JPS59140784A - Chroma-key device - Google Patents

Abstract

PURPOSE:To obtain a chroma-key device which prevents erroneous composition of an achromatic color area by detecting the distance of a foreground color video signal from the position of an achromatic color on color coordinates, and correcting the level of a key signal according to the distance. CONSTITUTION:Foreground color video data FG.VID and background color video data BG.VID applied to an interface 1 are adjusted to proper phases by a timing signal TRS and outputted. The data FG.VID is applied to a correcting circuit 3 to correct the level of color difference data on each sample according to the distance from the origin (achromatic color) of each sample of the data FG.VID on the color coordinates. The level-corrected signal is compared with a reference point corresponding to a back color, sample by sample, by a key signal generating circuit 4 to generate a key signal with a specific level.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] This invention relates to a chromakey device that uses differences in five hues to fit a subject in a foreground image into a foreground image.

"Background Art and its Problems" A foreground color video signal in which a subject is positioned in front of a screen of a predetermined color, for example, blue, and a background color video signal 0 of an image to be a background are supplied.

2. Description of the Related Art A four-key system is known that generates a key signal corresponding to the outline of a subject from a foreground color video signal, and uses this key signal to form a composite image in which the subject is positioned in front of a background.

Generally, in a digital chromakey device, when one image is given as component signals (luminance signal Y1 color difference signals U, V), reference points (Uo, V) corresponding to the back color are
O) is set in the (, ty-v) color coordinate system, and a key signal for switching is formed based on this reference point. The key signal is clipped at two levels (peak clip level PL and base clip level BL), shaped, and used as a switching signal.

If the difference between the peak clip level PL and the pace clip level EL is small, the quantization noise of the key signal becomes large. Therefore, in order to increase this level difference, if the peak clip level PL is set high, the key signal near the origin of the nine-color coordinate system (i.e., the achromatic color area) will also have an intermediate level,
Therefore, in one composite image, achromatic objects are visible through the image, resulting in an unnatural image.

Therefore, the color difference signal U belonging to the achromatic color area is used in advance.

It is conceivable to replace each of (2) with a larger predetermined value. However, when a key signal is formed from color difference signals processed in this way, there are cases where objects with extremely strong J bells exist adjacent to each other.1For example, when the subject is a semi-transparent glass tip, .

White dots flicker in this area of the glass, resulting in an unsightly image.

[Object of the Invention] An object of the present invention is to provide a chromakey device that prevents erroneous compositing of achromatic color areas.

Further, the present invention aims to provide a chroma key device that prevents the adverse effect on a composite image caused by stepwise changes in key signals between an achromatic color area and other color areas. Summary of the Invention The present invention detects the distance on a color s standard between a foreground color video signal and a signal corresponding to a predetermined color to generate a key signal corresponding to a subject. What is the distance from the achromatic color position on the color coordinates of ? This device includes a detection circuit and a correction circuit that substantially performs level correction on the key signal according to the detected distance.

Embodiment Hereinafter, an embodiment will be described in which the present invention is applied to a digital comb keying device that targets digital video signals of Y, U, and V signal systems.

In FIG. 1 showing the overall configuration of a digital chromakey device, reference numeral 1 denotes an interface to which foreground color video data FG, VID and background color video data, # data IBG, VID are supplied together with respective timing reference signals TR8. This color video data is obtained by sampling the luminance signal Y1 and the color difference signals U and V, which are formed by matrix calculation of the output (R, G, B) of a color television camera, at a sampling frequency of, for example, 14 (near). It consists of each component. The interface 1 looks at the timing signals (horizontal synchronization signal, vertical synchronization signal, etc.) decoded from each timing reference signal TBS, sets the phases of the two color video data FG, VID, and BG, VID appropriately, and outputs them to the subsequent stage. Output to.

2 is a back color data forming circuit. Back color data is formed from the foreground color video data FG and VID, and this back color data is supplied to the color canceller 6. Also, foreground color video data FG,
The VID is supplied to a key signal forming circuit 4 via a correction circuit 3.

As will be described later, the correction circuit 3 determines the origin (achromatic color) of each sample of the foreground color video data on the (U, V) color coordinates.
The level of color difference data of each sample is corrected according to the distance from the sample.

The key signal forming circuit 4 generates a reference point (U(1+Vo)) corresponding to the back color and the foreground color video data FG, V.
A comparison operation is performed for each sample corresponding to the ID, and a key signal of a predetermined level is generated. The key signal itself generated in this way contains many disturbances and cannot be used as is, so in Keep, Tanatsusa 5 performs waveform shaping processing such as clipping, adjusting the edge timing of this clip output, and adjusting the gain. is applied, and a key signal KE'Y'' is obtained from the key processor 5.

The color canceller 6 removes the packed color from the foreground color video data FG and VID based on this key signal KEY. For example, when the subject is transparent, the transparent pack color is removed. Specifically, the back color data is amplitude-modulated using the key signal KEY, and the modulated output is subtracted from the foreground color video data FG and VID. This backlash removal is done only for the U and V signals, and the WI degree signal Y is simply passed through T. (), VID is supplied.

The delay circuit 7 has a delay amount corresponding to the time T required for the above-mentioned waveform processing in the keep processor 5.

Then, the outputs CAN and VI of the color canceller 6 and the background color video data BG and VID are supplied to the mixer 8, and mixing of the two is performed based on the key signal KEY. This mixing method involves simply switching and outputting two color video data (CAN, MID and BC) and VI.

At the boundary between the two, a cross-fade method can be used in which the level of one is gradually decreased while the level of the other is gradually increased. The digital filter 9 is used to adjust the waveform of the output of the mixer 8.

The interface 10 is for outputting to the outside the achromatic CAM and VID from the color canceller 6, the composite color video data KYD and 'VID from the digital filter 9, the respective timing reference signals, and the key signal KEY. It is something.

Furthermore, a microprocessor 1i, a CRT monitor 12
1 and a console 13 are provided, and the console 13
It is possible to translate the user's key input from the computer and transmit it inside the system, and to perform the necessary arithmetic processing in each circuit block. What is the composition of the parts?

The configuration is as shown in FIG. - As an example, the third
1fflA, a foreground 16 in which a subject 15, for example, a person, is positioned in front of a back screen 14 painted with a back color, for example, blue, and a background 17, shown in FIG. 3B.
I think about that.

The pack color data forming circuit 2 outputs the pack color data DB corresponding to the pack screen 14 to be used.
is formed and fed to a multiplier indicated at 18 in FIG. In this multiplier 18.

A key signal KEY having a phase opposite to that of the key signal KEY, that is, a gain of O in the region of the subject 15, is supplied. therefore,
As shown in FIG. 3C, data in which the back color exists in the hatched area excluding the subject 15 is generated and is subtracted from the foreground color video data FG and VID by the subtracter 19. As a result, as shown in the third diagram, video data CAN and MID that correspond to the subject 15 and from which the back color has been removed are generated from the subtracter 19 and are output to the multiplier 20.
added to. In addition, background color video data BG, V
ID is applied to multiplier 21 and amplitude modulated by key signal KEY, thus as shown in FIG. 3E.

An output is produced from the multiplier 21 in which the data of the region of the subject 15 in the background 17 has been removed. On the other hand, since the key signal KEY is supplied to the first multiplier 20, the multipliers 20 and 2
By supplying the output data of 1 to the adder 22,
As shown in Figure 3F.

Output video data KYD and VID corresponding to an image in which the subject 15 is embedded in the background 17 are obtained.

Here, an example of a method for forming a key signal will be explained with reference to FIG. 4. First, (U.

■) Set the reference point corresponding to the back color on the color difference coordinates (U
o, vo), the axis formed by connecting the origin and this point is the X axis, and the axis perpendicular to it and passing through (Uo + Vo) is the Y axis. Reference point (Uo, vo) and arbitrary (U',
If the contributions to the X and Y axes are expressed as x and y in a vector connecting the V) value, then
) sinθy= (V-VO) casθ-(UUo
) sin θθ=tab-” (”7U0). Therefore, a key signal is defined by a function as shown by a linear equation.

K=alxl+blyl (a, b are any positive constants.) When K on the left side is constant (t & K o, the (U, V) values that satisfy it are lined up on the sides of the rhombus. The key signal generated in this way is processed by the key processor 5.
The waveform is shaped by a clip circuit. For example, a transparent tip is in front of the back screen 14 as subject 1.
If you take an image of the foreground 16 that is falling down as 5, the back color is transparent in the center of the tip, so the 5th
As shown in FIG. A, a key signal K is generated from the key signal forming circuit 4, which has a high level corresponding to the contour of the subject 15 and a slightly low level at the center thereof. The key signal is 2, for example 8
It is a digital signal with one bit as one sample, but the fifth
In the figure, a signal analog waveform is shown for convenience of explanation. Then, a clip operation is performed using the base clip level BL and the peak clip level PL as threshold levels. In other words, everything below EL is set to level 0.

All levels above PL are set to level 255.

As a result, the key signal KEY as shown in FIG. 5B is generated as I&. FIG. 5C shows the key signal KEY of opposite phase.

The above-mentioned base clip level BL and peak clip level PL are shown in FIG. 4, respectively.

In the same figure, everything contained within the small diamond marked 23 is 0, and everything outside the large diamond marked 24 is 255.
1. The part located between the two is.

This is an intermediate level area that determines the slope of the key signal. From the grasshopper screen 14 to the subject 15, at the boundary there is a change T from +(U6 g'/1)) to a position corresponding to the color of the subject fK15.

If the peak clip level PL is reduced by /11 so as not to fall into the achromatic region T centered on the origin 0, and placed within the second quadrant, erroneous compositing in the achromatic region will not occur. However, in that case, two clip levels BL and P
The level difference between L becomes small, the amplitude of the key signal becomes insufficient, and 8/N deteriorates. Therefore, as shown in FIG. 4, the peak clip level PL is increased.

The correction circuit 3 in this embodiment includes a distance detection circuit that calculates the distance from the origin 0 of the color difference data (U, v) of the foreground color video data FG and VID, and a color difference detection circuit that calculates the distance from the origin 0 of the foreground color video data FG and VID. It is composed of a level correction circuit that corrects the level of data.

Examples of the configuration of the distance detection circuit are shown in FIGS. 6, 7, and 8, respectively. The example shown in FIG. 6 is two color difference data U, V.
squared υ2. The value of v2 is calculated by the calculation circuits 41 and 42, and both are added by the addition circuit 43, and the square root of the addition output is calculated by the calculation circuit 44.

This square root (He 17), that is, the data D of the distance from the origin
is supplied to a comparison circuit 45 and compared with a reference value D0 that defines an achromatic color area. If the distance from the origin is smaller, it is determined that the distance is within the achromatic color area, and data D of this distance is taken into the latch 46, and if not, the latch 46 is cleared.

The distance detection circuit shown in FIG. 7 has a simplified configuration in accordance with the actual situation.

In FIG. 7, 47.48 is an RO that converts the color difference data U and V transmitted in a two's complement code into absolute values.
M is shown respectively. This ROM 47.

The absolute values IUI and IVI output from the comparator circuit 4
9.50 and compared with the reference value. Since this reference value Do defines the achromatic color area near the origin, it is expressed by at most 5 bits. Comparison circuit 49
．． When the absolute value of the color difference data from 50 to 9 is equal to or smaller than the reference value Do, an output of H (?il level) occurs. The outputs of these comparison circuits 49 and 50 are supplied to an AND gate 51.

Also, the absolute value of the color difference data is supplied to the ROM 52.
.

be done. This ROM52 is IUI +IVI2
That is, distance data D'4;6 from the origin is output, and this distance data D is supplied to the latch 53.

The output of the AND gate 51 is supplied to the clear terminal of the latch 53. The output of this AND gate 51 is + (
-1) o≦U≦Do r-Do≦v≦I) o) In the case of CD, it becomes dH, and in other cases it becomes L. Therefore, in the case of color difference data included in the achromatic color area defined by the above-mentioned inequality, the distance data D is taken into the latch 53.

The distance detection circuit shown in Fig. 8 is based on the absolute value IU of one color difference data.
1. ) Vl is determined in the ROM 47, 48, and both are supplied to the adder circuit 54, and the distance data D is generated at the output thereof. This distance data D (-1Ul+1Vl)
7chi 55 and comparison diagram 1i856. A reference value Do is supplied to the comparison circuit 56, and a comparison output is generated which becomes H when (D≦Do) and becomes L otherwise, and this comparison output is supplied to the clear terminal of the latch 55.

This distance data D draws a square locus centered on the origin in the I (U, V) coordinate thread.

Similarly, the price is near 1M. also draws a square locus centered at the origin. This standard value. The locus becomes an achromatic region.

If the distance from the origin is determined strictly, the trajectory will be a circle. However, as mentioned above, approximation using a square locus causes almost no problems.

The configuration of the distance detection circuit can be simplified.

Using the distance data D obtained in this way.

Correct the level of color difference data. FIG. 9 shows an example of the configuration of the correction circuit 4, which includes both a distance detection circuit and a level correction circuit.

In FIG. 9, a distance detection circuit similar to that in FIG. 8 is used. However, a ROM 57 is provided between the adder circuit 54 where the distance data D is generated and the latch 55. This R
The OM 57 generates correction data from the given starting distance data D and the maximum value LO. This correction data is supplied to one input terminal of ALUs 58 and 59 via latch 55.

Nine color difference data U j, , V are supplied to the other terminal of the ALU 58 * 59, respectively. This ALU 5 B
.

59 are supplied with control signals DM+*DM, respectively. These 0-R signals DMH and DM2 are.

This changes depending on in which quadrant of the (U*V) coordinate system the reference point (Uo + Vo) is set.

In other words, the corrected color difference data appearing in the output of the ALU58159 needs to be as far away from the diamond-shaped area as possible with the reference point (Uo + "6) as the center.

For this, depending on the quadrant in which the reference point is set,
It is necessary to switch the operation performed by ALU 58 #59 to addition or subtraction.

The color difference data appearing at the outputs of At and ty58.59 is supplied to the key signal generation circuit 4. Since the color difference data included in the achromatic color area is corrected to move away from the reference point (UosVo) by combining correction data in the ALU 58159, the level of the key signal is also increased. Therefore, erroneous composition in which an achromatic subject appears transparent does not occur.

Formation of correction data in the ROM 57 will be explained with reference to FIG. Cv, v) coordinate system, a large diamond 24 (
(see FIG. 4), the reference value Do defines a square achromatic region 25 centered on the origin O. This achromatic area 2
ROM 57 generates correction data for the color difference data contained within 5.

This correction data changes non-linearly as shown by 26 in the range (O≦D≦DO) with respect to the distance data D from the origin O, and (-D≦D≦0). 2 in the range of
7, it changes non-linearly. (D=
In the case of O), the correction data has the maximum value L0.

In addition, as in this embodiment, when the reference point is set in the quadrant ①, the control signals DJ +, DM
z, as shown in FIG.

ALU5 B performs subtraction operation (A-B), and ALU5
9 is controlled to perform an addition operation (A + B). Therefore, in the corrected color difference data generated in the output of I ALU 5 B a 59, the U component increases in the negative direction,
Since the (2) component increases in the positive direction, it is moved toward the (2) quadrant.

Unlike this embodiment, if the reference point is set in the first quadrant, as shown in FIG.

Both ALUs 5B and 59 perform a subtraction operation, and a correction is made to move toward the second quadrant. If the reference point is set in the 2nd quadrant, the corrected color difference data will move to the 2nd quadrant, and if this is set in the 2nd quadrant,
The corrected color difference data is moved to the second quadrant.

FIG. 12 shows another embodiment of the invention. Unlike the above embodiment, this embodiment directly corrects the key signal generated from the key signal forming circuit 4.

The correction data is generated in the same manner as in the previous embodiment. However, a delay circuit 60 is provided at the output of the adder circuit 54 to compensate for the delay occurring in the key signal forming circuit 4. The key signal from the key signal forming circuit 4 and the correction data from the latch 55 are supplied to an adder circuit 61, and a corrected key signal is generated at the output of the adder circuit 61.

Note that the gain characteristics shown in FIG. 10 regarding the correction data are examples, and various changes including linear characteristics are possible.

"Effects of the Invention" According to this invention, when setting the lip level to include an achromatic color area and improving the "/N" of the key signal,
Erroneous compositing in achromatic areas can be prevented. Furthermore, when correcting data included in the achromatic color area, the amount of correction is changed according to the distance from the origin in the present invention.

Continuity with surrounding data can be maintained. Therefore, in the case of a translucent glass object, it is possible to prevent the object from appearing flickering.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a color canceller and mixer in an embodiment of the invention, and FIG. 3 is an embodiment of the invention. Figures 4 and 5 are route maps and waveform diagrams used to explain the operation of key signal formation and waveform shaping, and Figures 6, 7, and 8 are color difference data diagrams. An example of a distance detection circuit that calculates the distance from the origin. FIG. 9 is a block diagram showing the structure of an example of a correction circuit in an embodiment of the present invention, and FIGS. 10 and 11 are block diagrams showing the structure of another example and still another example. FIG. 12, which is a route diagram used to explain the operation of the circuit, is a block diagram showing the configuration of another example of the correction circuit. 3... Correction circuit, 4... Key signal forming circuit. 5...Keep four processors, 8...Mixer. 14... Pack screen, 15...
subject. 16...Foreground, 17...Background, 47.
48...°...ROM for obtaining absolute value, 57.
...ROM that generates correction data, 58 e
59 ``”” ALU. Agent Masato Sugiura Figure 5 Figure 6 O Figure 8 n

Claims (1)

[Claims] A foreground color video signal in which a subject is positioned in front of a screen of a predetermined color and a background color video signal are provided;
The distance on color coordinates between the foreground color video signal and the signal corresponding to the predetermined color is detected to generate a key signal corresponding to the subject, and the key signal causes the subject to appear in front of the background. The chroma key device is configured to form a composite color video signal where the foreground color video signal is located, and includes a circuit for detecting the distance between the foreground color video signal and the achromatic color position on the color coordinates, and a correction circuit that substantially performs level correction on the key signal.