JPH0584112B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION "Field of Industrial Application" This invention relates to a Y/
This invention relates to a C separation circuit.

"Background technology and its problems" One of the conventional Y/C separation circuits is that the subcarrier is inverted for each line within the same field, in other words, the subcarrier frequency is offset by _H /2 ( _H : line frequency). Things that take advantage of having something are known. This is because the luminance signal is concentrated at a frequency that is an integer multiple of _H , and the chroma signal is concentrated at a frequency that is an integer multiple of _H /2, so a 1H (H is one line period) delay element is used. It is thought that Y/C separation can be performed with a comb-shaped filter, and the above-mentioned comb-shaped filter is combined with a bandpass filter that separates components near the subcarrier frequency.

In this two-dimensional Y/C separation circuit, deterioration in image quality is noticeable in the case of an image without line correlation. For example, in the case of an image where the upper half up to a certain line on the screen is blue and the lower half below it is red, at this color boundary,
Bleeding occurs.

Therefore, in the two-dimensional Y/C separation circuit, in order to minimize the deterioration in image quality, the presence or absence of line correlation is determined from, for example, three consecutive lines of signals (passed through a hand bus filter), and the determination result is An adaptive control is required to control the Y/C separation operation accordingly. For example, if three consecutive lines of an NTSC signal passed through a bandpass filter are T, M, and B, these signals can be used to generate a chroma signal (C) and a luminance signal (Y) using the following algorithm. can be separated.

C=1/2 [M-1/2(T+B)], Y=[M+1/2(T+B)]...C=1/2(M-T), Y=1/2(M+T)...C = 1/2 (M-B), Y = 1/2 (M + B)... The formula is when the luminance signal and chroma signal can be considered uniform, that is, all of the T, M, and B signals are correlated in the vertical direction. Applicable if you have. Also, when T and M are almost the same and M and B have no correlation, the formula is applied; when T and M are uncorrelated and M and B are almost the same, the formula is applied. Applicable. Furthermore, if there is no correlation between the three signals T, M, and B, the following processing is performed.

(a) When the sample data of the color subcarrier in signal M that is separated by one period has approximately the same value,
Regarded as a color difference signal transient, C=M, Y=0.

(b) When the sample data of the color subcarrier in signal M that is separated by one period are not nearly equal,
Regarded as a luminance signal transient, C=0, Y=M.

Furthermore, the subcarrier of the NTSC system has a phase inversion between two consecutive frames. Taking advantage of this property, Y/
C isolation circuits are also known. In a Y/C separation circuit using two frame memories, if the corresponding signals of three consecutive frames are F ₁ , F ₂ , F ₃ , then C = 1/2 [F ₂ - 1/2 (F ₁ + F ₃ )], Y=1/2 [F ₂ +1/2 (F ₁ +F ₃ )], Y/C separation can be performed. Furthermore, in a configuration using one frame memory, Y/C separation can be performed by C=1/2 (F ₂ -F ₁ ) and Y=1/2 (F ₂ +F ₁ ).

A Y/C separation circuit using such a frame memory can perform complete Y/C separation when there is no movement between frames. In other words, if there is movement, the image becomes blurry.

“Object of the invention” This invention provides a two-dimensional Y/
In the C separation circuit, Y/C allows the parameters necessary for adaptive control to be set to the best values.
The purpose is to realize a separation circuit. Also,
This invention is based on the fact that Y/C separation using a frame memory can perform complete separation in still images, and when there is movement, it uses a two-dimensional Y/C separation circuit with adaptive control. , it is an object of the present invention to provide a Y/C separation circuit that can perform good Y/C separation. Furthermore, this invention provides an unused two-dimensional Y/
Using a C separation circuit, compare the performance of Y/C separation for multiple parameters and find the best parameters that minimize the separation error. ,
This parameter is adopted.

"Summary of the Invention" This invention determines the presence or absence of line correlation from a composite color television signal of different lines passed through a bandpass filter, and when there is a line correlation, chromatizes the signal passed through a comb filter. In a two-dimensional Y/C separation circuit that extracts the signal as an output, and extracts the signal not passed through the comb filter as the chroma output when there is no line correlation, and extracts the chroma output from the composite color television signal as the luminance output, The error in Y/C separation from this chroma output and luminance output is evaluated with respect to multiple reference values, and the one that minimizes this separation error can be set as the reference value for determining the presence or absence of line correlation. It is.

In addition, the present invention provides a Y/C separation circuit using a frame memory in addition to the two-dimensional Y/C separation circuit described above, and when there is no movement and the Y/C separation circuit using this frame memory is configured. First, the separation performance of the two-dimensional Y/C separation circuit is evaluated, a reference value that minimizes the separation error is determined, and this reference value is adopted.

Embodiment FIG. 1 shows the overall configuration of an embodiment of the present invention. In the figure, 1 is an input terminal for, for example, a digitalized NTSC composite color television signal, and 2 is a frame memory. The Y/C separation circuit 3 used is an adaptive control type two-dimensional Y/C separation circuit. The luminance signal (Y) and chroma signal (C) output from the two Y/C separation circuits 2 and 3 are supplied to the selector 4. This selector 4
The output of one Y/C separation circuit is selected by
It is taken out to output terminals 5Y and 5C.

Further, among the three consecutive frames of signals from the Y/C separation circuit 2, the first frame signal _F1 and the last frame signal _F3 are supplied to the motion detection circuit 6. This motion detection circuit 6 generates a detection signal whose level changes depending on the presence or absence of motion.
The selector 4 is controlled by this detection signal. In other words, when there is no movement, the selector 4 selects the output of the Y/C separation circuit 2 using frame memory, and when there is movement, the selector 4 selects the output of the two-dimensional Y/C separation circuit 3. select. The motion detection circuit 6 is supplied with a detection reference value γ from a terminal 7 .

Further, the luminance signal and chroma signal output from the Y/C separation circuit 3 are supplied to the separation evaluation circuit 8. The output of this separation evaluation circuit 8 is supplied to a reference value generation circuit 9. The separation evaluation circuit 8 is a two-dimensional Y/
A separation error of the C separation circuit 3 is detected, and a reference value generating circuit 9 generates a reference value ε that minimizes this separation error. In addition, the separation evaluation shows no movement and Y/C.
This is done only when separation circuit 2 is used, so
For this control, a detection signal from the motion detection circuit 6 is supplied to a reference value generation circuit 9.

FIG. 2 shows an example of the Y/C separation circuit 2 using a frame memory. A digital composite color television signal from an input terminal 1 is supplied to a frame memory 10, and the output of this frame memory 10 is supplied to a frame memory 11. Three consecutive frames of signals F ₁ , F ₂ , F ₃ are taken out by the frame memories 10 and 11 and supplied to an adder circuit 15 via multipliers 12 , 13 and 14, respectively. The coefficient of multiplier 12 is 1/4, the coefficient of multiplier 13 is 1/2, and the coefficient of multiplier 14 is 1/4.

The output of this adder circuit 15 is taken out as a chroma signal, and is also phase inverted and supplied to an adder circuit 16. The signal F ₂ is supplied to this adder circuit 16, and the output of the adder circuit 16 is taken out as a luminance signal Y. This frame memory 10, 11
The delay of the Y/C separation circuit 2 using F2 coincides with the time of the signal _F2 . Therefore, the signal F ₂ is taken out to the terminal 17 and supplied to the two-dimensional Y/C separation circuit 3.

FIG. 3 shows the configuration of an example of the motion detection circuit 6. As shown in FIG. The signal F ₃ supplied to the input terminal 1 and the phase-inverted version of the signal F ₁ output from the frame memory 11 are supplied to the adder circuit 18, and the difference between the two (δ=F ₃ −F ₁ ) is calculated. This difference component is supplied to the comparison circuit 19 and compared with the reference value γ from the terminal 7. When (δ<γ), no movement is detected, and when (δ>γ), movement is detected. As described above, the selector 4 is switched by this detection signal.

Further, as the Y/C separation circuit 2 using a frame memory, a structure using one frame memory 10 as shown in FIG. 4 can also be used.
The signals F ₁ and F ₂ of different frames obtained by the frame memory 10 are supplied to the adder circuit 20, and the output of the adder circuit 20 is halved by the multiplier 21 to obtain a luminance signal output. It will be done. Furthermore, a chroma signal output is obtained by supplying the phase-inverted version of the signal F ₁ and the signal F ₂ to an adder circuit 22, and halving the output of the adder circuit 22 by a multiplier 23.

An example of an adaptive control type two-dimensional Y/C separation circuit 3 is shown in FIG. In FIG. 5, a digital composite color television signal is supplied to an input terminal indicated at 17. Further, 25 indicates a bandpass filter for passing the band of the chroma signal,
The output of this bandpass filter 25 is supplied to a series connection of a 1H (one horizontal interval) delay circuit 26 and a 1H delay circuit 27, and the signals B, M, and T extracted from between these stages are sent to multipliers 28 and 29. , 30, respectively, and is also supplied to a correlation determination circuit 31 as shown by the broken line. A predetermined coefficient, for example 1/2, is supplied to the multipliers 28, 29, 30, and the output thereof is supplied to the adder circuit 36 via switch circuits 33, 34, 35. This switch circuit 3
3, 34, and 35 are controlled by the determination output of the correlation determination circuit 31. Then, the output of the adder circuit 36 is taken out as a chroma signal. Further, the composite color television signal is supplied to an adder circuit 39 via a delay circuit 37 and a 1H delay circuit 38 for correcting the delay in the bandpass filter 25. A reverse phase chroma signal C is applied to this adder circuit 39, and its output is taken out as a luminance signal.

The above-mentioned correlation determination circuit 31 uses the absolute values of the three signals to determine the presence or absence of correlation, controls the switch circuits 28, 29, and 30 based on the result of this determination, and performs the following processing. I am doing it.

If |T|≒|M|≒|B| holds true, then based on the formula, C=1/2 [M-1/2 (T+B)], Y=1/2 [M+1/2 (T+B)] | If only T|≒|M| holds true, based on the formula C=1/2(MT), Y=1/2(M+T) If only |M|≒|B| holds true, based on the formula Based on C = 1/2 (M - B), Y = 1/2 (M + B) When |T|, |M|, |B| do not have an approximately equal relationship at all (a) In the signal M When sample data separated by one cycle of color subcarriers has approximately the same value, it is regarded as a color difference signal transient, and C=M, Y=0.

(b) When the sample data of the color subcarrier in signal M that is separated by one period are not nearly equal,
Regarded as a luminance signal transient, C=0, Y=M.

In the correlation determination circuit 31 described above, the determination reference value ε
is supplied from terminal 32. For example, (|T|−|
When the relationship M|<ε) holds, (|T|≒
|M|).

The above-mentioned adaptive control type two-dimensional Y/C separation circuit 3
This method performs correlation determination by comparing absolute values. Unlike this method, if the chroma signals of adjacent lines in the same field have the same level and a phase difference of 180 degrees, it is determined that there is a correlation, and otherwise it is determined that there is no correlation. has been proposed by the applicant.
Figure 6 shows the procedure for determining the presence or absence of this correlation, and shows the sum (or difference) of signals on adjacent lines.
It is determined whether the absolute value of is smaller than the reference value ε, and if this relationship holds, it is determined that there is a correlation, and if not, it is determined that there is no correlation.

FIG. 7 shows the configuration of a two-dimensional Y/C separation circuit 3 designed to perform such a correlation determination. Basically, a Y/C separation circuit similar to that shown in FIG. 5 is constructed, and a correlation determination circuit 31 shown surrounded by a broken line performs the determination shown in FIG. 6.

The calculation circuit 40 of the correlation determination circuit 31 calculates |M+T
The calculation circuit 41 sequentially performs the calculation of |M+B| for each sample. These calculation outputs are supplied to level comparison circuits 42 and 43 and compared with a reference value ε from terminal 32. This level comparison circuit 42
The comparison outputs a,,b, of 43 correspond to those shown in the flowchart of FIG.
Four decision outputs ab, , a, and b are formed. The meaning of each of these judgment outputs is |M+T
The output a that becomes high level when |>ε holds, and the output that becomes high level when it does not hold,
The output that becomes high level when |M+B|>ε is established is designated as b, and the output that becomes high level when this is not the case is designated as b, so that the outputs are as shown below.

ab... There is no correlation between the signals T, M, and B of the three lines.

...There is a correlation among all of the signals T, M, and B.

a... There is a correlation between signals M and B.

b... There is a correlation between signals M and T.

The above-mentioned judgment output is output from the AND gates 48, 49,
50 and 51. The output of the AND gate 48 is supplied to an AND gate 52, and the signal passed through the AND gates 49, 50, 51, and 52 is supplied to an OR gate 53, from which a chroma signal C appears.

In order to prevent the comparison operation of the level comparison circuits 42 and 43 from being affected by noise, the comparison result is made valid when the same comparison result occurs continuously for a plurality of sample data. You can do it like this.

Further, the adder circuits 54 and 55 generate the difference signal M
-B, M-T are formed, dividing circuits 56 and 57
Therefore, it is reduced to 1/2, so 1/2 (MB), 1
/2 (M-T) signals are formed, and the addition circuit 58 and division circuit 59 further form 1/2 [M-1/2 (T+B)] signals. The respective outputs of these divider circuits 56, 57 and 59 are supplied to the aforementioned AND gates 49, 50 and 51.

Also, the signal M appearing at the output of the 1H delay circuit 26
is supplied to the AND gate 52 and also to the delay circuit 60 and the arithmetic circuit 61. When the output of this delay circuit 60 is M', the arithmetic circuit 61 calculates |M-M'|, and the output is supplied to the level comparison circuit 62 and compared with the reference signal R from the terminal 63. . The delay circuit 60 has a delay amount of one cycle of the color subcarrier,
M' is a signal one cycle before M. When the levels of signals M and M' are almost equal (|M-M'|<
A comparison output is generated from the level comparison circuit 62 that is high level when the signal is R (R) and low level otherwise, and is supplied to the AND gate 48.
Therefore, when it is determined that ab is at a high level and there is no correlation among the three signals, and when the levels of M and M' are approximately equal, the signal M is is taken out as a chroma output, and the luminance signal Y from the adder circuit 39 is set to zero. Similarly, when it is determined that there is no correlation between the three signals, and M and
When the levels of M' are not substantially equal, the AND gate 52 is turned off, the chroma signal C is set to 0, and the luminance signal Y is set to the signal M.

Further, when the output of the judgment circuit 31 is at a high level, the output of the division circuit 59 is selected as the chroma signal, and when the judgment output a is at a high level, the output of the division circuit 56 is selected as the chroma signal, and the judgment The output of the divider circuit 57 is selected when the output b is at a high level.

In the above explanation, the presence or absence of correlation was determined using the signals of three adjacent lines of the same field, but the presence or absence of correlation was determined using the signals of four adjacent lines. You can also do it.

As shown in FIG. 1, the separation evaluation circuit 8 combines the outputs of the two-dimensional Y/C separation circuit 3 using an adder circuit 64 to form a composite color television signal.
The separation error β is detected by supplying this phase-inverted signal and the signal F ₁ of the first frame taken out from the Y/C separation circuit 2 using a frame memory to the addition circuit 65. Y/C separation circuit 2
The composite color television signal obtained by combining the outputs of is the signal _F2 of the second frame. However, since there is no movement, if the two-dimensional Y/C separation circuit 3 operates well, (F ₁ ≈F ₂ ), and the separation error β becomes extremely small.

This separation error β occurs for each sample data of the digitized signal, and
The signal is passed through a multiplication circuit 66 and a digital integration circuit 67. The digital integration circuit 67 includes a register 68 and an adder circuit 69, and is configured to add data input by the adder circuit 69 and data fielded back from the register 68. This digital integration circuit 67 generates the effective value of the separation error β.

The reference value generation circuit 9 includes registers 70, 71, 7
2, 73, a comparison circuit 74, a memory 75 for storing a plurality of reference values, a selector 76, and a control circuit 77. This selector 76 and control circuit 7
The detection signal of the motion detection circuit 6 is supplied to the motion detection circuit 7. The selector 76 selects the memory 75 when there is no movement.
Selects the reference value read from and supplies it to the two-dimensional Y/C separation circuit 3, and when there is movement, register
Select the reference value stored in 2D Y/C
Supplied to separation circuit 3. Further, the control circuit 77
Loading data into each register, memory 7
5. Controls the read operation of 5.

When there is no movement, the control circuit 77 reads from the memory 75 a reference value ε that changes sequentially at appropriate intervals, for example, every frame, and supplies it to the two-dimensional Y/C separation circuit 3 via the selector 76. Error β of Y/C separation performed using the reference value at this time
is detected by the separation evaluation circuit 8. The effective value of this error β for one frame is taken into the register 70. The register 70 takes in the separation error value slightly before the next reference value is read out from the memory 75 and stores it.

The register 71 stores the smallest separation error value among the separation error values corresponding to each of the past reference values. That is, the new separation error value from register 70 and the minimum value stored in register 71 are compared by comparator circuit 74, and only if the new separation error value is smaller is this added to register 71. It is captured. This register 71 is filled with a large value at the beginning of the evaluation of a series of reference values. Furthermore, the contents of the register 71 are as described above.
When updated, the reference value appearing in the selector 76 is stored in the register 73.

By performing the above operations for each of the series of reference values, the minimum separation error value is stored in register 71.
At the same time, the reference value that caused this minimum separation error value is stored in the register 73. Then, the contents of the register 73 are transferred to the register 72, and when there is movement, the reference value stored in the register 72 is selected by the selector 76. The two-dimensional Y/C separation circuit 3 uses this reference value ε to determine the correlation as described above. Therefore, the reference value ε is a value that optimizes the separation operation of the two-dimensional Y/C separation circuit 3.

"Application example" In order to evaluate the separation performance of the two-dimensional Y/C separation circuit 3, a composite color television signal for testing is supplied to the two-dimensional Y/C separation circuit.
An optimal reference value may be set. In this case, there is no need to combine a Y/C separation circuit using a frame memory.

Furthermore, in addition to the reference value for line correlation determination, the band of a bandpass filter included in the two-dimensional Y/C separation circuit may be optimized. The band of this bandpass filter is set to, for example, 0.5 MHz as a reference, and is switched by a control signal, and adjusted to a band where the separation error is minimized.

Further, the separation evaluation circuit and the reference value generation circuit may be configured using a microcomputer.

Furthermore, the present invention can also be applied to PAL color television signals.

"Effects of the Invention" According to the present invention, it is possible to automatically set the reference value used for determining the correlation and the band parameters of the bandpass filter to values that minimize the separation error. 2D Y/
A C separation circuit can be realized. Further, according to the present invention, which combines the two-dimensional Y/C separation circuit and the Y/C separation circuit using a frame memory and uses both selectively depending on the presence or absence of movement, blurring does not occur. Y/C, which eliminates the insufficiency of Y/C separation by a two-dimensional Y/C separation circuit.
C separation can be carried out.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of one embodiment of this invention, and FIGS. 2 and 4 are Y/C using frame memory.
A block diagram of an example of a separation circuit and another example, FIG. 3 is a block diagram of an example of a motion detection circuit, and FIG. 5 is a block diagram of an example of a motion detection circuit.
A block diagram of an example of a dimensional Y/C separation circuit, FIGS. 6 and 7 are a flowchart and a block diagram used to explain a two-dimensional Y/C separation circuit. 1...Input terminal for digital composite color television signal, 2...Y/Y using frame memory
C separation circuit, 3...2-dimensional Y/C separation circuit, 5Y
... Luminance signal output terminal, 5C ... Chroma signal output terminal, 6 ... Motion detection circuit, 8 ... Separation evaluation circuit, 9 ... Reference value generation circuit, 10, 11 ... Frame memory, 25 ... Bandpass filter, 2
6, 27...1H delay circuit, 31...Correlation judgment circuit, 32...Input terminal for reference value for correlation judgment.

Claims (1)

[Claims] 1. In a Y/C separation circuit that separates a luminance signal and a chroma signal from a composite color television signal, the presence or absence of line correlation is determined by comparing the sum or difference of the composite color television signals of different lines with a reference value. The judgment circuit and the output of this judgment circuit when there is a line correlation extract the signal passed through the comb filter as a chroma output, and the output of the judgment circuit when there is no line correlation extract the signal from the comb filter. a switching circuit that extracts the signal that is not passed through as a chroma output;
A circuit that removes the chroma output from the composite color television signal and extracts it as a luminance output, evaluates the Y/C separation error from the luminance output and the chroma output, and sequentially determines a reference value for determining the presence or absence of the line correlation. and a signal generating circuit that selectively provides a reference value that minimizes the separation error by changing the signal generation circuit.