JPS607287A - Y/c separating circuit - Google Patents

Abstract

PURPOSE:To realize a Y/C separating circuit with high quality without deterioration in the resolution while suppressing a residual image by using the addition of a detected output SMD in response to the movement of a picture between frames before and after to a frame difference signal SF as a chrominance carrier signal by means of an adder. CONSTITUTION:A movement information signal SM is applied to a serial parallel converting circuit 30, and a frame difference signal with respect to four picture element points parted by 2 sampling clocks at a line and the next line respectively appears at th output. This frame difference signal is fed to an orthogonal converting circuit 31, where a pattern of a frame difference picture in a region of the four picture element points is discriminated, four outputs signals having a level depending on the pattern and the four output signals are applied respectively to nonlinear circuits 32A-32D. The output of the nonlinear circuits 32A-32D is fed to a parallel serial converting circuit 34 via an inverting circuit 33, and the output of the circuit 34 is fed to an adder 16 as a movement detecting output SMO.

Description

[Detailed description of the invention] Industrial applications This invention applies to Y/C molecules (11 circuits).

Background technology and the time between si'l + 1 point Y/C min 1'I (\, as a circuit,';, iS l 1
When using a comb filter supported by a one-frame delay Zushi (1) and a pair of adders (21, (3) as shown in column 1-9), the video signal fed to the zigzag (4) ( If the signal containing the luminance signal Y and the transmitted color signal C) is a still image, the video signal delayed by one frame and the current video signal are the same, so the luminance signal Y and the transmitted color signal C are the same. can be completely separated.

However, when a video signal for a moving image is supplied to the wheel (4), the delayed video signal and the current video signal are different, so complete Y/C separation cannot be performed, and afterimages may occur. be.

On the other hand, as shown in FIG.
/C minute. In the t circuit σω, when moving a moving image, 7. jli Since the high frequency component of the nail signal Y remains, the high frequency component of the luminance signal Y output from the adder (21J) is
There is a risk that the trade quality will deteriorate even a little.

In this way, in the conventional method, both the suppression of afterimages and the prevention of n1 production from becoming gj are satisfied at the same time (the 11th generation is not folded and 1I is not folded).

Purpose of the Invention Therefore, the present invention proposes a high-quality Y/C component P:(k circuit) that suppresses afterimages and does not cause deterioration of M image quality.

SUMMARY OF THE INVENTION Therefore, in the present invention, the input color video signal) V-
The motion detection circuit has a comb filter that forms a frame difference signal, a motion detection circuit that detects the motion of the human image, and an addition circuit that adds the motion detection output and the frame difference signal. A circuit for forming a motion information signal that obtains motion information, and this motion! 1 out of the j'j news signals
Only when there is little arrogance, the Y/C component 61F circuit is constructed with the non-bell-shaped characteristic. In this case, by using the output obtained from the Goku-kan power circuit as the P/IJ detection output, the carrier color signal is divided by t(1) from the adder.

Embodiment Next, an example of the Y/C portion Pr9 circuit according to the present invention will be explained in detail with reference to FIG. 3 and subsequent figures.

In FIG. 3, (marked) is a one-frame delay element, and this and an adder +121 form a comb filter (13) to form a frame difference signal SF. That is, the frame difference signal SF is supplied to the terminal (14). The video signal (Y0C) is delayed: (
; After being delayed by 17V-times in child (11), this delayed image f7'i, (ii (YF-CF) and the current image signal (YF-CF)
(e) is added to form a frame difference signal SF. This frame difference signal SF is expressed by the following equation.

sF =y-+-C-(YF-CF)...
(IJ) In this invention, the adder (16) adds the detection output SMO corresponding to the movement of the image between the preceding and following frames to the frame difference A1''F8F, and then calculates the sum of the frame difference A1''F8F.
, C' and use it as 7. The α force is its output terminal. The purpose of using Itenyo Deba 8MO is to suppress afterimages and achieve perfect Y/C nI1. In addition to realizing 1.
This is to prevent deterioration of 94 degrees in the evening.

The carrier color signal C' is transferred to the adder (1).
(191
indicates the first way out.

The detection circuit f2 (11 detects the movement between frames, and also outputs the detection output (motion position IR signal))
This is provided to consider predetermined input/output characteristics for Sr, q, and the detection output after the characteristics are assigned is SM, 0.

The motion information signal SM can be expressed as follows.

SM=Y-4'4-(YF-CF)...(2)
Here, CI indicates a carrier color signal whose phase has been inverted by the (1) inversion port P8 (211, which will be described later).Furthermore, the detection output 8MO is divided by the following equation.

Sp, U0-8 to 1 = Y'-C'(-(Y'F-CF) ......(3
) (3), the output SMO that was lost in the equation is the 10th bar 1.
Input and output as shown in! The 1st count is also +J'-15. Therefore, the output of the power unit (16), 1', is 1 as follows.

...... (4) Here, when there is a lot of movement information (when Spa is small), the detection output 8wro becomes an output proportional to the movement 11" i-information signal SM, so pp; as shown in Figure 10. As shown in FIG. Carrier telegram when there is a lot of information - No. C, c
= c -4-Ci (6) Since this carrier color signal CIC does not include any signals processed between frames, no residual color occurs.

On the other hand, when the motion information is non-'7i ((5-1-0), the input/output t]) in FIG.
As if 1jil is 1゛, SMOζO Therefore, 17 color sending signal C' at this time is C=Y-YF
+ (C+CF) -c -4-cF... (J) This carrier color signal d is subjected to two-dimensional processing; J-t, /, since it does not contain two signals, the resolution is degraded. There's nothing to do.

In a section where it is difficult to distinguish between a moving image and a still image, the detection output 5rvxo of the non-retention type /lHi is obtained and combined with the frame difference signal SF (formula (4) decoy). in this case,
The motion detection circuit (20) is one that has undergone orthogonal transformation as described in 5 below, and when there is little motion information, the input is sufficiently suppressed before being output. Since motion information can be detected and the nonlinear motion detection output SP, 1° is synthesized, there is very little influence on image quality.

In order to achieve the above-mentioned operation, a motion detection circuit (2(
1) is accomplished by the motion information signal S'M forming circuit (2OA) and the characteristic imparting circuit FW5 (20J3).

The movement ('iV signal forming circuit (20A) is provided with a phase inversion circuit ('21J) that inverts the phase of only the carrier color signal C included in the current image protection signal (Y+C), and the inverted image signal ( Y - CI) is supplied to the adder (22!) and added to the delayed video signal (YF + CF) as shown in the figure. This addition output SM calculates the magnitude of the image motion information between frames. This is the Igogo (motion information signal) that is displayed.

The characteristic imparting circuit (20B) applies the motion 'l'i'j to the motion information signal only when there is little motion information as described above.
This is to provide an output characteristic such that the level of the information signal 8M is suppressed and output (see FIG. 10). This characteristic assignment is performed over the entire band of the video signal.

In order to provide uniform characteristics over the entire band and to effectively remove noise contained in parts with little motion information and impart the above characteristics to signals containing only motion information, this invention uses Hadamard transform, etc. This is done using an orthogonal converter circuit.

Note that the input video signal (Y+C) is a digital video signal obtained by P (J4 modulation) of the analog video signal in order to facilitate delay and arithmetic processing. □4) Color sub (general delivery
The video signal (Y-1-C), which is sampled at a frequency Iq three or four times that of
Next supplied. Therefore, the motion information signal SM is also a serial code digital signal.

The motion information signal Spa is supplied to the serial-to-parallel converter circuit C (11), so that, as will be described later, its output contains only 2 sampling clocks for each of the new line and the next line. A 7-frame multiplication signal regarding ii'ii X points (sampling points) appears.This frame difference signal is supplied to the orthogonal transform circuit 61). The pattern of the frame difference signal of the slope of the point is determined, and four output signals of a level determined according to the pattern are generated. ) are supplied respectively.

Non-1f4J! The outputs of the shape circuits (32A) to (32D) are supplied to the parallel p-column conversion circuit (3 (1) via the inverse conversion circuit (33) of the pedestal described later, and this output is used as the motion detection output f,
SP, q, and o are supplied to 161.

An example of the color signal inversion circuit (21) is shown in FIG.

This is an IH (1 horizontal period) delay circuit (41, (++)
adder (42) and 4-level attenuator G8 and subtracter (1
A) and a bandpass filter (45) are used to extract a carrier color signal having twice the level of the original carrier color signal and having an opposite phase.1. ', 15+, pjE circuit ((transparent to the nationality signal and carrier color signal via fil.

By adding it to the adder (471), only the position a of the color signal to be added is inverted from the color video signal (y-
c4).

FIG. 5 shows an example configuration of the serial-parallel conversion circuit (301 and orthogonal conversion circuit 01). The motion information signal SM supplied from the input terminal 60 in the form of a serial code is input to the field rli j7' by a delay circuit I5B of II+ and a delay circuit ('i21, e'i31) of two sampling clocks 2τ. It is converted into frame difference data at a total of four pixel points, two pixel points separated by two sampling clocks 2τ of each of the two 2-in pixels.

As shown in FIG. 6A, l of the nth line; "i, l
Let the two pixel points X″′J be a++22, and
1) 83 r the two adjacent double shields on the right eye line
a 4, when the input terminal ()) is supplied with the difference data Sa4 at the pixel point a4,
As shown in FIG. 5, data Sa is output from the delay circuit 6.
3 is plastic.''11, Data Sa2 appears at the output of the IH delay circuit 1iL5511, and data 8al appears at the output of the delay circuit group.

These frame difference data Sal□Sa4 are supplied to the old AC conversion circuit ((+1).

This orthogonal transform circuit 31) has an output of 61 children (31
The conversion output S! shown below for each of a) to (31d)! ~S
This is a performance n1 circuit configured to generate 4.

Sr = Sal -1-Sa2 +sa3+ 5a4
82 == Sal −Sa2 + Sa3 ba4S
3= Sal -4-8a2- Sa3-5a484
= 5a1-Sa2-Sa3+Sa4 The level of the conversion output 5l-84 of this orthogonal transformation circuit (31j) corresponds to the frame difference video pattern of the (2X2) area formed by the six pixels a1 to a4.Sixth In each of FIGS. B to E, if the perspective area is set to a size of +1, and the other areas are set to a size of -1, the converted outputs 81 to S4 will be as shown below.

[Figure 6B: Pattern where the total distribution area is +1] 81 = 4, 5
2=83=84=0 [Figure 6 C: 1,000 vertical pattern] 52=4,
51=83=84=0 [Figure 6 D=+1 is the horizontal pattern] 53=4. '
81=82=84=0 [Fig. 6 E: +1 direction pattern] S4= 4
', S1 = 82 = 83 = Q Therefore, if +1 corresponds to the white level of the video signal and -1 corresponds to its black V-pel, then
It is possible to form a level-related conversion output 5l-84 according to the frame difference video pattern of the area represented by four pieces of data.

Here, if it is assumed that random dot-shaped noise components exist, such noise has components in all directions, so each of the output signals SI-84 has a level of J or less. -+ noise is included. In addition, since the movement information nokukuon has a specific pattern, it is expected that it will appear more often in any of the output signals 81-84. Therefore 1'+11 of movement
The level difference between 1!3 and noise becomes clear.

When there is little movement information, the S/N is inherently poor, so
If the motion information is simply detected from the frame difference signal, it is difficult to detect the motion component of the original image from this motion information.゛Motion information Jin - No. SM, but the movement is ``i''j fj! Since it is captured as a continuous change in motion information, noise is suppressed and the original motion information signal can be reliably detected.

For this reason, even if the motion information signal SF,l is indistinguishable between a still image and a moving image, the motion information detection accuracy is insufficient, and the information signal SF, which contains a lot of noise, may be mistakenly replaced with the original motion information signal. This eliminates the need to add the frame difference signal SF to the frame difference signal SF. The detection accuracy increases up to the order of orthogonal transformation.
View. Also, by performing orthogonal transformation, the probability of outputting a signal that is difficult to distinguish between a still image and a moving image as an original motion information signal is extremely low, and this makes it difficult to distinguish between a still image and a moving image. r f”! : It is possible to suppress the deterioration of the degree.

An example of non-reducible circuits (32A) to (32D) to which noise-suppressed output signals 81 to S4 are supplied is shown in FIG.
c Indicator o Non-n-type circuit (32A) to (32D) C;p, As shown in 1015>-1, an input with very little image movement that makes it difficult to distinguish between a moving image and a still image (1 '!The output signals S1 to S4 are outputted only when the information is inputted.Therefore, the nonlinear circuits (32A) to (32D) are configured to output the output signals S1 to S4 only when such specific information is input. A level conversion circuit (consisting of 551, an addition ohm, and a dexterity) outputs the characteristics as shown in FIG.

Conversion coefficient 1 (K=1−αl) of level conversion circuit 65)
SMl ......(!riα) is a constant, which is the gate between the output after orthogonal transformation and the level of the output after orthogonal transformation [A81 to S4 (motion information signal SM) and the transformation coefficient [z is shown in Fig. 8.The V-bell conversion 1 negative output obtained from the level conversion circuit Kasumi by this conversion coefficient K is By, It, SM =
K-8i...(ill) However, since i is 1 to 4, the output signals 331 to S4 are as shown in FIG.
level is +-La or higher, that is, movement (77C"j
α is large or zero in some cases, and otherwise lZ2 as shown in the figure.
Next, a π-shaped level conversion output SM is obtained.

As a result of adding this level conversion output SM and the output signal Si with the polarity shown, the non-0 type circuits (32A) to (32D)
The nonlinear output By has a non-glandular characteristic as shown in FIG.

Each of the non-fr Q type circuits (32A, ) to (32D)
He1 type Deba SYx +8'/2rsYs+sY< is the 1st
The signal is supplied to an inverse conversion circuit (33) as shown in FIG.

This inverse transform circuit (33) reverses the flow of the signal from the input to the output of the orthogonal transform circuit 4'll shown in FIG.
Tai 1 "7 generations, and its human power J, +i, l child (60a
) ~ (60d) nonlinear output S3 according to each pattern
't~Sy4 is supplied, its output has 1l
ljl 4 of the non-consolidated output at each of exemption points a1 to a4
Outputs 4Ya1 to 4Ya4 having twice the level appear.

The output of this inverse conversion circuit (33) is the LH delay circuit (631, (14
41 and 3-1's addition i'') device; 5) ~G force
Furthermore, the motion detection output 5xio at each pixel point is passed through the upper level fin H device (6!
1st order -, 14, will be. And this motion detection output SMO
is combined with the input color video signal in the adder (IB).

In this way, in the nonlinear circuits (32A) to (32D), 1
When it is a static image or similar motion information, it is almost completely suppressed, making it impossible to distinguish between a still image and a moving image. It has some nonlinear characteristics. This is based on the following reasons.

For example, as shown in Figure 11, if you completely set all of the movements from a still image to fF+fr4 to zero, where it is impossible to distinguish between a still image and a complete moving image, the amount of movement in the image is small/l: sharp and large. The difference is getting bigger and 1r1ri 1t
This is because the influence on r cannot be ignored.

Note that the pattern classification area is (2X2) l', j,
However, it may be performed for a larger area such as (2×4).

Effects of the Invention As explained above, the present invention can suppress images and at the same time prevent deterioration of angular clarity. Also, fix the movement information! : Since the noise component is suppressed by the transformation, the motion 17'i? 14 Even if there are few months, 441 movement information can be reliably output. In addition, since the motion information has several nonlinear characteristics, it is necessary to separate the
1'l Into color sending signal C + 1 + 4 Te + output output Sp, io
If the entrance into the world becomes small and virtuous, it is possible to improve both qualities.

[・TIN 'l' theory on page 1, SuJ Figure 1 and; 4
'+; 2 Fl/l are respectively conventional) 13 circuits for y/C, block 1''/, 1, i', :+'
3 j゛? j is 'YlC part F; f': times 1'6 (7) - (rll not shown)o) riri according to this invention
Figure 5 is a block diagram showing an example of a 1rT parallel variable i negative circuit and an intersecting negative circuit. Iit + i i' = butter 7
)(!II Show 1:'a,'11”z 7 Figure PI
Non-p,! Blocked '1 showing an example of the shape Pus,
Bow' hair 8 figure - 11th is 'c' in! + Preservation and release 1.
111 K 1jl<Zuru/IV l'4: lu4,
``(!' 12 Figure C is a block diagram showing Ili of the AC inverse conversion circuit and the parallel-to-serial conversion circuit.

(10) is a Y/C separation circuit, 03) is a comb filter, (2 (31 is a motion detection circuit, (20A) is a motion information signal forming circuit, (20B) is a characteristic output circuit, and c3il+ is an orthogonal transformation circuit Ca1l are orthogonal transformation circuits, (32A) to (
3zl)) is a nonlinear circuit, I3:(l is an orthogonal inverse transform circuit, and (3(1) is a parallel column transform circuit F3.

Representative Italy H16 Times Hide Matsukuma Figure 1 3 Figure 2 6 J d Figure 4

Claims (1)

[Claims] A comb-shaped filter that forms a frame difference signal of an input color video signal, a motion detection circuit that detects the center gap between the frames, and a motion detection circuit that detects the Q difference detection output and the frame difference signal. The motion detecting circuit is composed of one circuit for the force to be fixed, and the motion detecting circuit detects the motion l' that calculates the motion information between the arenes and
A circuit for forming the iY information signal and a circuit for applying nonlinear characteristics only when there is little motion information among the motion information (r4"-I) constitute a 41゛・f circuit, and this /I
? By using the output output from the color detection circuit as the above-mentioned motion output output, it is possible to output the 4[theta] color sending signal from the 4T device. Y/C
Molecule 3 (T circuit.