the maximum and minimum values multiplied by the coefficient are input to the signal suppression circuit 17, and the maximum and minimum values multiplied by the coefficient are set as the upper and lower limits, and the suppression circuit is within that range. 17 is output with a signal delayed by 1H.
[0038]
The maximum and minimum values of the input line and the line delayed by 2H are detected and used as the upper and lower limits of the line delayed by 1H, and overshoot and undershoot that cause unnaturalness of the image after contour enhancement are suppressed. is doing.
[0039]
The above description will be specifically explained by borrowing the image data shown in FIG. 1. If the B line is below the A line and the C line is the lower line, the respective inputs to the maximum value and minimum value detection circuits will be described. Are the A line and the C line, and the input to the signal suppression circuit is the B line signal from the 1H delay circuit in addition to the input from the coefficient multiplier circuit. In other words, the main circuit shown in FIG. 2 can also be used in image processing for contour enhancement of the B line signal.
[0040]
In this embodiment, the upper and lower limits of the signal to be suppressed are not constant values but values around the interpolated pixels, so that the values are flexible in the screen and a natural image is obtained. Here, the total number of pixels in the horizontal direction of the upper and lower lines for finding the maximum value and the minimum value is 14 pixels in the above example, but the number of pixels may have to be further reduced depending on the circuit scale. In this case, the upper and lower limits of the interpolation data can be varied using the coefficient multiplication circuits 15 and 16, and the same effect can be obtained even if the number of pixels in the horizontal direction is small.
[0041]
As a result, an image in which overshoot and undershoot are observed due to contour emphasis can be obtained without suppressing the effect of contour emphasis, and overshoot and undershoot can be suppressed to obtain a natural image.
[0042]
【The invention's effect】
According to the present invention, the following effects can be expected. In other words, conventionally, the interpolation data used in the moving image portion has been averaged using the upper and lower pixels that exist, but this method has a frequency region that is lost, resulting in an image with no sense of resolution. On the other hand, in the present invention, a lost frequency region is compensated, and overshoot and undershoot are suppressed, and a natural image can be obtained. Therefore, it is possible to provide a high-quality progressive image with a high resolution even in the moving image portion.
[0043]
In addition, since interpolation data is generated in the current field without using signals in the front and rear fields, a complicated circuit configuration or a large-scale circuit is not required, and the circuit scale can be suppressed by using a coefficient multiplier circuit. it can.
[0044]
The image suppression circuit can also be applied to obtain a natural image by suppressing overshoot and undershoot that cause unnaturalness of an image after image processing such as general edge enhancement.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating an example of input image data.
FIG. 2 is a configuration example of an image processing apparatus according to an embodiment of the present invention.
FIG. 3 is a diagram showing frequency bands that can be reproduced by intra-field LPF interpolation.
FIG. 4 is a diagram showing frequency characteristics of BPF 2 employed in the present embodiment shown in FIG.
5 is a diagram showing frequency characteristics of the LPF 9 used in the present embodiment shown in FIG.
FIG. 6 is a diagram showing an interpolated image of an interlaced image having diagonal lines after adding three adders according to the present embodiment.
FIG. 7 is a diagram illustrating an example of an image interpolated by an upper and lower average value of an interlaced image having an oblique line.
FIG. 8 is a diagram illustrating an example of an image obtained by interpolating an interlaced image having a horizontal edge.
FIG. 9 shows an interlaced image (a) of an interlaced image having a diagonal line before passing through the interpolation value suppression circuit 8 and an interlaced image having a diagonal line after passing through the interpolation value suppression circuit 8 according to the present embodiment; It is a figure which shows the interpolated image (b).
FIG. 10 is a diagram showing a circuit configuration block using a maximum value detection circuit, a minimum value detection circuit, a coefficient multiplication circuit, and a signal suppression circuit in the video signal processing apparatus according to the embodiment of the present invention.
[Explanation of symbols]
1 LPF in the field
2 BPF for frequency component extraction
3 Adder 4 to 13 Maximum value detection circuit 5 to 14 Minimum value detection circuit 6 to 7, 15 to 16 Coefficient multiplication circuit 8 Interpolation data suppression circuit 9 LPF for noise reduction by circuit 8
l0 signal selection circuit 11-12 1H delay circuit 17 signal suppression circuit

Claims (2)

In a video signal processing apparatus for converting an interlaced image into a progressive image,
Generating means for generating interpolation reference data based on an averaging process of upper and lower pixels existing in the current field;
An extraction means that is a band-pass filter process for extracting a frequency component that can be lost in the averaging process of the upper and lower pixels by using a pixel in the vicinity in the vertical direction with respect to the interpolation target pixel ;
Adding means for adding the interpolation reference data output from the generating means and the correction value output from the extracting means;
Suppressing means that is provided on the output side of the adding means and suppresses overshoot and undershoot generated in the interpolation data output from the adding means;
Maximum value detecting means for detecting a maximum value from two line signals existing above and below the interpolation line;
Minimum value detecting means for detecting a minimum value from two line signals existing above and below the interpolation line;
Maximum value coefficient multiplication means for multiplying the maximum value obtained from the maximum value detection means by a coefficient (1 + K);
Minimum value coefficient multiplication means for multiplying the minimum value obtained from the minimum value detection means by a coefficient (1-K),
The suppression means suppresses the interpolation data output from the addition means to be within the values obtained from the minimum value coefficient multiplication means and the maximum value coefficient multiplication means,
A video signal processing apparatus characterized by performing interpolation to improve a decrease in resolution using only signals in the current field. The video signal processing device according to claim 1,
The video signal processing apparatus, wherein the suppression processing by the suppression means is nonlinear processing, and a reduction means for reducing noise generated by the nonlinear processing is provided on the output side of the suppression means.

Images