CN1015859B - Colour name encoding method for colour signal - Google Patents

Abstract

The present invention belongs to the technical field of color signal digital processing and provides a color signal digital encoding method, particularly to a color name encoding method which is a new component encoding method characterized in that color picture signals are encoded in two paths, wherein the two paths are a luminance encoding path and a color name encoding path. The effect is equal to the method of using one color name encoding path for replacing two color difference encoding paths in the component encoding at present. The compressed encoding amount is about 1/4 to 1/3. Another advantage of the method is that chrominance replacement can be carried out to the image. The method is suitable for all digitization technical fields of color signals, such as a color TV broadcasting system, a satellite TV image transmission system, a high-definition TV system, an analog component encoding mac transmission system and a computer image processing system.

Description

Colour name encoding method for colour signal

The invention belongs to colour signal digital processing technology field.

In the prior art, the colour signal digital coding method can reduce two kinds: a kind of is the full signal coding method, and it is with the anolog TV signals Direct Digitalization; Another kind method is the component coding method, and this method is that color picture signal is decomposed into luminance signal and color difference signal, divides three road codings, one road luminance coding, two road aberration coding.

At present, the transmission of analog color tv signal has multiple different systems: " quadrature amplitude balance system " " " NTSC " system; " order transmits colored and the memory system " " " SECAN " system and " phase-aoternation line system " " " PAL " system etc., the difference of these standards is that the processing mode of its line frequency, field frequency, colour signal is different, mutual conversion between the standard is also cumbersome, therefore, the full signal coding method is as just a kind of interim form.

Transmission standard for unified Digital Television, Consultative Committee on International Radio (CCIR) has passed through 601 suggestions about TV broadcast center digital television signal coding standard in 82 years, this proposed recommendations replaces the full signal coding method with component coding, and determines that/50 standards of 625 row and/60 standards of 525 row can compatible digital coding method parameters; The advantage of this suggestion is: this regulation helps three kinds of conversions between standard, and weak point is that it divides three road codings, and it is bigger than full signal encoding amount, and this makes transmission, record that inconvenience all be arranged.

Utilize human vision property, suitably the encoding rate of compressed signal.Colour signal can utilize three fundamental physical quantities to represent, i.e. " brightness ", " tone ", " saturation " (" tone " claims colourity again with " saturation "), and people's vision is inequality to the resolving power of three physical quantitys, tone susceptibility to the difference color matching is also inequality, for example: yellowish green interval is the highest, and is blue purple interval minimum; And it is least responsive to " saturation "; For example: when the brightness of signal changed 1-2%, people's vision is distinguishable to be drawn, and during tone variations 2%, people's vision only can perceive, but saturation is when changing 10-20%, and people's vision does not but reflect; Saturation changes with brightness again and changes, as shown in Figure 1a: wherein x represents saturation, and y represents brightness, and its rule is: only corresponding particular luminance value when saturation is for a hundred per cent in same tone, when brightness was below or above this particular value, saturation all can reduce; Vision according to the experiment people is relevant with brightness to the resolving power of color, and is strong to the resolving power of color when brightness is high, when brightness is hanged down, a little less than the resolving power to color; In television system, when brightness is the highest, tone is a white, and when brightness was minimum, tone was a black, as shown in Figure 2, people's vision is also relevant with tone to the resolution capability of color, and is when being bordering on yellow (its wavelength is a 585nm rice) and dark green (its wavelength is a 485nm rice), the most responsive, in visible-range, the color that people's vision can be distinguished is about 165 kinds.

In colour TV image, the variation of brightness is continuous, and when digitized processing, the inter-stage difference of quantification should be less than the resolution threshold value of people's vision.The variation overwhelming majority of colourity disperses, and is discontinuous, and obvious boundary line is arranged.When digitized processing, can utilize this specific character, the compressed encoding amount; Aberration coding in 601 suggestions has used the quantification progression identical with luminance coding, it is to transmit luminance signal with one road 8bit coding, two road 8bit coding transmits color difference signal, luminance signal is divided equally 220 grades, saturation signal is divided into 110 grades, tone is divided into 690 grades, and compare the classification of color harmony saturation with luminance signal meticulous, can further compress.

For overcoming the deficiency of component (aberration) compiling method, colour signal digital coding method-color name compiling method that the present invention proposes, be that a kind of new component coding method is characterized in dividing two road codings with color picture signal, one road luminance coding and one road color name coding, two road aberration that are equivalent to utilize one road color name coding to replace in the present component coding are encoded, reduced by one road coding, compressed the about 1/4-1/3 of encoding amount.This method is compared with the aberration compiling method, and another advantage can be carried out " colourity replacing " to image, and this is that the aberration compiling method can't realize.

The color name compiling method replaces the aberration compiling method, is applicable to all colour signal digitizing technique fields, comprises the color TV presentation system, the satellite television graphic transmission system, high-definition television system, analogue component coding mac system transmission system, computer image processing system.

According to the principle of three primary colours, this method is to utilize one group of binary code N1 ... Nn represents the colourity of mixing by different proportion respectively, just to every kind of binary numeral of colourity name commonly used.In fact, this method is that the color region that picture tube in the chromatic diagram of international lighting association recommendation can show is divided into some zonules, and a kind of colourity is represented in each zonule, uses a binary numeral name, is decided to be the color name compiling method.The area dividing mode can be taked even division, also can take non-homogeneous dividing mode, for reaching the purpose of compressed encoding amount, adopts non-homogeneous mode to divide, according between human eye vision sensitivity characteristic dividing regions as table 1, shown in Figure 3, be 3bit color name coding schematic.

The color name compiling method is will pass through vision signal E ' R, E ' G, E ' B and the E ' Y of γ correction according to different proportion to the Signal Processing mode, with binary code difference called after luminance coding and color name coding.

As shown in Figure 4, after the encoded device BM of signal sending end vision signal handles, send luminance coding and color name coding, receiving terminal is after decoder yM handles, the luminance coding that receives be multiply by the relevant parameter of this color name coding representative, to restore corresponding video signals.For analogue component coding mac system,, the color name coding encodes as long as being converted into analog quantity.

For increasing special effects, can carry out the color name replacement artificially, for example: guaranteeing to change the colourity of clothes and environment under the constant situation of the colour of skin; Or refer in particular to colourity partly in the autotelic change picture, promptly " colourity replacing " can apply in a flexible way.

As the digital tv encoder method, the present invention meets 601 suggestions as much as possible, also adopts 8bit luminance coding Y, 8bit color name coding N, and the bit number can increase or reduce as required.Luminance coding Y also takes the uniform quantization mode, luminance coding and color name encoded sampled frequency meet 601 suggestions, color name coding N takes the non-uniform quantizing mode, utilize statistical method, to in the TV image common to the tone saturation carry out statistical classification, according to the probability of its appearance, select bigger 220 of probability, carry out the color name scale.

After tone is determined, because the correlation between saturation and the brightness has multiple different interval division method, but accompanying drawings, equal proportion zonal coding method shown in Fig. 1 a, 1b, 1c, equisaturation zonal coding method, rhombus zonal coding method, or the like.

Accompanying drawings is as follows:

Fig. 1: brightness and yellow saturation graph of a relation

Fig. 1 a: equal proportion zonal coding method

Fig. 1 b: equisaturation zonal coding method

Fig. 1 c: rhombus zonal coding method

Fig. 2: human eye spectrum identification threshold

Fig. 3: 3bit color name coding colourity diagram intention

Fig. 4: color name compiling method coding, decoding schematic diagram

Fig. 5: equal proportion zonal coding device circuit diagram

Fig. 6: equal proportion subregion decoder circuit figure

Fig. 7: equisaturation zonal coding device circuit diagram

Fig. 8: equisaturation subregion decoder circuit figure

Fig. 9: rhombus subregion decoder circuit figure

Below several coding methods are introduced respectively:

One. equal proportion zonal coding method:

The characteristics of this method are that the equal proportion with vision signal E ' R, E ' G, E ' B and E ' Y is divided into area principle, and ratio is identical in the same district, and as shown in Figure 1a, its color name coding N is expressed as follows parameter:

kR＝0.299E′R/E′Y±m

N kG＝0.587E′G/E′Y±m

kB＝0.114E′B/E′Y±m

M represents the subregion precision in the formula, and value is 0-10%.

Account form: as the N60% yellow:

E′Y＝0.299×100%+0.587×100%+0.114×40%＝0.9316

kR＝0.299E′R/E′Y＝0.299/0.9316＝0.32

N kG＝0.587E′G/E′Y＝0.587/0.9316＝0.63

kB＝0.114E′B/E′Y＝0.114×40%＝0.0456/0.9316＝0.05

Be KR=0.32

The yellow KG=0.63 of N60%

KB＝0.05

All the other tone saturations are changed with reference to the method, in the practicality decimal number is converted to binary code, weave into corresponding color name coding N at coding side according to known vision signal E ' Y, E ' R, E ' G, E ' B, at KR, KG, KB reduction become corresponding video signals E ' R, E ' G, the E ' B of decoding end according to luminance coding Y and color name coding N representative.

E′R＝kR·E′y/0.299

E′G＝kG·E′y/0.587

E′G＝kG·E′y/0.114

With 60% yellow is example, when E ' Y=0.5

kR＝0.32

N60% kG＝0.63

kB＝0.05

E′R＝0.32×0.5/0.299＝0.535

E′G＝0.63×0.5/0.587＝0.536

E′B＝0.05×0.5/0.114＝0.219

So just be converted to vision signal E ' R, E ' G, E ' B according to luminance signal E ' Y in the vision signal and color name coding N.

Equal proportion zonal coding device:

Circuit theory as shown in Figure 5, vision signal E ' Y through gray correction, E ' R, E ' G is adjusted at the 1VP-P(peak-to-peak value) send into input 3, E ' R and E ' G are through overdamping, under the control of isochronous controller 6, enter sampling retainer 4, make 0.299E ' R and 0.587E ' G enter 8 analog to digital converters 7,8, standard electric pressure side at transducer is sent into E ' Y, convert thereof into and be 0.299E ' R/E ' Y and two group of 8 bit sign indicating number of 0.587E ' G/E ' Y, just kR and kG weave into corresponding color name coding N through encoder 14, and every kind of coding is represented one group of kR, the kG parameter, because kR+kG+kB=1 can be according to kR, kG obtains kB.

In order to meet 601 suggestion requirements, reserve 00010000 and 11110000 as last following boundary belt.

Equal proportion subregion decoder:

Its circuit as shown in Figure 6,8bit luminance coding Y and 8bit color name coding N send into decoder data terminal 16, luminance coding Y is through digital to analog converter 21, convert luminance signal E ' Y to, send into the standard electric pressure side of digital to analog converter 22,23,24 again, through decoder 18 color name coding N is translated into kR, kG, kB three road 8bit codes, through digital to analog converter 22,23,24 operational amplifiers 11,12,13, be converted to E ' R, E ' G, E ' B, more than each step all under isochronous controller 6 control, keep synchronously with external circuit.

Two. equisaturation zonal coding method

The characteristics of this method are that saturation is identical in the same district by between the saturation dividing regions.With 60% saturation yellow is example, and a point is dark saturation point shown in Fig. 1 b, and saturation increases with the increase of brightness before a point, crosses after a point, and brightness increases, and saturation is constant, in this method, and the representative of color name coding following parameter: N:Ya, kRa, kGa, kBa

Wherein Ya represents the brightness that dark saturation point a is ordered, and corresponding signal is E ' Ya

kRa＝0.299E′Ra/E′Ya±m

kGa＝0.587E′Ga/E′Ya±m

kBa＝0.114E′Ba/E′Ya±m

To represent subregion precision value be 0-10% to m in the formula.

Below be that example is calculated each parameter value with 60% yellow, Ya is 100% yellow value * 60%.

E′Ya＝（0.299×100%+0.587×100%）60%＝0.5316

KRa＝0.299E′Ra/E′Ya＝0.337

KGa＝0.587E′Ga/E′Ya＝0.663

KBa＝0

That is: Ya=0.5316

The yellow KRa=0.337 of N60%

KGa＝0.663

KBa＝0

All the other tone saturations are with reference to this method conversion.

In encoder-side, weave into color name coding N according to known E ' Y, E ' R, E ' G, E ' B, its relation is as follows:

When E ' Y≤E ' Ya

KRa＝0.299E′R/E′Y±m

KGa＝0.587E′G/E′Y±m

KBa＝0.114E′B/E′Y±m

As E ' y＞E ' ya

KRa＝[0.299E′R-0.299（E′y-E′ya）]/E′ya±m

KGa＝[0.587E′G-0.587（E′y-E′ya）]/E′ya±m

KBa＝[0.114E′B-0.114（E′y-E′ya）]/E′ya±m

In decoder end, obtain vision signal E ' R, E ' G, E ' B according to the parameter of known luminance coding Y and color name coding N representative.

When E ' y≤E ' ya

E′R＝kRa·E′y/0.299

E′G＝kGa·E′y/0.587

E′B＝kBa·E′y/0.114

When E ' ya＜E ' y

E′R＝[kRa·E′ya+0.299（E′y-E′ya）]/0.299

E′G＝[kGa·E′ya+0.587（E′y-E′ya）]/0.587

E′B＝[kBa·E′ya+0.114（E′y-E′ya）]/0.114

Below be example with 60% yellow, calculate the numerical value of vision signal:

If E ' y=0.7316

Ya=yellow 100% * 60%=0.886 * 60%=0.5316

Ya＝0.5316

The yellow kRa=0.337 of N60%

kGa＝0.663

kBa＝0

When E ' y＞E ' ya

E′R＝[0.337×0.5316+（0.7316-0.5316）×0.299]/0.299

＝0.799

E′G＝0.800

E′B＝0.200

The encoder circuit of this partition method as shown in Figure 7; vision signal E ' R, E ' G, E ' B through the gray scale calibration send into encoder input 5; through sampling retainer 4; send into No. three analog to digital converters 7,8,9; be converted to three tunnel 8 binary codes; encoded device 15 is weaved into 220 kinds of color name coding N, reserves upper and lower boundary belt in accordance with regulations.

The decoder circuit of this partition method is sent brightness and color name coding into the data input pin 17 of decoder as shown in Figure 8, and through decoder 19, color name coding N resolves into Ya, kRa, kGa, kBa four road 8bit coding; Y, Ya is through digital to analog converter 21,22 enter comparator 10 after converting analog quantity to, compare, when E ' Y＜E ' Ya: K2 connects, E ' Y is sent into digital to analog converter 23,24,25 standard electric pressure side, as standard source, kRa, kGa, kBa sends into digital to analog converter 23,24,25, be converted into kRaE ' y, kGaE ' y, kBaE ' y, by operational amplifier 11,12, be output as E ' R after 13 computings, E ' G, E ' B, when: E ' y＞E ' ya: E ' y and E ' ya subtract each other formation (E ' y-E ' ya) enter operational amplifier 11 by resistance decrement by comparator 10,12,13 respectively with kRaE ' ya, kGaE ' y, kBaE ' y, addition, at E ' R, E ' G, E ' B output has the 1V voltage-stabiliser tube, the voltage stabilizing amplitude limit prevents to export too high.

Encoder is all synchronoused working with external circuit under the control of isochronous controller 6.

Three. rhombus zonal coding method

The characteristics of this method are that saturation increases (a point is the saturation breakover point) with brightness before a point, increase with brightness in a point back saturation to reduce, shown in Fig. 1 c; Determine that the method that a orders is as follows, true qualities are saturated breakover point A absolutely, and absolutely connects a straight line between saturation point B with its corresponding complementary colours, and the point that all the other percentage saturated lines are crossing with it is a point.

Rhombus zonal coding method, color name coding N represents following parameter:

Ya-represents brightness value E ' Ya of breakover point a

Ratio before the breakover point a:

kRa＝0.299E′R/E′y±m

N kGa＝0.587E′G/E′y±m

kBa＝0.114E′B/E′y±m

Ratio behind the breakover point:

kRb＝（0.299E′R-kRa·E′Ya）/（E′y-E′Ya）±m

N kGb＝（0.587E′a-kGa·E′Ya）/（E′y-E′Ya）±m

kBb＝（0.114E′B-kBa·E′Ya）/（E′y-E′Ya）±m

M represents subregion precision, value 0-10% in the following formula

Be example with 60% yellow below, calculate above various value.

The blue look 100% of the yellow 100%-of Ya=()/2 * 60%+0.5

＝（0.886-0.114）/2×60%+0.5＝0.7316

Obtain the numerical value of E ' R, E ' G, E ' B according to the relation between each amount of equisaturation zonal coding method.

Result: E ' R=0.8

E′G＝0.8

E′B＝0.2

kRa＝0.299E′R/E′Y＝0.299×0.8/0.7316＝0.327

kGa＝0.587×0.8/0.7316＝0.642

kBa＝0.114×0.2/0.7316＝0.031

kRb＝（0.299-0.327×0.7316）/1-0.7316＝0.223

kGb＝（0.587-0.642×0.7316）/1-0.7316＝0.437

kBb＝（0.114-0.031×0.7316）/1-0.7316＝0.34

All the other tone saturations all can be calculated with reference to the method.

At coding side, weave into color name coding N according to vision signal E ' y, E ' R, E ' G, E ' B, restore vision signal in decoding end according to luminance coding Y and color name coding N equally.

When E ' y≤E ' ya

E′R＝kRa·Ey/0.299

E′G＝kGa·Ey/0.587

E′B＝kBa·Ey/0.114

When E ' y＞E ' Ya

E′R＝[kRa·E′ya+kRb（E′y-E′ya）]/0.299

E′G＝[kGa·E′ya+kGb（E′y-E′ya）]/0.587

E′B＝[kBa·E′ya+kBb（E′y-E′ya）]/0.114

With 60% yellow is example, when establishing E ' y=0.9

E′R＝[0.327×0.7316+0.223（0.9-0.7316）]/0.299＝0.925

E′G＝[0.642×0.7316+0.437（0.9-0.7316）]/0.587＝0.925

E′B＝[0.031×0.7316+0.34（0.9-0.7316）]/0.114＝0.701

Coding circuit is similar to equisaturation zonal coding device, just in encoder, changes according to the coding formula of rhombus encoder.The decoder principle as shown in Figure 9, luminance coding Y and color name coding N send into its data input pin, translate corresponding Ya through decoder 20, kRa, kGa, kBa, kRb, kGb, 7 road 8bit numeric data codes such as kBb, send into digital to analog converter 21,22, Y, Ya enters comparator 10 through conversion and compares, when Y≤Ya, K1 connects, luminance signal E ' y is sent into transducer 23,25,27 standard electric pressure side, at this moment, because comparator 10 is output as negative, E ' y-E ' ya is zero, so digital to analog converter 24,26,28 no-outputs, digital to analog converter 23,25,27 output is through operational amplifier 11,12, after 13 computings, be output as E ' R, E ' G, E ' B; When Y＞Ya, K2 connects, E ' ya is sent into the standard electric pressure side of digital to analog converter 23,25,27, at this moment, comparator 10 is output as E ' y-E ' ya, and this voltage signal is sent into the standard electric pressure side of digital to analog converter 24,26,28, the addition in operational amplifier of each transducer, be output as E ' R, E ' G, E ' B output is established the voltage stabilizing didoe pressure limiting, prevents that output voltage is too high.29 is 1: 1 operational amplifier among the figure.

In circuit working, also need there be isochronous controller and register etc. to finish data transaction and computing.

In above method, for rate of compression coding, do not influencing under the view reception effect situation, the quantity of color name coding N reduces as far as possible, for guaranteeing conversion accuracy, parameters such as Ya, kR, kG, kB are when being converted to binary code, quantize progression and can try one's best greatly, as use 8bit, 10bit etc.

More than introduced several examples of name encoding method for colour, in fact used name encoding method for colour can design multiple coded system, can produce multiple integrated circuit, also can finish conversion with computer program.

Table 1:3bit color name coding schedule

Signal 0.299E ' R/E ' Y 0.587E ' G/E ' Y 0.114E ' B/E ' Y N

Color

White 0.299 0.587 0.114 1 111

Red 0.299/0.299 00 0.299 100

Yellow 0.299/0.886 0.587/0.866 0 0.866 110

Green 0 0.587/0.587 0 0.587 010

Blue or green 0 0.587/0.701 0.114/0.701 0.701 011

Blue 00 0.114/0.114 0.114 001

Purple 0.299/0.413 0 0.114/0.413 0.413 101

Black 0000 111

Claims (5)

1, a kind of colour name encoding method for colour signal that is used for television transmission system is a kind of component coding method, and one of component of colour signal is luminance signal y, and it makes up in following ratio:

E′y=0.299E′R+0.587E′G+0.114E′B

Wherein E ' Y, E ' R, E ' G, E ' B represent the vision signal through gray correction, luminance signal adopts the uniform quantization mode, utilize one tunnel binary coding as luminance coding, it is characterized in that: another component of colour signal is the color name signal, color region in the chromatic diagram is divided into several regions, a kind of colourity is represented in each zonule, utilizes one tunnel binary numeral N1 ... Nn represents respectively to distinguish colourity respectively; To signal processing method is to encode with one tunnel binary coding called after color name according to different ratios through the vision signal EY of gray correction ER EG EB, to take non-homogeneous mode to divide to the division of the color region in the chromatic diagram.

2, name encoding method for colour according to claim 1 is characterized in that: be that equal proportion with vision signal E ' Y, E ' R, E ' G, E ' B is divided into area principle, ratio is identical in the same district.Color name coding N is expressed as follows parameter:

KR＝0.299E’R/E’Y±m

N KG＝0.587E’G/E’Y±m

KB＝0.114E’B/E’Y±m

Wherein subregion precision m value is 0-10%.

3, name encoding method for colour according to claim 1 is characterized in that: take equisaturation zonal coding method, promptly saturation is identical in the same district; Saturation increases with the increase of brightness before dark saturation point a, crosses after the dark saturation point a, and brightness increases, and saturation is constant, the following parameter of color name coding representative:

N：Ya、KRa、KGa、KBa

Wherein Ya represents the brightness of dark saturation point, and corresponding signal is E ' Ya

KRa＝0.299E’Ra/E’Ya±m

KGa＝0.587E’Ga/E’Ya±m

KBa＝0.114E’Ba/E’Ya±m

4, name encoding method for colour according to claim 1, it is characterized in that: take rhombus zonal coding method, the characteristics of this method are, before saturation breakover point a point, saturation increases with the increase of brightness, behind a point, saturation reduces with the increase of brightness, and the color name coding N of this method represents following parameter:

Ya-represents the brightness value E ' Ya breakover point a numerical value in the past of breakover point a:

KRa＝0.299E’R/E’Y±m

N KGa＝0.587E’G/E’Y±m

KBa＝0.114E’B/E’Y±m

The numerical value that breakover point a is later:

KRb＝（0.299E’R-KRa ^＊E’Ya）/（E’Y-E’Ya）±m

N KGb＝（0.587E’G-KGa ^＊E’Ya）/（E’Y-E’Ya）±m

KBb＝（0.114E’B-KBa ^＊E’Ya）/（E’Y-E’Ya）±m

5, a kind of coding/decoding method that is suitable for the described coding method of claim 1: the same ratio when luminance coding is pressed with coding becomes corresponding video signals through digital-to-analogue conversion;

It is characterized in that: become relevant parameter that this parameter be multiply by luminance signal E ' Y the color name code conversion and obtain the corresponding video signal.

Images