CN103354073A - LCD color deviation correction method - Google Patents

Abstract

The invention provides an LCD color deviation correction method. According to the method, a BP neural network based on genetic algorithm is introduced, wherein the genetic algorithm is used for performing rapid global searching on an initial weight value and a threshold value of the neural network, and thus a defect that conventional BP algorithm is easy to be trapped into a local minimum is overcome, a rate of convergence is fastened, and the generalization ability of the network is improved. Better quantification classification accuracy is provided by adopting Lab color space values as input output values of the neural network than directly adopting RGB color space vectors, and thus the LCD color deviation can be automatically corrected, and the LCD can have better color reproduction ability.

Description

A kind of LCD colour cast correcting method

Technical field

The invention belongs to the display technique field, relate to a kind of LCD colour cast correcting method.

Background technology

Along with the development of display technique, LCDs is widely used in the various electronic equipments.People require more and more higher to the color of LCDs.Chromatic rendition is an important indicator of evaluating liquid crystal display screen, but LCD display will cause color and the object itself of the image that gathers to have certain deviation under different photoenvironments; Under the different hardware device environment, also can cause the input and output image color error ratio to occur.How color error ratio is corrected, become a focus of current research.

The LCD display color adopts the RGB color space definition usually.But RGB color space and equipment are closely related, and the color that different equipment shows has difference, directly with the RGB color vector as input, can't obtain desirable treatment effect.The Lab color space has device independence, and this color space has reflected the colour vision impression preferably, and the aberration that it calculates and human eye vision perception are basically identical.

The BP neural network is widely applied in the color correction field in recent years.The BP neural network is a kind of Multi-layered Feedforward Networks by the Back Propagation Algorithm training, its learning rules are to use gradient descent method, constantly adjust weights and the threshold value of network by backpropagation, make the error sum of squares of network minimum, thereby arrive the purpose of anticipation learning.But in the process of study, crenellated phenomena can occur, even be absorbed in local minimum.

Summary of the invention

The present invention is directed to above-mentioned technical matters, propose a kind of LCD colour cast correcting method.Described method adopts the GA-BP neural network based on genetic algorithm and BP algorithm, can automatically correct the LCD color error ratio, makes LCD have better chromatic rendition ability.

The present invention adopts following technical scheme for solving the problems of the technologies described above:

A kind of LCD colour cast correcting method comprises the steps:

Step 1 take red, green and blueness as establishment of coordinate system RGB three-dimensional color space, and is chosen training sample in this space;

Step 2 is chosen the output calibration value that a standard screen is used for obtaining the BP neural network, selects the input value that a screen to be calibrated is used for obtaining the BP neural network else;

Step 3 is exported training sample through screen to be calibrated, use digital camera to obtain its RGB three-dimensional color space color value V _RGB, and this value is transformed into Lab color space value V _Lab, as the input value of BP neural network;

Step 4 through the output of standard screen, uses digital camera to obtain its RGB three-dimensional color space RGB color space value V ' training sample _RGB, and this value is transformed into Lab color space value V ' _Lab, as the output calibration value of BP neural network;

Step 5 creates the BP neural network, adopts to have the BP neural network of three input neurons, three output neurons and five intrerneurons;

Step 6 is optimized with initial weight and the threshold value of genetic algorithm to the BP neural network; Acquisition has the BP neural network of optimum weights and threshold value;

Step 7, training BP neural network: with V _LabAs the input of BP neural network, V ' _LabAs the output of BP neural network, according to predetermined excitation function, training function, maximum training iterations and neural metwork training permissible error, training BP neural network;

Step 8 is cured to the BP neural network that trains in the hardware configuration of LCD display, realizes LCD display color correction function.

Beneficial effect of the present invention: the present invention proposes a kind of LCD colour cast correcting method.Described method is introduced the BP neural network based on genetic algorithm, genetic algorithm is carried out fast global search to initial weight and the threshold value of neural network, solve the shortcoming that the traditional BP algorithm easily is absorbed in local minimum, accelerated speed of convergence, improved the generalization ability of network; Adopt Lab color space value as the input and output value of neural network, have better quantitative classification accuracy than direct employing RGB color space vector, can automatically correct the LCD color error ratio, make LCD have better chromatic rendition ability.

Description of drawings

Fig. 1 is a kind of LCD colour cast of the present invention correcting method workflow diagram.

Fig. 2 is the GA-BP neural network structure figure that the present invention adopts.

Fig. 3 is genetic algorithm improved BP neural network workflow diagram.

Fig. 4 is aberration correlation curve before and after proofreading and correct.

Embodiment

Below in conjunction with the drawings and specific embodiments, further specify a kind of LCD colour cast correcting method that the present invention proposes.

As shown in Figure 1, a kind of LCD colour cast of the present invention correcting method workflow diagram, its detailed process is as follows:

Step 1: the present invention is with the RGB color space little cubic space of 10*10*10 that evenly distributes, and total obtains 1000 samples.And from 1000 samples, pick out representative 400 as train samples.

Step 2: select a screen to be calibrated and a standard screen, use the high-definition digital camera that LCD demonstration image is taken, obtain respectively its color value, be designated as RGB.Operation platform is WindowsXP, and development language is VC++6.0, take the OpenGL shape library as the basis.

Step 3: screen color character vector to be calibrated extracts and standardization, by the CIE1976LAB colour difference formula RGB color space value is converted into Lab color space value, is designated as Lab.

At first with RGB(0-255) the color space value, be converted into XYZ (CIE1931 standard colorimetric system) color space value:

$\left\{\begin{array}{c}X=0.412453*R+0.357580*G+0.180423*B\\ Y=0.212671*R+0.715160*G+0.072169*B\\ Z=0.019334*R+0.119193*G+0.950227*B\end{array}\right.---\left(1\right)$

Then carry out normalized:

$\left\{\begin{array}{c}x=X/(255*0.950456)\\ y=Y/255\\ z=Z/(255*1.088754)\end{array}\right.---\left(2\right)$

Secondly the xyz value that obtains is converted to Lab color space value:

$\left\{\begin{array}{c}L=115*y^{1/3},y>0.00856\\ L=903.3*y,y<=0.00856\end{array}\right.---\left(3\right)$

$\left\{\begin{array}{c}a=500*(f\left(x\right)-f\left(y\right))\\ b=200*(\left(f\left(y\right)\right)-f\left(z\right))\end{array}\right.---\left(4\right)$

F (t)=t wherein ^1/3, work as t〉and 0.008856; F (t)=7.787t+16/116 is when t＜=0.008856

Partial data is as shown in table 1, the RGB color space value of comparing, and Lab has obvious numerical values recited and positive and negative differentiation.

Table 1 Color Characteristic

Color	R	G	B	L	a	b
							Red	255	0	0	69	80	67
Yellow	255	255	0	112	-22	94
							Blue	0	0	255	48	79	-109
Green	0	255	0	103	-86	83
							In vain	245	245	245	113	-2	0
Ash	124	125	128	91	0.1	-1

Orange	255	128	64	96	18	35
							Purple	128	0	255	65	89	-79
Palm fibre	128	64	0	76	10	66
							Black	8	7	3	35	-2	12

Step 4: treat that standard screen color character vector extracts and standardization, method obtains the Lab value as described in the step 3, is designated as L'a'b'.

Step 5: create initial BP neural network: the BP neural network of the present invention's design adopts the 3-5-3 structure, and its topological structure as shown in Figure 2.This structure has 3 input neurons and three output neurons.Lab color space value is as 3 inputs of BP neural network, and L'a'b' color space value is as 3 outputs of neural network, and the hidden layer neuron number there is no theoretic guidance, and according to many experiments, the present invention chooses 5 nodes.Excitation function adopts tansigmoid; The training function adopts trainbpx, utilizes the rapid bp algorithm Training Multilayer Neural Network, has namely adopted momentum or adaptive learning, can reduce the training time, and maximum training iterations is in 300, and the network training permissible error is made as 0.03.

Step 6: initial weight and the threshold value of using genetic algorithm optimization BP neural network: the GA-BP workflow as shown in Figure 3.With genetic algorithm initial weight and threshold value are carried out fast global search, the chromosome of genetic algorithm is exactly one of the BP network and connects weights or threshold value, each chromosome is encoded with specific mode, the satisfactory solution of genetic algorithm is exactly the weights of BP neural network and the approximate value of threshold value, the optimum weight threshold that obtains is assigned to not yet begin the BP network of training.

(1) coding and initialization colony: coded system common in the genetic algorithm is 2 scale codings and real coding.2 scale codings of comparing, real coding has the precision height, the advantage that speed of searching optimization is fast, the present invention adopts the real coding mode.To input the Lab value and be converted into the real number string.

(2) fitness function is definite, and this function is the unique information of genetic algorithm guidance search, and choosing of it determines the algorithm quality.This function should be easy to restrain and calculate and can not cause locally optimal solution.

(3) selection, intersection, mutation genetic operation

1, select: select the strong individuality of vitality to produce the process of new population from old population, the present invention adopts the algorithm of tournament selection method namely to select at random some individualities from population at every turn, and then best winning is that the father is individual, constantly repeats until produce N individuality.

2, intersect: two chromosomes that match are mutually pressed the mutual switching part gene of certain mode, thereby form new individuality.The introducing of crossover operator has determined the ability of searching optimum of genetic algorithm.Restructuring in the middle of the present invention adopts:

T ₂=S ₂+λ ₂(S ₁-S ₂) (5)

T ₁=S ₁+λ ₁(S ₂-S ₁) (6)

T in the formula ₁Be son individuality 1, T ₂Be son individuality 2, S ₁Be father's individuality 1, S ₂Be father's individuality 2, λ ₁, λ ₂Be scale factor.

3, variation: in the biological heredity process, can cause gene mutation in the accidentalia situation, thereby new species occur.Introduce mutation operator and produce new individuality, determined the local search ability of genetic algorithm.The present invention adopts and chooses at random several body with the definitive variation probability from population, and for the individuality of choosing, the number that produces between [0-9] at random on the whole position of need variation replaces legacy data.

Step 7: training GA-BP neural network: with training sample (training sample that the step 1 mode obtains) training weights and the BP neural network of threshold value behind genetic algorithm optimization.After satisfying the network training error, training process is finished automatically.

Step 8: the GA-BP neural network that trains is passed through dsp chip, be cured in the hardware configuration of LCD display, realize LCD display color correction function.

Fig. 4 is the comparison diagram of value of chromatism before and after proofreading and correct, and horizontal ordinate is the part sample of choosing, and ordinate is value of chromatism.By experimental results show that algorithm of the present invention has good correcting feature to the LCD image color.

Simulation result shows that this algorithm ability of searching optimum is strong, and fast convergence rate has overcome the BP neural network and easily has been absorbed in local minimum defective, has shown feasibility and the validity of the method.

Claims (1)

1. a LCD colour cast correcting method is characterized in that, comprises the steps:

Step 1 take red, green and blueness as establishment of coordinate system RGB three-dimensional color space, and is chosen training sample in this space;

Step 2 is chosen the output calibration value that a standard screen is used for obtaining the BP neural network, selects the input value that a screen to be calibrated is used for obtaining the BP neural network else;

Step 3 is exported training sample through screen to be calibrated, use digital camera to obtain its RGB three-dimensional color space color value V _RGB, and this value is transformed into Lab color space value V _Lab, as the input value of BP neural network;

Step 4 through the output of standard screen, uses digital camera to obtain its RGB three-dimensional color space RGB color space value V ' training sample _RGB, and this value is transformed into Lab color space value V ' _Lab, as the output calibration value of BP neural network;

Step 5 creates the BP neural network, adopts to have the BP neural network of three input neurons, three output neurons and five intrerneurons;

Step 6 is optimized with initial weight and the threshold value of genetic algorithm to the BP neural network; Acquisition has the BP neural network of optimum weights and threshold value;

Step 7, training BP neural network: with V _LabAs the input of BP neural network, V ' _LabAs the output of BP neural network, according to predetermined excitation function, training function, maximum training iterations and neural metwork training permissible error, training BP neural network;

Step 8 is cured to the BP neural network that trains in the hardware configuration of LCD display, realizes LCD display color correction function.

Images