Research On Characterization Method Of LCD Display Based On Spectrum

With the continuous advancement of technology,the display devices that can be used become more and more complex and diverse,but viewing the same photo between different media often has different visual effects.How to ensure the consistency of the viewing effect is very important.One of the solutions is To characterize the device,the display characterization is divided into spectral characterization and chromaticity characterization.This paper studies both of the two characterization models.For the chromaticity characterization model,this paper proposes an improved model based on the GOG model.The improved model uses a polynomial form to express the relationship between the normalized brightness and the driving value,and analyzes the improved model using 2,3,4,5,6 The effect of the power on the accuracy of the color difference,the results show that the use of the third-order polynomial can be close to the optimal prediction effect,and the improved model based on the thirdorder polynomial is used on the four display devices(three professional and one ordinary commercial)owned by the laboratory.Compared with the four models of GOG,PLCC,SRPM,and SRPPM,the driving value of the test sample is formed by any combination of 4,14,52,90,128,166,204,242 values.The results show that the improved algorithm based on the GOG model is the same as the GOG model.Under the test sample,there will be better prediction results,and the improved model can improve the forward prediction accuracy of the GOG model.For spectral characterization,there are two main research contents in this article.The first is to propose a spectral characterization model based on cross-terms.This model considers the mutual influence of two channels and analyzes the influence of the model training interval on the color difference.,The training interval includes 1,2,3,4,5,6...100,etc.,comprehensively consider the selection of 5 as the best training interval,and use this interval to train the cross-term spectral characterization model to compare with the above four models,the result it shows that this model is more suitable for NEC and BENQ displays than the other models,and can achieve more accurate prediction accuracy,but it performs generally on the other two displays,with larger color difference and spectral root mean square errors;the second is the core content of this article,namely the spectral characterization model of liquid crystal display based on singular value decomposition.The forward model uses training data to establish a single-channel drive value and a look-up table corresponding to the coordinates under the orthogonal base.The spectral radiance of any drive value can pass it is obtained by linear interpolation;the inverse model is to first find the corresponding base coordinates,and then obtain the corresponding coordinates according to the first base vector through inverse interpolation.For this model,the influence of the training interval on the color difference is also analyzed.The training interval includes 1,2,3,4,5,10,15,25,35,finally choose 15 as the best training interval for the four monitors,and compare with the above four models,it is found that the spectral characterization model using singular value decomposition is compared with others The four models have high chromaticity and spectral accuracy.The three models proposed in this paper propose new solutions for the display characterization model,especially the liquid crystal display spectral characterization model based on singular value decomposition,which has high prediction accuracy and has certain practical significance.