CN101330564B - Method for correcting signal and circuit thereof as well as method for correcting image signal and circuit thereof - Google Patents

Abstract

The invention provides a signal correction method and a circuit thereof. A plurality of adjustable default values are used for clamping a plurality of input signal values in the signal correction method, thereby producing a plurality of corresponding clamping signals. Then, the matrix multiplication of the generated clamping signals is carried out according to a matrix coefficient which is obtained by multiplying at least two or more preset correction matrix coefficients, thereby outputting a calculation result. Subsequently, the signals after the correction are generated by clamping the calculation result according to a plurality of corresponding preset critical values.

Description

Signal correction method and circuit thereof and picture signal modification method and circuit thereof

Technical field

The present invention relates to a kind of multiplier operation method and circuit thereof in order to corrected signal, and particularly relate to a kind of merging multiplier operation method and circuit thereof, can correctly export revised signal, and the picture signal modification method and the circuit thereof that use this signal correction method.

Background technology

On digital image signal process aiming, according to the real image processing integrated circuit of making of algorithm (Integrated Circuits, IC) in, regular meeting utilizes multiplier (Multiplier) to carry out the multiplying of signal.For instance, generally in the image processing integrated circuit in digital camera (Digital Camera), data image signal all can be through the handling process of several picture signals, as shown in Figure 1.

Please refer to Fig. 1, mainly is the handling process schematic diagram of key diagram as processing integrated circuit inside.In digital camera, after the sensing signal that the analog/digital converter that belongs to Sensor Analog Relay System front end (Sensor Analog Front End) is exported imageing sensor transfers red green and blue (R, G, B) tricolor signal of numeral to by simulation, then just carry out treatment of picture by the image processing integrated circuit.White balance gains for example shown in Figure 1 (White Balance Gain) revision program 102, sensor photoelectric characteristic correction (Inverse Optical Electronic Conversion Function) program 104, color filtering array interpolation (Color Filter Array Interpolation) 106, colour correction (ColorCorrection) 108, color gamut conversion (Color Space Transform) 110 and gamma correction (GammaCorrection) 112 obtain the picture signal I that needs at last _m, wherein, color gamut conversion 110 is converted to a brightness signal Y and two chroma signals Cb, Cr with R, G, B tricolor signal exactly.

Generally speaking, according to different algorithms, the front and back ordinal relation of above-mentioned handling process can be different, and project also may increase and decrease to some extent.Yet, in above-mentioned handling procedure, for example white balance gains correction 102, colour correction 108 or color gamut conversion 110 or the like, these handling procedures all need multiplier to carry out multiplying.But because each tame manufacturer is in order to save the cost of image processing integrated circuit, the multiplier of front and back level can be done merging, for example the multiplier that white balance gains correction 102 and colour correction 108 handling procedures are used merges, as shown in Figure 2, or the multiplier that colour correction 108 and color gamut conversion 110 handling procedures are used merges, as shown in Figure 3, or even the multiplier that white balance gains correction 102, colour correction 108 and color gamut conversion 110 threes are used all merges, as shown in Figure 4.The multiplier that Fig. 2, Fig. 3, Fig. 4 are all handling procedure merges schematic diagram.

Yet, merely the multiplier that the front and back level used is merged and do not have supporting way if having only, can cause many problems.Wherein topmost problem is that if there is the numerical value (irrational numerical value) above codomain to import in the whole expression formula, the result who obtains computing so at last produces the situation of overflowing, and wrong numerical value just occurs.On the phenomenon of reality, can be observed these wrong numerical value, can make the high luminance area of image abnormal color occur, cause the quality of entire image to descend, this situation will describe in detail down below.

Please refer to Fig. 5, is the general mlultiplying circuit of explanation.To be multiplier 502 carry out multiplying according to the two value of input signal IN and another signal A with general mlultiplying circuit, and by the output clamp circuit 504 restriction outputs codomain of OUT as a result.Usually the figure place of OUT has fixing maximum clamper value according to exporting as a result and this exports clamp circuit 504.For instance, if the figure place of output OUT is 8 (bits), so maximum clamper value just is 255; If the figure place of output OUT is 10, then maximum clamper value just is 1023.Yet after multiplier merges,, and do not do corresponding correction, just can produce unexpected numerical value if still take old mlultiplying circuit shown in Figure 5.

Below the two does real example explanation with 8 output figure place and white balance gains correction and colour correction.If consider to carry out the white balance gains correction under the situation that the multiplier at said two devices merges as yet, with regard to mathematic(al) representation:

${\left[\begin{array}{c}R\\ \mathrm{<figure-callout\; id="1"\; label="G\; B\; o"\; filenames="H2007101122129E00061.png,H2007101122129E00081.png"\; state="\{\{state\}\}">G</figure-callout>}& \\ B\end{array},\right]}_o=\left[\begin{array}{ccc}{R}_{\mathrm{gain}}& 0& 0\\ 0& G_{\mathrm{gain}}& 0\\ 0& 0& B_{\mathrm{gain}}\end{array}\right]\&times;{\left[\begin{array}{c}R\\ G\\ B\end{array}\right]}_i---\left(1\right)$

Suppose (R _i, G _i, B _i)=(230,240,190), and the R signal gain coefficients R in the white balance gains correction matrix _Gain, G signal gain coefficient G _GainAnd B signal gain coefficient B _GainBe respectively 1.5,1 and 1.25, pass through white balance gains correction output (R afterwards so _O1, G _O1, B _O1)=(345,240,237.5).And then pass through the output clamp circuit and carry out the restriction of codomain, obtain (R at last _O1, G _O1, B _Oi)=(255,240,237.5).Then, utilize resulting numerical value (255,240,237.5) to carry out colour correction again:

${\left[\begin{array}{c}R\\ G\\ B\end{array}\right]}_{o2}=\left[\begin{array}{ccc}\mathrm{rr}& \mathrm{rg}& \mathrm{rb}\\ \mathrm{gr}& \mathrm{gg}& \mathrm{gb}\\ \mathrm{br}& \mathrm{bg}& \mathrm{bb}\end{array}\right]{\left[\begin{array}{c}R\\ G\\ B\end{array}\right]}_{o1}---\left(2\right)$

Consider following color correction coefficient:

$\left[\begin{array}{ccc}\mathrm{rr}& \mathrm{rg}& \mathrm{rb}\\ \mathrm{gr}& \mathrm{gg}& \mathrm{gb}\\ \mathrm{br}& \mathrm{bg}& \mathrm{bb}\end{array}\right]=\left[\begin{array}{ccc}1.3& -0.2& -0.1\\ 0.1& 1.4& -0.5\\ -0.2& -0.3& 1.5\end{array}\right]---\left(3\right)$

So, can obtain (R _O2, G _O2, B _O2)=(259.75,242.75,233.75), and then pass through the output clamp circuit and carry out the restriction of codomain, normally exported numerical value (R at last _O2, G _O2, B _O2)=(255,242.75,233.75).

If but the two directly merges with white balance gains correction and colour correction, can obtain merging matrix:

${\left[\begin{array}{c}R\\ \mathrm{<figure-callout\; id="2"\; label="G\; B\; o"\; filenames="H2007101122129E00091.png"\; state="\{\{state\}\}">G</figure-callout>}& \\ B\end{array},\right]}_o=\left[\begin{array}{ccc}\mathrm{rr}& \mathrm{rg}& \mathrm{rb}\\ \mathrm{gr}& \mathrm{gg}& \mathrm{gb}\\ \mathrm{br}& \mathrm{bg}& \mathrm{bb}\end{array}\right]\&times;\left[\begin{array}{ccc}{R}_{\mathrm{gain}}& 0& 0\\ 0& G_{\mathrm{gain}}& 0\\ 0& 0& B_{\mathrm{gain}}\end{array}\right]\&times;{\left[\begin{array}{c}R\\ G\\ B\end{array}\right]}_i---\left(4\right)$

Wherein

$\left[\begin{array}{ccc}\mathrm{rr}& \mathrm{rg}& \mathrm{rb}\\ \mathrm{gr}& \mathrm{gg}& \mathrm{gb}\\ \mathrm{br}& \mathrm{bg}& \mathrm{bb}\end{array}\right]\left[\begin{array}{ccc}{R}_{\mathrm{gain}}& 0& 0\\ 0& G_{\mathrm{gain}}& 0\\ 0& 0& B_{\mathrm{gain}}\end{array}\right]=\left[\begin{array}{ccc}1.95& -0.2& -0.125\\ 0.15& 1.4& -0.625\\ -0.3& -0.3& 1.875\end{array}\right]---\left(5\right)$

And the hardware implementation mode of existing directly merging multiplier as shown in Figure 6, and employed image correction parameter is then shown in the matrix parameter of equation (5).Please refer to Fig. 6, be existing merging multiplier architecture, comprise the combination of multiplier, adder and the output clamp circuit of many groups.For example 602 among Fig. 6 is expressed as multiplier, and 604 are expressed as adder, and 606 are expressed as the output clamp circuit.Suppose input signal (R _i, G _i, B _i) be similarly (230,240,190), after the matrix operation through equation (5), obtain the result and be (376.75,251.75,215.25), and then carry out the restriction of codomain through the output clamp circuit, obtain (R at last _O2, G _O2, B _O2)=(255,251.75,215.25), so this is incorrect value, correct output valve should be (255,242.75,233.75).

If same situation is hypothesis (R _i, G _i, B _i) be (255,255,255) of high brightness, after the matrix operation through equation (5), the result who obtains is (414.375,235.875,325.125), and then carries out the restriction of codomain through the output clamp circuit, obtains (R at last _O2, G _O2, B _O2)=(255,235.875,255).Yet this numerical value can make actual image present inclined to one side purple, and therefore (255,235.875,255) that obtain at last are incorrect color numerical value, and correct output valve should be (255,255,255).

Therefore, merging under the preceding topic of multiplier with the reduction hardware cost, how exporting correct numerical value, and then guarantee correct picture color, is the target that the dealer made great efforts in this field.

Summary of the invention

The present invention proposes a kind of merging multiplier operation method and circuit thereof, can still can obtain correct operation result under the preceding topic that reduces hardware cost.

The present invention proposes a kind of merging multiplier operation method and circuit thereof, if be used in the computing of image processing, can correctly export revised picture signal.

The present invention proposes a kind of merging multiplier operation method and circuit thereof, and the high luminance area that can avoid image is through producing the phenomenon of abnormal color behind the union operation.

In the signal correction method of the present invention, be applicable to merge the multiplying operation that wherein, the multiplying of merging operation comprises a plurality of multiplying parameters at least, obtain one after multiplying each other and merge the multiplying parameter and above-mentioned multiplying parameter merged in advance.This signal correction method comprises carries out a clamper step in advance, in order to according to a plurality of values of adjusting a plurality of input signals of default value clamper, produces corresponding a plurality of clamper signals according to this.According to merging the multiplying parameter, respectively above-mentioned clamper signal is carried out one and merge the multiplying step, export a plurality of operation results according to this.According to corresponding a plurality of preset critical, above-mentioned operation result is carried out the clamper operation respectively, the result of output is revised input signal.Wherein, described preset critical is the maximum that can represent corresponding to described input signal.

The present invention also proposes a kind of picture signal modification method, be applicable to and merge the multiplying operation, wherein, the multiplying operation of this merging comprises a plurality of multiplying parameters at least, obtain one after multiplying each other and merge the multiplying parameter and described multiplying parameter merged in advance, this picture signal modification method comprises: carry out a clamper step in advance, in order to according to a plurality of values of adjusting a plurality of image input signals of default value clamper, produce corresponding a plurality of clamper signals according to this; Merge the multiplying parameter according to this, described clamper signal is carried out one merge the multiplying step, export a plurality of operation results according to this; And according to corresponding a plurality of preset critical described operation result is carried out clamper respectively and operate, the result of output is revised image input signal, wherein, described preset critical is the maximum that can represent corresponding to described image input signal.

The signal correcting circuit that the present invention proposes comprises the input clamp device, merges multiplier and output clamp device.This input clamp device is in order to receiving a plurality of input signals, and according to the default value adjusted of a plurality of correspondences, the value of the above-mentioned input signal of clamper is exported a plurality of clamper signals according to this respectively.In order to merge the multiplying parameter above-mentioned clamper signal is carried out a merging multiplying and merge multiplier according to one, export a plurality of operation results according to this, wherein, a plurality of multiplying parameters are used in described merging multiplying operation at least, obtain merging the multiplying parameter and the multiplying parameter merged in advance after multiplying each other.And the output clamp device is exported revised input signal according to this in order to the value according to the above-mentioned operation result of a plurality of preset critical difference clampers.Wherein, described preset critical is the maximum that can represent corresponding to described input signal.

The present invention also proposes a kind of image signal correcting circuit, comprising: an input clamp device, and in order to receiving a plurality of image input signals, and according to the default value adjusted of a plurality of correspondences, the value of the described image input signal of clamper is exported a plurality of clamper signals according to this respectively; One merges multiplier, in order to merge the multiplying parameter described clamper signal is carried out a merging multiplying according to one, export a plurality of operation results according to this, wherein, a plurality of multiplying parameters are used in this merging multiplying operation at least, obtain this merging multiplying parameter and described multiplying parameter merged in advance after multiplying each other; And an output clamp device, in order to according to a plurality of preset critical value of the described operation result of clampers respectively, export revised image input signal according to this, wherein, described preset critical is the maximum that can represent corresponding to described image input signal.

Above-mentioned signal correction method and circuit, if be used in picture signal modification method and the circuit, then described multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a colour correction matrix coefficient and a white balance gains matrix coefficient.This merges the multiplying parameter and gets for the color correction matrix coefficient is multiplied by the white balance gains matrix coefficient.

When described image input signal comprises trichromatic R, G, B signal, comprise a R signal gain coefficient, a G signal gain coefficient and a B signal gain coefficient in this white balance gains matrix coefficient.And the described default value of adjusting comprises the maximum that can represent corresponding to image input signal, can adjust default value for three that get divided by R signal gain coefficient, G signal gain coefficient and B signal gain coefficient respectively.

In above-mentioned the picture signal modification method and circuit, described a plurality of multiplying parameter comprises a colour correction matrix coefficient and a colour gamut transition matrix coefficient, wherein, described merging multiplying parameter gets for the color gamut conversion matrix coefficient is multiplied by the color correction matrix coefficient.

In above-mentioned the picture signal modification method and circuit, described a plurality of multiplying parameter comprises a white balance gains matrix coefficient, a colour correction matrix coefficient and a colour gamut transition matrix coefficient, wherein, described merging multiplying parameter is that the color gamut conversion matrix coefficient is multiplied by the color correction matrix coefficient in regular turn and the white balance gains matrix coefficient gets.

In above-mentioned the picture signal modification method and circuit, described preset critical is the maximum that can represent corresponding to image input signal.

For above and other objects of the present invention, feature and advantage can be become apparent, preferred embodiment cited below particularly, and conjunction with figs. are described in detail below.

Description of drawings

Fig. 1 is the handling process schematic diagram of image processing IC interior.

Fig. 2, Fig. 3, Fig. 4 are that the multiplier of handling procedure merges schematic diagram.

Fig. 5 is general mlultiplying circuit.

Fig. 6 is existing merging multiplier architecture.

Fig. 7 is the flow chart according to the picture signal modification method of one embodiment of the invention.

Fig. 8 is the circuit diagram according to the image signal correcting circuit of one embodiment of the invention.

Fig. 9 is the image signal correcting circuit according to another embodiment of the present invention.

The reference numeral explanation

102: the white balance gains correction

104: the sensor photoelectric characteristic correction

106: the color filtering array interpolation

108: colour correction

110: color gamut conversion

112: gamma correction

502,602,832: multiplier

504,606,842,844,846: the output clamp circuit

604,834: adder

702-706: step

810: the input clamp device

812,814,816: clamp circuit

818,820,822: register

830: merge the matrix multiplication device

840: the output clamp device

A, B, B _i, B _O1, B _{O1 '}, B _O2, Cb, Cr, G, Gi, G _O1, G _{O1 '}, G _O2, R, R _i, R _O1, R _{O1 '}, R _O2, I _m, IN, I _N1-I _Nn, O _Ut1-O _Utn: signal

OUT: output

910: the input clamp device

930: merge the matrix multiplication device

940: the output clamp device.

Embodiment

The present invention proposes a kind of merging multiplier operation method and circuit thereof, can still can obtain correct operation result under the preceding topic that reduces hardware cost.The multiplying operation that merges in the above-mentioned merging multiplying operation comprises a plurality of multiplying parameters at least, obtain merging the multiplying parameter and above-mentioned multiplying parameter merged in advance after multiplying each other, and merge multiplying parameters according to these and merge multiplying, so can reduce the cost of hardware, in addition, more can after merging multiplying, can possess correct numerical value.

If with merging multiplier operation method proposed by the invention and circuit application thereof in the picture signal modification method.And these multiplying parameters can be for example color correction matrix coefficient and white balance gains matrix.This merges multiplying parameter can be to merge color correction coefficient to be multiplied by the white balance gains coefficient and to get, or merging the color gamut conversion matrix coefficient is multiplied by the color correction matrix coefficient and get, or merging color gamut conversion coefficient is multiplied by color correction coefficient and white balance gains coefficient in regular turn and gets.

In picture signal modification method of the present invention and the image signal correcting circuit, merge the cost that multiplier can reduce hardware, in addition, more can after merging multiplying, can possess correct numerical value, and then guarantee the correct of picture color.

The merging multiplier computing circuit that the present invention proposes comprises the input clamp device, merges matrix multiplication device and output clamp device.This input clamp device carries out the operation of clamper respectively according to a plurality of default values of adjusting to a plurality of input signals.This can adjust default value promptly is to avoid the wrong operation of clamper in advance in order to merge multiplying.In order to merge the multiplying parameter above-mentioned clamper signal is carried out a merging multiplying and merge multiplier according to one, export a plurality of operation results according to this, wherein, a plurality of multiplying parameters are used in described merging multiplying operation at least, obtain merging the multiplying parameter and the multiplying parameter merged in advance after multiplying each other.And the output clamp device is exported revised input signal according to this in order to the value according to the above-mentioned operation result of a plurality of preset critical difference clampers.

For describing merging multiplier operation method and circuit thereof proposed by the invention in detail, under will with the picture signal modification method do an actual implementation execute the example introduction, so also, this embodiment is not in order to limit merging multiplier operation method proposed by the invention and the exercisable category of circuit.Such as any need be based on the consideration that reduces cost, multiplier operation method and circuit thereof must be merged, all method proposed by the invention and circuit can be used, except the cost that can reduce hardware, in addition, more can after merging multiplying, can possess correct numerical value.

Fig. 7 is merging multiplier operation method and the circuit thereof according to preferred embodiment of the present invention, is used in picture signal modification method schematic flow sheet.In this flow process, at first, can adjust default value according to N is individual, the value of clamper N image input signal produces N clamper signal according to this, and wherein, N is a positive integer (step 702).Then, the N * N matrix coefficient of at least two default correction matrix multiplication gained of foundation carries out the multiplication of matrices computing to above-mentioned N clamper signal, exports N operation result (step 704) according to this.Then, the value according to N this N of preset critical clamper operation result produces N output signal (step 706) according to this.

In order more to clearly demonstrate the present invention and to contrast with prior art, the following image input signal of hypothesis earlier is three primary signals (being N=3), be respectively red (R) signal, green (G) signal and blueness (B) signal, and with 8 output figure places is example, and the two does the real example explanation with white balance gains correction and colour correction.Under the preceding topic of these hypothesis,, can design an image signal correcting circuit accordingly, as shown in Figure 8 in order to reach the purpose of the described method of Fig. 7.

Fig. 8 is merging multiplier operation method and the circuit thereof according to a preferred embodiment of the present invention, is used in the circuit diagram of image signal correcting circuit.Please refer to Fig. 8, it comprises input clamp device 810, merges matrix multiplication device 830, reaches output clamp device 840.Input clamp device 810 is represented it with Ri, Gi and Bi respectively in order to receive above-mentioned R signal, G signal and B signal.This input clamp device 810 can three of foundations can be adjusted the default value value of above-mentioned three image input signals of clamper accordingly, exports three clamper signals according to this, is respectively R _O1, G _O1And B _O1Merge matrix multiplication device 830 and multiply by 3 * 3 matrix coefficients of white balance gains matrix coefficient gained, above-mentioned three clamper signals are carried out the multiplication of matrices computing, and export three operation results according to this, be respectively R in order to foundation color correction matrix coefficient _O1', G _O1' and B _O1'.Output clamp device 840 is exported three output signals according to this in order to the value according to three above-mentioned three operation results of preset critical clamper, is respectively R _O2, G _O2And B _O2

In input clamp device 810, include clamp circuit 812,814,816, and register 818,820,822.Clamp circuit 812,814,816 is received signal Ri, Gi and Bi respectively, and can adjust the value that default value comes the image input signal that clamper receives according to one separately, to export clamper signal R respectively _O1, G _O1And B _O1Register 818,820,822 is then in order to deposit the required default value adjusted of clamp circuit 812,814,816 respectively.In merging matrix multiplication device 830, comprise the many groups multiplier and the adder that design in the matrix array mode, for convenience of description, at this only with multiplier 832, adder 834 explanations.Then comprise three the output clamp circuits 842,844 and 846 that merge matrix multiplication device 830 output valves corresponding to three in the output clamp device 840.

The present invention proposes a kind of circuit that merges the multiplier computing, arrangement and design as Fig. 8, comprise input clamp device 810, merge matrix multiplication device 830 and output clamp device 840 or the like, assembly that it is formed and quantity, all can adjust, be not limited to this embodiment according to the image input signal of required calculating.

Suppose the R signal gain coefficients R in the white balance gains correction matrix _Gain, G signal gain coefficient G _GainAnd B signal gain coefficient B _GainBe respectively 1.5,1 and 1.25, so above-mentioned three can to adjust default value be exactly that the maximum that can represent of the figure place (8) according to output signal gets divided by R signal gain coefficient, G signal gain coefficient and B signal gain coefficient respectively, just respectively with 255/1.5,255/1 and 255/1.25, obtain at last 170,255 and 204 etc. three can adjust default value.Can be deposited with respectively in register 818,820 and 822 and can adjust default value 170,255 and 204, use to offer clamp circuit 812,814 and 816 respectively.

Certainly, R signal gain coefficients R _Gain, G signal gain coefficient G _GainAnd B signal gain coefficient B _GainCan be different along with the photograph mode of digital camera.The R signal gain coefficients R in the night photography pattern for example _Gain, G signal gain coefficient G _GainAnd B signal gain coefficient B _GainMay be respectively 1.5,1 and 1.25, but the R signal gain coefficients R in the close-range photography pattern _Gain, G signal gain coefficient G _GainAnd B signal gain coefficient B _GainMay be again other three coefficient values.Therefore, the numerical value of being deposited in the register 818,820 and 822 can be different along with the photograph mode of digital camera.

Consider (R now _i, G _i, B _i)=(230,240,190), after indivedual process clamp circuits 812,814 and 816 carry out the clamper of codomain, can obtain clamper signal (R _O1, G _O1, B _O1) be (170,240,190).Then consider following color correction coefficient again:

$\left[\begin{array}{ccc}\mathrm{rr}& \mathrm{rg}& \mathrm{rb}\\ \mathrm{gr}& \mathrm{gg}& \mathrm{gb}\\ \mathrm{br}& \mathrm{bg}& \mathrm{bb}\end{array}\right]=\left[\begin{array}{ccc}1.3& -0.2& -0.1\\ 0.1& 1.4& -0.5\\ -0.2& -0.3& 1.5\end{array}\right]---\left(6\right)$

Therefore the two merging matrix of white balance gains correction and colour correction can be expressed as:

$\left[\begin{array}{ccc}\mathrm{rr}& \mathrm{rg}& \mathrm{rb}\\ \mathrm{gr}& \mathrm{gg}& \mathrm{gb}\\ \mathrm{br}& \mathrm{bg}& \mathrm{bb}\end{array}\right]\left[\begin{array}{ccc}{R}_{\mathrm{gain}}& 0& 0\\ 0& G_{\mathrm{gain}}& 0\\ 0& 0& B_{\mathrm{gain}}\end{array}\right]=\left[\begin{array}{ccc}1.95& -0.2& -0.125\\ 0.15& 1.4& -0.625\\ -0.3& -0.3& 1.875\end{array}\right]---\left(7\right)$

Then merge matrix multiplication device 830 and clamper signal (170,240,190) is carried out the multiplication of matrices computing, to obtain operation result R according to the merging matrix of equation (7) gained _{O1 '}, G _{O1 '}And B _{O1 '}, be shown below:

${\left[\begin{array}{c}R\\ \mathrm{<figure-callout\; id="1"\; label="G\; B\; o"\; filenames="H2007101122129E00061.png,H2007101122129E00081.png"\; state="\{\{state\}\}">G</figure-callout>}& \\ B\end{array},\right]}_{o^{}}=\left[\begin{array}{ccc}1.95& -0.2& -0.125\\ 0.15& 1.4& -0.625\\ -0.3& -0.3& 1.875\end{array}\right]\left[\begin{array}{c}170\\ 240\\ 190\end{array}\right]---\left(8\right)$

And operation result (R _{O1 '}, G _{O1 '}, B _{O1 '})=(259.75,242.75,254.25).At last, three the output clamp circuits 842,844 and 846 of output in the clamp device 840 all according to preset critical respectively to operation result R _O1', G _O1' and B _O1' carry out the clamper of codomain, to obtain output signal (R _O2, G _O2, B _O2).And preset critical is exactly output signal R _O2,, G _O2,, B _O2The maximum that figure place separately can be represented (promptly 255) therefore can obtain (R _O2, G _O2, B _O2)=(255,242.75,254.25), this is correct value.

If consider image input signal (R now _i, G _i, B _i) be the situation of high brightness, (R at this moment _i, G _i, B _i) be (255,255,255), after indivedual process clamp circuits 812,814 and 816 carry out the clamper of codomain, can obtain clamper signal (R so _O1, G _O1, B _O1)=(170,255,204).Process merging matrix multiplication device 830 carries out the multiplication of matrices computing according to the merging matrix of equation (7) gained to clamper signal (170,255,204) again, can obtain operation result (R _{O1 '}, G _{O1 '}, B _{O1 '})=(255,255,255).At last, 840 couples of operation result (R of output clamp device _{O1 '}, G _{O1 '}, B _{O1 '}) carry out the clamper of codomain, to obtain output signal (R _O2, G _O2, B _O2).And output signal (R _O2, G _O2, B _O2)=(255,255,255), this is correct value.Therefore, by the example of Fig. 7 and Fig. 8 as can be known, the present invention can export correct numerical value really, can avoid the reduction of image quality.

Spirit based on Fig. 7 and the described embodiment of Fig. 8, merge 830 default correction matrix coefficients that can use of matrix multiplication device and can certainly comprise color correction matrix coefficient and color gamut conversion matrix coefficient, and the matrix coefficient of the two merging just gets for the color gamut conversion matrix coefficient is multiplied by the color correction matrix coefficient.In like manner, merge 830 default correction matrix coefficients that can use of matrix multiplication device and can also comprise white balance gains matrix coefficient, color correction matrix coefficient and color gamut conversion matrix coefficient, and the matrix coefficient that above-mentioned three merges is that the color gamut conversion matrix coefficient is multiplied by the color correction matrix coefficient in regular turn and the white balance gains matrix coefficient gets.Yet, must be noted that under the situation that the default correction matrix coefficient that merging matrix multiplication device 830 is used is had some change the described default value of adjusting that is applied in the input clamp device 810 also should be done change accordingly.

The described embodiment of Fig. 7 and Fig. 8 is not in order to restriction the present invention just in order to explain orally.Because image input signal may not be a tricolor signal also, but other image input signal therefore by the teaching of Fig. 7 and the described embodiment of Fig. 8, can be set at image input signal N, and image signal correcting circuit can be done correction slightly, as shown in Figure 9.

Fig. 9 is according to merging multiplier operation method proposed by the invention and circuit thereof, is used in the schematic diagram of another image signal correcting circuit.I among this figure _N1-I _NnBe expressed as image input signal, O _Ut1-O _UtnBe expressed as output signal.Please refer to Fig. 9, it comprises input clamp device 910 corresponding to image input signal, merges matrix multiplication device 920, and output clamp device 940.This input clamp device 910 is in order to receive above-mentioned I _N1-I _NnSignal, and have right clamp circuit of n and register.Merge matrix multiplication device 930 for having the union operation unit of n * n matrix coefficient, in order to the predefined matrix coefficient of foundation union operation, for example the color correction matrix coefficient multiply by the n * n matrix coefficient of white balance gains matrix coefficient gained, a said n clamper signal is carried out the multiplication of matrices computing, and export n operation result according to this.Output clamp device 940 has n output clamp circuit, in order to the value of a foundation n preset critical clamper said n operation result, exports n output signal according to this.

What deserves to be mentioned is, though image signal correcting circuit has been depicted a possible kenel in the above-described embodiments, but knowing this operator should know, each manufacturer is all different for the design of wherein input clamp device, and therefore application of the present invention is when being not restricted to this possible kenel.In other words, come the value of the above-mentioned N of a clamper image input signal accordingly so long as utilize the input clamp device can adjust default value according to N, thereby can avoid abnormal numerical value to run into merging matrix multiplication device and carry out computing, be to have met spiritual place of the present invention just.

The present invention is because of before merging N image input signal of matrix multiplication device reception, just utilize earlier the input clamp device can adjust default value and come the value of the above-mentioned N of a clamper image input signal accordingly, therefore can avoid abnormal numerical value to run into merging matrix multiplication device and carry out computing according to N.Thus, the present invention just can export correct numerical value, and then avoids the high luminance area of image abnormal color to occur and reduced the quality of image.

Though the present invention discloses as above with preferred embodiment; right its is not in order to limit the present invention; anyly have the knack of this skill person; without departing from the spirit and scope of the present invention; when can doing a little change and retouching, so protection scope of the present invention is as the criterion when looking the claim person of defining of the present invention.

Claims (15)

1. signal correction method, be applicable to merge the multiplying operation, wherein, the multiplying of this merging operation comprises a plurality of multiplying parameters at least, obtain one after multiplying each other and merge the multiplying parameter and described multiplying parameter merged in advance, this signal correction method comprises:

Carry out a clamper step in advance,, produce corresponding a plurality of clamper signals according to this in order to according to a plurality of values of adjusting a plurality of input signals of default value clamper;

Merge the multiplying parameter according to this, described clamper signal is carried out one merge the multiplying step, export a plurality of operation results according to this; And

According to corresponding a plurality of preset critical described operation result is carried out the clamper operation respectively, the result of output is revised input signal,

Wherein, described preset critical is the maximum that can represent corresponding to described input signal.

2. picture signal modification method, be applicable to merge the multiplying operation, wherein, the multiplying of this merging operation comprises a plurality of multiplying parameters at least, obtain one after multiplying each other and merge the multiplying parameter and described multiplying parameter merged in advance, this picture signal modification method comprises:

Carry out a clamper step in advance,, produce corresponding a plurality of clamper signals according to this in order to according to a plurality of values of adjusting a plurality of image input signals of default value clamper;

Merge the multiplying parameter according to this, described clamper signal is carried out one merge the multiplying step, export a plurality of operation results according to this; And

According to corresponding a plurality of preset critical described operation result is carried out the clamper operation respectively, the result of output is revised image input signal,

Wherein, described preset critical is the maximum that can represent corresponding to described image input signal.

3. signal correction method as claimed in claim 2, wherein, described image input signal comprises trichromatic R, G, B signal.

4. picture signal modification method as claimed in claim 3, wherein, each this multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a colour correction matrix coefficient and a white balance gains matrix coefficient, wherein, this merging multiplying parameter gets for this color correction matrix coefficient is multiplied by this white balance gains matrix coefficient.

5. picture signal modification method as claimed in claim 4 wherein, comprises a R signal gain coefficient, a G signal gain coefficient and a B signal gain coefficient in this white balance gains matrix coefficient.

6. picture signal modification method as claimed in claim 5, wherein, the described default value of adjusting comprises the maximum that can represent corresponding to described image input signal, can adjust default value for three that get divided by this R signal gain coefficient, this G signal gain coefficient and this B signal gain coefficient respectively.

7. picture signal modification method as claimed in claim 3, wherein, each this multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a colour correction matrix coefficient and a colour gamut transition matrix coefficient, wherein, this merging multiplying parameter gets for this color gamut conversion matrix coefficient is multiplied by this color correction matrix coefficient.

8. picture signal modification method as claimed in claim 3, wherein, each this multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a white balance gains matrix coefficient, a colour correction matrix coefficient and a colour gamut transition matrix coefficient, wherein, this merging multiplying parameter is multiplied by this color correction matrix coefficient in regular turn for this color gamut conversion matrix coefficient and this white balance gains matrix coefficient gets.

9. signal correcting circuit comprises:

One input clamp device, in order to receiving a plurality of input signals, and according to the default value adjusted of a plurality of correspondences, the value of the described input signal of clamper is exported a plurality of clamper signals according to this respectively;

One merges multiplier, in order to merge the multiplying parameter described clamper signal is carried out a merging multiplying according to one, export a plurality of operation results according to this, wherein, a plurality of multiplying parameters are used in this merging multiplying operation at least, obtain this merging multiplying parameter and described multiplying parameter merged in advance after multiplying each other; And

One output clamp device in order to the value according to the described operation result of a plurality of preset critical difference clampers, is exported revised input signal according to this,

Wherein, described preset critical is the maximum that can represent corresponding to described input signal.

10. image signal correcting circuit comprises:

One input clamp device, in order to receiving a plurality of image input signals, and according to the default value adjusted of a plurality of correspondences, the value of the described image input signal of clamper is exported a plurality of clamper signals according to this respectively;

One merges multiplier, in order to merge the multiplying parameter described clamper signal is carried out a merging multiplying according to one, export a plurality of operation results according to this, wherein, a plurality of multiplying parameters are used in this merging multiplying operation at least, obtain this merging multiplying parameter and described multiplying parameter merged in advance after multiplying each other; And

One output clamp device in order to the value according to the described operation result of a plurality of preset critical difference clampers, is exported revised image input signal according to this,

Wherein, described preset critical is the maximum that can represent corresponding to described image input signal.

11. image signal correcting circuit as claimed in claim 10, and each this multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a colour correction matrix coefficient and a white balance gains matrix coefficient, wherein, this merging multiplying parameter gets for this color correction matrix coefficient is multiplied by this white balance gains matrix coefficient.

12. image signal correcting circuit as claimed in claim 11, wherein, described image input signal comprises trichromatic R, G, B signal, and comprises a R signal gain coefficient, a G signal gain coefficient and a B signal gain coefficient in this white balance gains matrix coefficient.

13. image signal correcting circuit as claimed in claim 12, wherein, the described default value of adjusting comprises the maximum that can represent corresponding to described image input signal, can adjust default value for three that get divided by this R signal gain coefficient, this G signal gain coefficient and this B signal gain coefficient respectively.

14. image signal correcting circuit as claimed in claim 10, wherein, each this multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a colour correction matrix coefficient and a colour gamut transition matrix coefficient, wherein, this merging multiplying parameter gets for this color gamut conversion matrix coefficient is multiplied by this color correction matrix coefficient.

15. image signal correcting circuit as claimed in claim 10, wherein, each this multiplying parameter is a matrix coefficient, and these a plurality of multiplying parameters comprise a white balance gains matrix coefficient, a colour correction matrix coefficient and a colour gamut transition matrix coefficient, wherein, this merging multiplying parameter is multiplied by this color correction matrix coefficient in regular turn for this color gamut conversion matrix coefficient and this white balance gains matrix coefficient gets.

Images