JPH0714219B2 - Luminance signal Color signal separation device - Google Patents

Description

The present invention relates to a device for separating a luminance signal and a chrominance signal of a composite video signal.

(Prior Art) As a conventional luminance signal / color signal separation device, for example, there is one shown in FIG. In FIG. 6, the composite video signal (Y _H + Y _L + C) is input to the terminal 38, and the high pass filter 39 passes the high pass component (Y _H + C). One of the high frequency components (Y _H + C) is supplied to the subtractor 44 and subtracted from the composite video signal (Y _H + Y _L + C) to obtain the low frequency luminance signal (Y _L ). Further, the high frequency component (Y _H + C) is supplied to the vertical comb filter 40 and the horizontal comb filter 41, and separated into a high frequency luminance signal (Y _H ) and a color signal (C) by the respective filters. To be done. Then, the separated high-frequency luminance signal (Y _H ) and color signal (C) are supplied to the switch circuit 43. Switch circuit 43
In the case where the vertical direction correlation is strong, the output of the vertical comb filter 40 is selected according to the correlation of the high frequency component (Y _H + C) determined by the correlation determination circuit 42, and the horizontal direction When the correlation is strong, the output of the horizontal comb filter 41 is selected. FIG. 7 shows a video signal, that is, a pixel, which is input to the above-described luminance signal / color signal separation device and which is sampled at 4 times F _sc . In FIG. 7, the video signals of the l−1 line, the l line, and the l + 1 line are shown, and τ is 1/4 · F _sc . Then, the above-mentioned luminance signal / color signal separation device performs separation as follows.

IF | G ₂₄ -G ₂₆ | ≦ | G ₂₂ -G ₂₈ |; Y = (G ₂₂ + 2G ₂₅ + G ₂₈ ) / 4 C = − (G ₂₂ -2G ₂₅ + G ₂₈ ) / 4 IF | G ₂₄ -G ₂₆ |> | G ₂₂ -G ₂₈ |; Y = (G ₂₄ + 2G ₂₅ + G ₂₆ ) / 4 C = − (G ₂₄ -2G ₂₅ + G ₂₆ ) / 4 As described above, G pixel In the case of determining the correlation between the horizontal direction and the vertical direction, the correlation between | G ₂₄ -G ₂₆ | and | G ₂₂ -G ₂₈ |

(Problems to be Solved by the Invention) However, in the above-described configuration, the determination of the correlation is made only by comparing the difference between the values of the pixels separated by one period of the subcarrier or the pixels separated by two horizontal lines. In consideration of the difference in sensitivity of human eyes to brightness and color and the difference in sensitivity to vertical and horizontal directions, the above-described correlation determination method has a problem that it does not necessarily match human visual characteristics.

Further, the determination circuit for performing the correlation determination and the filter for performing the luminance signal / color signal separation use three sample points, and in the filter in the vertical direction, the three sample points extend over three horizontal lines. Therefore, for example, l-
If the line and the l line have a correlation, but the l + 1 line has no correlation with other lines, and if the horizontal correlation is broken even within the l line, either the vertical direction filter or the horizontal direction filter is used. There is a problem that the separated luminance signal or chrominance signal deteriorates. Further, when noise is mixed in the video signal and the correlation is lost, there is a problem that deterioration due to noise directly appears on the screen.

In view of the above points, the present invention has a correlation determination unit that satisfies the visual characteristics of human eyes, is less affected by noise, and can perform correlation determination between any two pixels, and uses two or more sample points. It is an object of the present invention to provide a luminance signal color signal separation device that has a filter for performing luminance signal color signal separation according to the above, and performs luminance signal color signal separation correctly in any pattern.

(Means for Solving Problems) In order to solve the above problems, the present invention provides a high-frequency luminance signal and a chrominance signal of a composite video signal sampled with a clock that is an even multiple of a subcarrier synchronized with a burst signal. The difference between the pixel 1/4 cycle before the subcarrier and the pixel 1/4 cycle after the subcarrier of the currently processed pixel of the high-frequency video signal is calculated as the difference value between the currently processed pixel and It is a luminance signal color signal separation device that calculates the correlation value between pixels in the vicinity of the difference value, determines the correlation with the pixels in the vicinity, and performs the luminance signal color signal separation by the calculation with the pixel having the strongest correlation. Then, the above-mentioned difference value is added to and subtracted from neighboring pixels to calculate the correlation value of the color signal and the high-frequency luminance signal component, and the addition value and the subtraction value are respectively weighted. The multiplier and this weighted sum A correlation circuit consisting of a minimum value circuit that selects the minimum subtraction value, a multiplication circuit that weights the outputs of a plurality of correlation circuits that calculate the correlation value with neighboring pixels, and this weighted correlation A judgment circuit that selects the pixel with the smallest value and judges the pixel with the strongest correlation, an averaging circuit that calculates the average value of the output of this judgment circuit for the past fixed period, and this average value and the output of the judgment circuit are compared, and they differ greatly. In this case, a correlation and noise determination circuit including a second determination circuit that determines that the output of the determination circuit is noise, a memory that stores the minimum correlation value, and a luminance signal using a plurality of pixels in the vicinity. The output of the filter that uses the pixel with the strongest correlation is selected from the plurality of filters that perform color signal separation, and if noise is mixed in the pixel being processed, it is selected one clock before. Characterized by comprising a switch circuit for selecting the output of the filter or filter that is set in advance.

(Operation) According to the present invention, since the correlation value of the high-frequency luminance signal and the color signal is independently calculated and appropriately weighted by the above-described configuration, the correlation is determined, so that it matches the visual characteristics of human eyes. The correlation determination can be performed, and the correlation determination of two adjacent pixels can also be performed. Further, in the case of the correlation determination, since the correlation is determined using the history of the correlation already determined, it is possible to reduce the error of the correlation determination due to noise. Further, since the filter for performing the luminance signal color signal separation performs the separation using two or more sample points, the luminance signal color signal separation can be performed correctly even in the case of a picture that changes drastically by combining with the above correlation determining unit. be able to.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of a luminance signal / color signal separation device of the present invention. In FIG. 1, 1 is a terminal for inputting a composite video signal, 2 is a high-pass filter for passing a high-frequency video signal, and 3 to 7 are separated from each other by a half week period of a subcarrier of the sampled high-frequency video signal. An arithmetic circuit for calculating the difference between two values, 8 to 11 are arithmetic circuits (k) (k = 1 to n)
And the output of the arithmetic circuit (0) are added and subtracted to perform more appropriate weighting, and the smaller one of the weighted addition value and subtraction value is output, and 12 is the correlation circuit (k) (k
= 1 to n), a multiplier circuit that appropriately weights the outputs, and 13 has already selected the minimum value of the output of the arithmetic circuit (k) (k = 1 to n) that is appropriately weighted by the multiplier circuit 12. Correlation and noise determination circuit to select based on the history of the minimum value, 14 is a memory that stores the output of the correlation and noise determination circuit 13, 15 is an averaging circuit that averages the output of the memory 14, 16 to 19 are A filter that performs color signal separation based on the correlation with neighboring pixels, 20 selects the output of the filter that uses the pixel selected by the correlation and noise determination circuit 13 from the output of the filter (k) (k = 1 to n) Switch circuit,
Reference numeral 21 is a subtractor for subtracting a color signal from the composite video signal, 22 is a terminal for outputting a color signal, and 23 is a terminal for outputting a luminance signal.

The operation of the luminance signal / color signal separation device of the present embodiment configured as described above will be described below.

The composite video signal (Y _L + Y _H + C) sampled at an even multiple of the sub-carrier is supplied to the terminal 1, and the high pass filter 2
Causes the high frequency image signal (Y _H + C) to pass. The high frequency image signal (Y _H + C) is supplied to the arithmetic circuit (k) (k = 0 to n),
A difference value between two values separated by a half cycle of the subcarrier is calculated. Here, the arithmetic circuit (0) is the pixel G _{0 of} interest.
The difference value of is calculated, and the calculation circuit (1), the calculation circuit (2),
The arithmetic circuit (n) is a circuit for calculating the difference value between the neighboring pixels G ₁ , G ₂ , ..., Gn. Output ES0 of arithmetic circuit (0)
And the output ESk (k = 1 to 1) of the arithmetic circuit (k) (k = 1 to n)
n) are respectively supplied to the correlation circuit 1 (l = 1 to n). In the correlation circuit 1 (l = 1 to n), first, the following calculation is performed to obtain the correlation value PESl of the color signal and the correlation value DESl of the high frequency luminance signal.

PESl = | (ESl + ES0) | DESl = | (ESl-ES0) | (l = 1 to n) Furthermore, weighting based on the difference in luminance and color sensitivity of human visual characteristics is applied to correlation values PESl and DESl. To do. If the correlation values obtained as a result are PES'l and DES'l, the smaller correlation value PES'l and correlation value DES'l, that is, the stronger correlation value is selected and output. If the value with the stronger correlation is set to OSKl, the correlation circuit (l) indicates that the correlation between the high-frequency luminance signal correlation value and the color signal correlation value with the neighboring pixel Gl is correlated. The OSKl (l = 1-n) having the stronger value is output. In the arithmetic circuit 12, each pixel Gl (l =
1-n) are weighted according to the difference in the sensitivity of the human eye to the positions of the correlation circuits (1) (1-n). Then, in the correlation and noise determination circuit 13, the one having the smallest value is selected from the outputs of the correlation circuits (l) (1 to n) weighted in the multiplication circuit 12, and the number l of the pixel Gl is selected. Is output to the switch circuit 20. The selected minimum correlation value output from the correlation and noise determination circuit 13 is supplied to the memory 14, and the averaging circuit 15 further calculates the average of the already selected minimum correlation values. This average value is supplied to the correlation and noise determination circuit 13, and when the minimum correlation value is selected, if the difference between the selected value and the average value exceeds a preset value, the selection is incorrect. In this case, the preset pixel number or the pixel number selected one clock before is output to the switch circuit 20. In this way, even when noise is mixed in the image signal, its influence can be minimized. In the filter (l) (l = 1 to n), pixels Gl and G respectively
Color signal separation is performed using ₀ . This filter (l) (l =
1 to n) are comb filters. The output of the filter (l) (l = 1 to n) is supplied to the switch circuit 20, where the output of the film using the pixel having the strongest correlation with the pixel G ₀ selected by the correlation and noise determination circuit 13 is selected. And output from the switch circuit 20. The color signal thus obtained is supplied to the terminal 22 as a color signal, and is also supplied to the subtractor 21, where it is subtracted from the video signal and the brightness signal is supplied to the terminal 23. As described above, according to the present embodiment, the correlation between the luminance signal and the chrominance signal in the high-frequency video signal is independently examined by providing the arithmetic circuit and the correlation circuit, and the correlation between the pixels is determined by the smaller one of them. Since the sex is checked, the correlation can be determined without error even in a video signal in which a large number of high-frequency luminance signal components exist near the frequency of the color signal. Further, by providing the correlation and noise determination circuit, the memory, and the averaging circuit, even if noise is present in the video signal, malfunction due to it can be eliminated.

FIG. 2 shows the arithmetic circuit in FIG. 1 in detail, in which 25 is a subcarrier half-cycle delay circuit (τ = 1/4 · FS).
C) and 26 are subtractors. In this arithmetic circuit, terminal 2
The difference between the signals supplied to 4 and separated by 2τ period is calculated and output from the terminal 27.

FIG. 3 shows the correlation circuit in FIG. 1 in detail, in which 29 is a subtractor for subtracting the signal from the terminal 32 from the signal from the terminal 28, and 33 is the signal from the terminal 28 and the signal from the terminal 32. Adders for adding signals, 30 and 34 for absolute value circuits, 31 and 35 for appropriate weighting multipliers, and 36 for a minimum value (MIN) circuit for selecting the minimum value. If the signal supplied to the terminal 28 is ES and the signal supplied to the terminal 32 is ESO, | ES in the subtractor 29, the adder 33, and the absolute value circuits 30 and 34
−ESO | and | ES + ESO | are calculated, and after appropriate weighting is performed by multipliers 31 and 35, respectively, the smaller value is selected in the minimum value circuit and the correlation value is set to the terminal.
It is output to 37.

The operation of the arithmetic circuit and the correlation circuit shown in FIGS. 2 and 3 will be described with reference to actual waveforms in FIGS. 4 (a) and 4 (b). FIG. 4 (a) shows three horizontal lines l.
-1,1, l + 1 shows the color signal on line 1
The color signals of line 1 and line l have the same magnitude, but l
The color signal of +1 line is 1/2 of the color signal of other lines.
In addition, S ₀₁ is the sample point of interest currently being processed,
S ₀₂ and S ₀₃ are sample points in the vicinity, S ₁ , S ₂ , S ₃ , S ₄ , S ₅ and S ₆ are sample points used for correlation judgment, and sample points on the same line are separated by τ period. . When the signal of FIG. 4 (a) is supplied to the arithmetic circuit, S ₁ -S ₂ (= DC ₁ ), S ₃ -S ₄ (= DC ₂ ), S
₅ -S ₆ (= DC ₃ ) calculation result is output. As shown in the figure, DC ₁ , DC ₂ , and DC ₃ are shown so that the arrows are upward when the values are positive and downward when the values are negative. Thus, there is a relationship of DC ₁ = -DC ₂ and DC ₃ = DC _1/2 .

Now, when DC ₁ and DC ₂ are supplied to the correlation circuit, DC ₁ and
DC ₂ is input to terminal 32 and the following operations are performed.

_{MUL1 = | DC 1 + DC 2} | * CO1 DIF1 = | DC 1 + DC 2 | * CO2 CO1 and CO2 is the factor of the weighting, both for simplicity assumed to be 1. Since the DC ₁ = there is a relationship between the _{-DC 2, MUL1 = 0, DIF1} = 2 | DC 1 | become. And MUL1 and DIF1
, The minimum value of MUL1 (= 0) is selected by the minimum value circuit and output. Next, when DC _2, DC ₃ is supplied to a correlation circuit, DC ₃ to the terminal 32, the terminal 28 DC ₂ is input, the same operation is performed as follows.

_{MUL2 = | DC 2 + DC 3} | * CO1 DIF2 = | DC 2 -DC 3 | * because it is _{CO2 DC 3 = -DC 2/2} , MUL2 = | DC 2 | / 2, DIF2 = 3 | DC 2 | It becomes / 2, and MUL2 = | DC ₂ | / 2 is selected and output by the minimum value circuit. Where MUL1
When Comparing the MUL2 MUL1 = 0, MUL2 = | DC 2 | a / 2, MU
It shows that L1 has a stronger correlation, which is l-1
It correctly shows that the color signals of the line and the l line have the same magnitude but the magnitudes of the color signals of the l + 1 line and the l line are different. As described above, the adder 33, the absolute value circuit 34, and the multiplier 35 in the correlation circuit correctly calculate the correlation value of the color signal.

Next, in FIG. 4 (b), l-line, l-line, l + 1
The high-frequency luminance signal in the line is shown, and the magnitudes of the luminance signals of the l-1 line and the l line are the same, but l + 1
The magnitude of the luminance signal of the line is half that of the other lines. S ₀₄ is the sample point of interest currently being processed, S ₀₅ and S ₀₆ are neighboring sample points, and S ₇ , S ₈ , S ₉ , and S
₁₀ , S ₁₁ and S ₁₂ are sample points used for correlation determination, and sample points on the same line are separated by 2τ period.

When the signal of FIG. 4 (b) is supplied to the arithmetic circuit, S ₇ -S ₈ (= D
C ₄ ), S ₉ -S ₁₀ (= DC ₅ ), S ₁₁ -S ₁₂ (= DC ₆ ) calculation results are output. As shown in the figure, DC ₄ , DC ₅ , and DC ₆ are arrows pointing upward if the polarity is positive and pointing downward if the polarity is negative. This DC _4, DC _5, DC ₆ is as implicated that _{DC 4 = DC 5, DC 6} = DC 4/2. When DC ₄ and DC ₅ are fed to the correlation circuit, the terminals
DC ₄ is input to 28 and DC ₅ is input to terminal 32, and the next operation is performed.

_{MUL3 = | DC 4 + DC 5} | * CO1 DIF3 = | because it is _{* CO2 DC 4 = DC 5,} MUL3 = 2 | | DC 4 -DC 5 DC 5 |, DIF3 = 0 becomes the minimum value circuit DIF3 (= 0) is selected and output.

Then DC ₅ and DC ₆ are applied to the correlation circuit, and DC ₆
However, DC ₅ is input to the terminal 32 and the same operation is performed.

_{MUL4 = | DC 6 + DC 5} | * CO1 DIF4 = | because it is _{* CO2 DC 6 = DC 5/} 2, MUL4 = 3 | | DC 6 -DC 5 DC 5 | / 2, DIF4 = | DC 5 | / 2
Next, DIF4 a minimum value circuit _{(= | DC 5 | / 2} ) is selected and output.

Now, comparing DIF3 and DIF4 here, DIF3 = 0, DIF4 =
| DC ₅ | / 2, indicating that DIF3 has a stronger correlation. This is because the luminance signals of the l-1 line and l line are the same, and the luminance signal of the l + 1 line is the same. Correctly shows that is different from that of the other lines.

As described above, the subtractor 29 in the correlation circuit, the absolute value circuit 30,
The multiplier 31 correctly calculates the correlation value of the luminance signal.
By using the arithmetic circuit and the correlation circuit in this way, the correlation value between the color signal and the luminance signal between the two pixels can be correctly calculated.

FIG. 5 shows the positions of neighboring pixels, the circles in each line indicate the pixels, and the interval between the pixels is τ period. FIG. 5 (a) shows l-1 line, l line, l + 1.
This is an example of the case of using the signals of three horizontal lines of lines, and as shown in the figure, 8 of G ₁ , G ₂ , G ₃ , G ₄ , G ₅ , G ₆ , G ₇ , and G ₈ are used.
Pixels in one neighborhood can be selected, and the number of pixels can be increased. G ₀ is the pixel of interest,
H ₁ , H ₂ , ..., H ₁₆ are the target pixel G ₀ and neighboring pixels G ₁ to G ₈
Pixels used to calculate the difference value of

In an example where FIG. 5 (b) is of increased to five the number of horizontal lines, a fifth diagram G ₉ of l-2 line and the l + 2 lines compared to _{(a), G 10, G} 11, G 12 12 pixels can be selected in total, and both the number of pixels and the number of horizontal lines can be increased, and H ₁₇ , H ₁₈ , ..., H ₂₄ are neighboring pixels.
It is a pixel used to calculate the difference value of G ₃ , G ₁₀ , G ₁₁ , and G ₁₂ .

(Effects of the Invention) As described above, according to the present invention, when the correlation determination of two pixels is performed, the luminance signal correlation value and the chrominance signal correlation value are independently calculated, and appropriate weighting is performed. The minimum value is used as the correlation value, and the correlation determination is performed after weighting the multiple correlation values thus obtained according to the position of the pixel, so the correlation determination that matches the visual characteristics of the human eye must be performed. You can In addition, when performing the correlation determination, since the determination is performed using the history of the correlation that has already been determined, the influence can be reduced even when the correlation is broken due to mixing. Since the filter used for the luminance signal color signal separation uses two or more sample points to perform the separation, the luminance signal color signal separation can be performed correctly by combining with the above-described correlation determining unit even when the pattern changes drastically. it can.

[Brief description of drawings]

FIG. 1 is a block diagram of a luminance signal / color signal separating apparatus according to an embodiment of the present invention, FIG. 2 is a block diagram of an arithmetic circuit, FIG. 3 is a block diagram of a correlation circuit, and FIGS. 4 (a) and 4 (b). ) Is a waveform diagram for explaining the operation of the arithmetic circuit and the correlation circuit, FIGS. 5 (a) and 5 (b) are diagrams showing the positions of pixels used for luminance signal color signal separation, and FIG. 6 is a conventional luminance signal color. FIG. 7 is a block diagram of the signal separation device, and FIG. 7 is a diagram showing the positions of the pixels used in the device of FIG. 2 ... High-pass filter, 3-7 ... Arithmetic circuit, 8-11
...... Correlation circuit, 12 …… Multiplication circuit, 13 …… Correlation and noise judgment circuit, 14 …… Memory, 15 …… Averaging circuit, 16 to 19
...... Filter, 20 ...... Switch circuit, 25 …… 2τ delay circuit, 30,34 …… Absolute value circuit, 31,35 …… Multiplier, 36 …… Minimum value circuit.

Claims (1)

[Claims] 1. A subcarrier of a currently processed pixel of a high-frequency video signal composed of a high-frequency luminance signal and a chrominance signal of a composite video signal sampled at a clock that is an even multiple of a sub-carrier synchronized with a burst signal. The difference between the pixel before the cycle and the pixel after the 1/4 cycle of the subcarrier is calculated as the difference value of the pixel currently being processed, and the correlation value between the pixels in the vicinity of this difference value is calculated, and the pixel in the vicinity A luminance signal color signal separation device that determines the correlation with the pixel having the strongest correlation and performs the luminance signal color signal separation by calculating the color signal by performing addition and subtraction with the neighboring pixels of the difference value. The adder and the subtracter for calculating the correlation value of the high-frequency luminance signal component, the multiplier for weighting the addition value and the subtraction value, and the minimum value of the weighted addition value and the subtraction value Correlation circuit consisting of minimum value circuit to be selected, and A multiplication circuit that weights the outputs of a plurality of correlation circuits that calculate correlation values with neighboring pixels, and a determination circuit that selects the smallest weighted correlation value and determines the pixel with the strongest correlation And an averaging circuit that obtains the average value of the output of this determination circuit for the past certain period of time, and this average value and the output of the determination circuit are compared, and if there is a large difference, it is determined that the output of the determination circuit is noise A correlation and noise determination circuit including a circuit, a memory that stores the minimum correlation value, and a pixel that has the strongest correlation among a plurality of filters that perform luminance signal color signal separation using a plurality of pixels in the vicinity. If the output of the filter used is selected and noise is mixed in the pixel being processed, the switch that selects the output of the filter selected one clock before or the output of the preset filter is selected. Luminance signal and color signal separating apparatus characterized by comprising a latch circuit.