CN113781395A

CN113781395A - Lighting color quality evaluation method suitable for Chinese traditional painting

Info

Publication number: CN113781395A
Application number: CN202110907912.7A
Authority: CN
Inventors: 党睿; 杨洋
Original assignee: Tianjin University
Current assignee: Tianjin University
Priority date: 2021-08-09
Filing date: 2021-08-09
Publication date: 2021-12-10
Anticipated expiration: 2041-08-09
Also published as: CN113781395B

Abstract

The invention belongs to the field of illumination of cultural relics, and provides a theoretical basis for improving illumination quality on the basis of cultural relic protection, providing a research basis for color preference of a special luminous environment of a museum, establishing a color perception model suitable for illumination of the museum, and establishing an illumination strategy of the museum. The technical scheme adopted by the invention is that the illumination color quality evaluation method is suitable for Chinese traditional painting, psychophysical experiments are adopted to obtain data of color preference of a subject for each illumination condition, a support vector machine is adopted to convert the data from a low-dimensional space to a high-dimensional space, a hyperplane is constructed for classification, and a chromaticity preference grade chart is established for guiding the illumination design of a museum. The invention is mainly applied to the illumination occasions of cultural relics.

Description

Lighting color quality evaluation method suitable for Chinese traditional painting

Technical Field

The invention belongs to the field of cultural relic illumination, and particularly relates to an evaluation model of chromaticity preference and color rendering quality of a light source to traditional Chinese painting colors in exhibition and display illumination of museums and art museums.

Background

By 2018, the total number of Chinese museums reaches 5136, wherein the Chinese traditional painting collection has more than 70 thousands of pieces and huge stock. And the Chinese painting has rich picture content, various colors and very high cultural and artistic values. However, as important constituent materials constituting the traditional Chinese painting, paper, silk and most of pigments are typical high light-sensitive objects, which are very vulnerable to illumination and thus cause irreversible damage. At present, a great deal of research and standards related to museum lighting damage mainly start with the lighting quantity (illumination intensity and exposure) of a light source, and aim to reduce the influence of lighting on cultural relics, reduce the lighting quality of the cultural relics to a certain extent and influence the exhibition effect of visitors. The museum cultural relic illumination environment further improves the illumination quality of the cultural relics and is widely accepted on the basis of strengthening the illumination protection of the cultural relics.

In order to improve the Color quality of a light source so as to meet the illumination requirement of appreciation in a museum, the invention researches from two aspects of the chromaticity preference and the Color development preference of the light source, firstly, the illumination and CCT (correlated Color temperature) are respectively used as the photometric index and the colourmetric index of the light source and are two important parameters influencing the Color preference of the museum, Dianzhe museum at Zhejiang university researches the comfort area of the illumination and the Color temperature in the LED illumination of the museum, but the recommendation range of the illumination and the CCT meeting the traditional Chinese painting Color preference is still unclear at present. Secondly, the color development quality of the light source directly influences the visual perception of the museum luminous environment, and is an important index for measuring the degree of the appearance expression of the illumination light source on the illuminated painting color compared with a reference light source (usually natural light). Historically, the Color Rendering Index R was specified by the International standards organization (CIE) before LED (light emitting diode) lighting became popular_a(CRI R_a) And metamerism indices to adjust the spectral quality of conventional light sources. In 2015, the american society of lighting engineering (IES) issued a dual index specification containing an average fidelity index, an average color gamut index and a color vector map. Color rendering Performance of light sources to the International Commission on illumination (CIE) proposed CIE R in 2017_fTo make the average CIE R of the current common lamps_fIn place of R_a. In 2018, IES published TM-30-18 as a substitute for IES TM-30-15. But to date, CIE R_fTM-30-18 is taken as two latest color development quality indexes, but the color developing method is only suitable for common lighting environments such as offices, classrooms and the like, and in addition, the colors of some special cultural relics are not in the range of 99 color samples of TM-30-18, so that the color developing method is not suitable for museumsA special light environment.

The research respectively researches the relation between the chromaticity index and the color development quality index in the museum luminous environment and the color preference of human eyes, aims to obtain the color preference characteristics of typical Chinese traditional painting under the working conditions of different illumination and correlated color temperature combinations, establishes a chromaticity preference evaluation model and draws a chromaticity preference grade chart, and provides a research basis for the color preference of the museum special luminous environment; secondly, the light source color rendering quality has an important influence on the visual perception of the light environment, the research changes three key color rendering indexes of spectral fidelity, color gamut and color gamut shape through psychophysical experiments, obtains the influence of the light source color rendering performance on color preference when the traditional drawing of the ornamental China is carried out, and establishes a color perception model suitable for the illumination of museums.

In conclusion, the patent finally provides an obtaining method suitable for a color evaluation model of the traditional Chinese painting by researching the chromaticity preference and the color development preference suitable for appreciating the traditional Chinese painting in museum illumination and applying a support vector machine and a stepwise linear regression method from the viewpoint of the chromaticity preference and the color development quality of a light source.

1. The current closest prior patents to this application are as follows:

(1) the invention discloses a method and a system for evaluating the quality of display illumination light based on illumination and correlated color temperature (patent number: CN 202010334650.5). The method and the system for evaluating the quality of display illumination light based on illumination and correlated color temperature comprise the measurement of the illumination of a light source to be evaluated; judging whether the illumination of the light source to be evaluated is in the illumination range applicable to the invention; collecting spectral power distribution of a light source to be evaluated; calculating the correlated color temperature of the light source to be evaluated in the uniform color space; judging whether the correlated color temperature of the light source to be evaluated is in the range of the correlated color temperature applicable to the invention; and for the light source to be evaluated, obtaining a corresponding estimated quantity value by combining a light quality estimation model according to the illumination and the correlated color temperature of the light source, and realizing the representation of the light quality of the display light source from the angles of light source whiteness perception and color preference.

The invention aims to realize the representation of the quality of an exhibition light source from the whiteness and color preference of the light source, mainly aims to judge whether the illumination and the color temperature of the light source to be measured are in the set illumination and color temperature range, the illumination and color temperature range is not suitable for the illumination requirements of traditional Chinese painting in a museum, and the patent does not quantitatively evaluate the color rendering performance of the light source.

(2) A full-spectrum LED illuminating lamp (patent number: CN201710098064.3) suitable for a museum discloses a full-spectrum LED illuminating lamp suitable for the museum, and belongs to the technical field of LED illumination. The control system consists of RGBW # LED four-color mixed light, a V-shaped groove paraboloid condenser and a light guide optical fiber. The device adopts a V-shaped groove paraboloid condenser and a light guide optical fiber to mix light of the RGBW four-color LED chip, and obtains a full-spectrum illuminating lamp with uniform light and color suitable for illumination of museums. The invention aims to solve the problem of white points of non-uniform mixed distribution of RGB # LED three-color mixed light, effectively improve the light mixing efficiency and obtain different requirements of different exhibits in a museum on illumination uniformity, glare, color temperature, color rendering, illumination and the like of a light source. The invention aims to provide a novel full-spectrum LED light source suitable for illumination of a museum, and different parameters can be adjusted according to different exhibits, so that the illumination requirement in the museum is met.

The invention aims to solve the white point problem existing in RGBLED three-color mixed light arrangement, the existing light source is reconstructed by reconstructing the internal structure of the light source, and a novel LED illuminating lamp is mainly provided.

(3) The invention discloses a method and a system for predicting multi-level illumination color temperature preference degree for traditional Chinese painting exhibition and display illumination (patent number: CN 202011430644.6). the invention discloses a method and a system for predicting multi-level color temperature illumination preference degree for traditional Chinese painting exhibition and display illumination, which comprises the steps of collecting the surface illumination of a traditional Chinese painting to be exhibited; judging whether the surface illumination of the traditional Chinese painting to be displayed is in the illumination range applicable to the invention; collecting spectral power distribution of a light source to be evaluated; calculating the correlated color temperature of the light source to be evaluated in the uniform color space; judging whether the correlated color temperature of the light source to be evaluated is in the range of the correlated color temperature applicable to the invention; and obtaining a corresponding estimated quantity value by combining an illumination preference estimation model according to the surface illumination of the traditional Chinese painting to be displayed and the correlated color temperature information of the light source to be evaluated, thereby realizing the representation of the illumination quality of the traditional Chinese painting display light source.

The invention aims to provide a multi-level illumination color temperature preference prediction scheme aiming at the display illumination of Chinese paintings, which realizes the representation of the display illumination preference of the Chinese paintings by taking the surface illumination of the Chinese paintings to be detected as a basis and taking a preference estimation model as a basis, but the invention only represents the preference of an illumination light source through illumination and color temperature, and actually, the color rendering quality of the light source is also important to the illumination preference, and the invention does not research the influence of the color rendering performance of the light source on the color preference when the Chinese traditional paintings are viewed.

(4) The invention provides a method for evaluating the quality of exhibition illumination light for visual color preference (patent number: CN201910308645.4), which can effectively realize the quantitative representation of light source color rendering for human eye visual preference perception, and comprises determining exhibition illumination mode, and summarizing exhibition illumination condition into three types of modes of showcase observation, open type observation and general type observation; measuring the relative spectral power distribution information of the light source in the visible light range, and calculating the color rendering index of the corresponding light source according to the determined display illumination mode; and according to the determined exhibition lighting mode, combining the corresponding existing light source color quality indexes, calculating a light source color rendering quantitative comprehensive evaluation index under different exhibition conditions, and finishing the light quality evaluation facing to the visual color preference.

The patent aims to calculate the relative spectral distribution of the light source according to different display lighting modes, and quantitatively evaluate the color rendering of the light source according to the display lighting modes and the existing evaluation indexes from the preference of audiences. The invention is a selection of an exhibition mode under the condition of equal illumination, only the color rendering preference of a light source is researched, but different illumination and color temperature can have the same influence on exhibition and appreciation of exhibits, and further have the influence on the color preference of audiences.

2. The current closest research to the present application is as follows:

(1) the visual influence of calligraphy works under the irradiation of LED light sources with different illumination intensities and color temperatures [ D ] 27 observers of Zhang Yuanming university of Dalian industry are invited to carry out psychophysical experiments under the irradiation of the LED light sources with different illumination intensities and color temperatures, subjective evaluation data of the observers on the calligraphy works with 4 calligraphy fonts under different light environments are obtained, and the influences of the calligraphy fonts and the lighting light sources on the preference degrees, the attraction degrees and the comfort degrees of the observers in the exhibition of the calligraphy works are analyzed and discussed. The experimental data result shows that: in the exhibition of calligraphy works, calligraphy fonts have no influence on the attractiveness and comfort of an observer; the comfort of an observer is greatly influenced by the color temperature, and the comfort is relatively high when the color temperature of the light source is 3500K and 4500K; the attraction degree of the calligraphy works to the observer is increased along with the increase of the illumination; the light source conditions that the observer has the highest preference degree for seal script, clerical script, regular script and running script are respectively as follows: 2500K 100lx, 3500K 150lx, 4500K 200 lx. The research is not based on the color preference of an observer and does not research the color rendering performance of a light source, the lighting environments of calligraphy works and painting works are completely different, and the final color temperature and illumination indexes are not suitable for traditional Chinese painting.

(2) The color quality of the LED illumination of the museum has a visual effect on the painting (D). Dial at Zhejiang university Zhai doctor studies the oil painting in the museum, and the color painting is commonly used to study the color quality of the LED illumination of the museum. Experimental studies a series of related psychophysical experiments were designed and performed, LED lighting with different illumination, color temperature, color rendering (color fidelity), and color gamut size and shape was studied, and data results were compared with results associated with other international research teams. The optimum color temperature for museum lighting may be summarized between 3000K and 4000K. Both the color fidelity and the color gamut of the illumination have a significant impact on the visual impact of the painting. Spectrally tunable illumination is a trend. The research takes gouache and oil painting works in modern painting as samples, and the pigment composition and the light sensitivity level of the research are greatly different from those of the traditional Chinese painting. The research is mainly from fidelity to the color rendering quality of the light source, the research on parameters such as color gamut size and color gamut shape is not deep enough, and an evaluation model of chromaticity preference and color perception is not established finally.

(3) Influence of illumination light source and paper color on exhibition and display preference of traditional calligraphy works [ D ] the Liu Qiang professor of Wuhan university studies on the illumination light source and exhibition and display preference of traditional calligraphy works in China, and the exhibition and display preference of calligraphy works is essentially the color and graphic preference problem in color combination preference. Aiming at the problem that the display and display mode of calligraphy works at the present stage lacks reasonable basis in the aspect of color science, the research analyzes and discusses the influence of an illumination light source and paper colors on the display and display preference of the traditional calligraphy based on subjective and objective experiments. The research firstly obtains the original information of the light source and the paper color through the measurement of the relative spectral power distribution of five typical LED lighting sources of correlated

color temperatures

2500,3500,4500,5500 and 6500K and the spectral reflectivity information of five typical calligraphy paper of white, yellow white, light white, red and orange, and respectively converts the original information into CIEXYZ and CIECAM02 color spaces. Then, the relevance analysis is carried out on the light source and the paper color attributes and 1000 groups of calligraphy exhibition preference data based on 40 observer psychophysics experiments, and influence factors of traditional calligraphy work exhibition preference are discussed from multiple statistical aspects such as the relevance analysis, the multiple regression analysis and the normal distribution analysis. Experimental results show that the exhibition lighting preference of the traditional calligraphy works has obvious particularity compared with the color preference of a common scene, the preference degree of the traditional calligraphy works is low in color attribute correlation such as lightness contrast and hue contrast, and the traditional calligraphy works are mainly influenced by light source factors. The influence of the illumination light source and the paper color on the exhibition preference of the traditional calligraphy works.

The calligraphy works used in the research are greatly different from the traditional Chinese painting, and particularly in the aspect of color composition, the traditional Chinese painting is more complex and has higher requirements on the color rendering performance of a light source. In addition, the research is mainly carried out on the exhibition and display preference of calligraphy works, the research under the combined working condition of different color temperatures and illumination is not involved, and an evaluation model of chromaticity preference and color rendering preference is not established.

4. The current national standard related to the present application is "museum lighting design specifications" (GB/T23863-:

(1) in item 4.3.1, a light source with a color temperature less than 3300K should be selected as the illumination light source.

(2) Item 4.3.2 in the places such as display paintings, colored fabrics, multi-color exhibits, etc. where the requirement for color discrimination is high, a light source with a general color rendering index (Ra) of not less than 90 should be used as the illumination light source.

(3) Item 6.0.1 reduces the UV and IR radiation in the lamp light so that the relative UV content of the light source is less than 5 μ W/1 m.

However, the standard has two problems:

(1) the chromaticity preference and the color rendering preference of the light source directly influence the visual perception of an observer, but the museum lighting standard does not have indexes for evaluating the appreciation from the chromaticity and the color rendering of the light source, which indicates that the current lighting standard is not applicable to the lighting evaluation of traditional Chinese paintings.

(2) The current standard began to be implemented in 2009 and there is currently no new revision. LEDs have not been applied to museum lighting before 2009, so the current standard is a relevant index established for traditional light sources such as metal halide lamps, tungsten halide lamps, fluorescent lamps and the like. However, LEDs are rapidly developed as a new-generation light source in recent years, and are beginning to be used in museum lighting in large quantities. The LED can meet the requirements of three indexes of 'no ultraviolet and infrared, color temperature lower than 3300K and general color rendering index greater than 90' in the standard, and has a plurality of types. However, considering the flexible spectrum composition of the LEDs, the spectrum differences of LEDs of different manufacturing principles, different manufacturers and different models are great, the indexes of color temperature, Ra and the like which affect the lighting quality are different, and the influence on viewing is also different, so that the lighting quality of the traditional chinese painting cannot be ensured from the two aspects of chromaticity requirement and color rendering requirement.

In summary, no achievement in the aspect of 'being suitable for Chinese traditional painting color evaluation model' from the aspects of chromaticity preference and color development quality is found at present. On this background, the present application proposes an acquisition method that can be applied to a chinese traditional painting color evaluation model.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention aims to:

(1) a light source chromaticity preference and a perception model of the traditional Chinese painting are established, and the influence of correlated color temperature and illumination on the color preference of the traditional Chinese painting is researched. The traditional Chinese painting has extremely high cultural value and artistic value, but is also a high-light-sensitivity article, and when the painting is displayed in a museum, the painting does not have corresponding standards to limit the range of illumination and correlated color temperature so as to meet the color preference of human eyes. The objective is to obtain the color preference characteristics of typical Chinese traditional painting under the working condition of different illumination and correlated color temperature combination through psychophysical experiments and a support vector machine model algorithm, establish a chromaticity preference evaluation model and draw a chromaticity preference grade chart, and provide a research basis for the color preference of a special light environment of a museum.

(2) A light source color development preference and a perception model of the Chinese traditional painting are established, and the influence of relevant color development indexes on the color evaluation of the Chinese traditional painting is explored. Although several methods for quantifying the quality of color rendering are proposed, there is still the disadvantage that it is not suitable for the special light environment of museums. Therefore, the target changes three color development key indexes of the fidelity, the color gamut and the color gamut shape of the spectrum through a psychophysical experiment, obtains the influence of the light source color development performance on the color preference when the traditional drawing of China is viewed, and establishes a color perception model suitable for the illumination of the museum.

The lighting device provides a theoretical basis for balancing the lighting strategy of cultural relic protection and viewing, and in response to a call for living the cultural relic, the lighting quality needs to be improved on the basis of cultural relic protection. The research starts from the influence of different illumination and color temperature on color preference and the color rendering performance of the illumination light source, aims to find the balance between illumination protection and illumination quality for painting cultural relics, and further provides a theoretical basis for the establishment of an illumination strategy of a museum.

Therefore, the technical scheme adopted by the invention is that the illumination color quality evaluation method is suitable for Chinese traditional painting, psychophysical experiments are adopted to obtain data for a subject to grade the color preference of each illumination condition, a support vector machine is adopted to convert the data from a low-dimensional space to a high-dimensional space, a hyperplane is constructed for classification, and a chromaticity preference grade chart is established for guiding the illumination design of a museum.

And calculating an arithmetic mean value of color preference of two Chinese traditional paintings to be evaluated, wherein the calculated model comprises three preference levels which are respectively 'dislike', 'acceptable' and 'like', and a C-SVC (C-Support Vector Classification) model is built by using Matlab2017a software and an LIBSVM tool kit. And in the training mode, the illumination and the CCT are used as characteristic variables, the processed preference level is used as a target variable, and parameter optimization is carried out.

In the chromaticity preference model, two indexes of CCT and illumination are used as input quantities, and chromaticity preference level is used as an output quantity.

The detailed steps are as follows:

1. light source

An 11-channel spectrally tunable light box, thomslate led tube, with dimensions of 38cm x 50cm x 55cm, was used to generate 28 light sources. The gray light absorption cloth is pasted on the inner wall of the lamp box, and four illumination values of 50, 100, 150 and 200lx are selected as variables. Seven CCT variables of 2700, 3000, 3500, 4000, 4500, 5000 and 5500K are respectively set at each illumination value, and 28(4 x 7) working conditions | Duv | ≦ 0.0051, CRI Ra of 97 +/-1, Rf of 96 +/-2 and Rg of 100 +/-3:

photometric and colourmetric information for the conditions in tables 5-128

1. Drawing sample

Selecting two representative paintings as samples, wherein the first original work is loquat mountain and bird pictures, the printing samples are adopted, the picture size is 27 x 27cm, yellow and brown are mainly used as main colors, and the printing samples also comprise green, white and black; the second picture is a true drawing drawn by a folk painter, and is a freehand drawing and a picture color measurement site thereof. The picture subject is flower and bird, the picture size is 35 x 35cm, mainly regard yellow of the substrate as the dominant color, also include red, green, black, white;

2. is tested

5 subjects participated in the experiment with a male to female ratio of 18: 7. The test subjects are 19 to 31 years old, the average age is 23.8 years old, the standard deviation is 2.9 years old, the test subjects pass the Ishihara color blindness test, a chair used in the test subject in the experiment is placed in front of an observation lamp box, and the visual angle of the test subject is ensured to be fixed. The picture is horizontally placed in the lamp box, the distance between the picture and the tested eyes is about 59 cm, and the tested eyes cannot see the light source;

3. process for producing a metal oxide

The number of comparisons generated using the pairwise comparison is expressed as:

in the formula, J is the comparison times of the experiment, R is the repetition times of the experiment, and N is the number of samples to be evaluated;

four illumination groups and a repeated illumination group, wherein each illumination group has 7 CCTs, so that each picture is tested to be compared 105 times, firstly, each participant is given 1 minute of time to adapt to the experimental environment, and then after the whole experimental process is learned, three pairs of training illumination conditions are randomly selected to be judged;

in order to avoid sequence deviation, each pair of each group is played in a random sequence, the occurrence sequence of two illumination conditions in each pair is random, the duration time of each illumination condition is 4 seconds, an experimenter informs that the former working condition is X and the latter is Y, only letters of preferred working conditions need to be answered orally when the tested result is judged, the result is recorded by the experimenter, the working condition to be tested can be played for multiple times until the tested result can make positive judgment, two minutes of rest are carried out between each illumination group, in order to test the validity of tested data, one illumination group is randomly selected to carry out repeated experiments after four illumination groups are judged, the number of times of misjudgment of each tested result is not more than three, during data statistics, the assumption is that each working condition group has two working conditions of X and Y, if X is more than Y, the more preferred X is that A is 2, and B is 0; if X < Y, then Y is more preferred, then the opposite is true; if X is Y, namely the like degrees of X and Y are the same, respectively marking 1 point; after all comparisons were completed, the experimenter played randomly the lighting conditions for each group and asked the participants to choose what is considered neutral: illumination conditions of neither cold nor warm CCT;

in the subjective questionnaire, subjects scored a color preference for each lighting condition, "color preference" representing a subject's preference for a painterly-appearing color; after the pair comparison is finished, the tested person takes a rest for five minutes, then the tested person needs to sequentially score 28 randomly sequenced working conditions by seven-level evaluation, and the score is-3 to 3 and comprises 0; the 28 working conditions are divided into four groups in a random sequence, the groups are used as units for scoring to prevent fatigue of the tested object, the rest is carried out for two minutes between each group, and the whole group of working conditions are played twice in turn before scoring of each group to enable the tested object to be familiar;

and after the grading data are obtained, converting the grading data from a low-dimensional space to a high-dimensional space by using a support vector machine, constructing a hyperplane for classification, and establishing a chromaticity preference level diagram.

The method also comprises the following data reliability analysis steps:

before statistics of scoring data results, the data reliability of an evaluator is judged according to a Kendall consistency coefficient method, and a calculation formula is shown in 4-2.

Wherein W is a consistency coefficient, and the larger the W value is, the stronger the data consistency is; n is the number of the evaluated objects; k is the number of the scorers or the standard number according to which the scorers score; s is the sum R of the rating of each evaluated object_iAnd the average of all these sums

Is a sum of squared deviations, i.e.

Where m is_iThe number of the duplicate ranks, n, in the evaluation result of the i-th evaluator_ijThe same grade number of the jth repetition grade in the assessment result of the ith evaluator; the calculation results are shown in tables 4-2 and 4-3.

Since the number of the scorers K is 25, the number of the evaluated objects N is 7 and exceeds the range of the Kendal consistency coefficient significance critical value table, the W value is converted into chi²Value, then X²Checking and calculating formula X²K (N-1) W, the data freedom d is N-1 to 6, when the significance is 0.01 by checking a chi-square test critical value table, if the experimental data chi²All values are greater than 16.81, indicating that all W values reach a very significant level:

TABLE 4-2 values and values of data obtained by pairwise comparison

Tables 4-3 values and values of the subjectively evaluated data

The invention has the characteristics and beneficial effects that:

1. a chromaticity preference evaluation model established by a support vector machine is utilized to draw a chromaticity preference grade chart, so that the corresponding relation between the color preference and the illumination intensity and CCT when the traditional painting of China is viewed can be effectively described, and meanwhile, a research basis is provided for formulating the illumination standard of a museum.

2. The understanding of human visual perception is further deepened, the relation between the light source color rendering performance index and the color preference is obtained, a perception model is built by utilizing a stepwise regression method, subjective evaluation perception can be predicted and represented, and meanwhile, a basis is provided for the color rendering quality evaluation index suitable for the museum, so that a professional can be assisted to make a more intelligent design decision.

3. According to the proposed evaluation model, reference can be provided for the establishment of the current museum lighting standard.

4. The invention can provide basis for museum lighting designers to select light sources.

5. The invention can provide reference for quantitative evaluation of museum lighting sources by related departments.

In conclusion, the method can provide guidance for the lighting strategy of the fragile painting in the museum, greatly improve the exhibition and display level of the Chinese traditional painting, and contribute to the propagation of Chinese traditional culture.

Description of the drawings:

fig. 1 is a technical roadmap.

Fig. 2 is a pictorial sample. In the figure, (a) a strongbox, (b) a freehand drawing, (c) a flower-bird drawing, and (d) a red character drawing.

Fig. 3 is an experimental flow chart.

FIG. 4 is an experimental scenario of a subject evaluating a pictorial representation.

In fig. 5, (a) is a scatter diagram and (b) is an optimum parameter of the SVM.

Fig. 6 is a chromaticity preference level map drawn according to the preference evaluation model.

FIG. 7 is two classification curves obtained from a scatter-fit of model discrimination.

Fig. 8 is the overall environment and geometry of a museum laboratory.

Fig. 9 is a preference model.

Fig. 10 is a naturalness model.

FIG. 11 is a verification of a predictive model.

Fig. 12 is a graph of the relative spectral power distribution for 28 operating conditions.

Detailed Description

2. Experimental light source

The experiment was performed in the darkroom optical laboratory at Tianjin university, using an 11-channel spectrally tunable light box THOUSLITE LEDCube to generate 28 light sources used in the experiment. The size of the light box is 38cm by 50cm by 55 cm. The gray light-absorbing cloth is adhered to the inner wall of the lamp box, so that the influence of glare on the irradiation effect is effectively avoided. CIE 157: 2004, the illumination values for the museum environment are specified to be in the range of 50-200lx, and therefore four illumination values of 50, 100, 150 and 200lx are selected as experimental variables in equal ratios in the range. Also according to CIE 157: recommended ranges for CCT in 2004 and ANSI/IES RP-30-17[81], seven CCT variables of 2700, 3000, 3500, 4000, 4500, 5000, 5500K are set at each illumination value, respectively. The 28(4 x 7) working conditions | Duv | ≦ 0.0051, CRI Ra 97 + -1, Rf 96 + -2, and Rg 100 + -3.

Photometric and colourmetric information for the conditions in tables 5-128

4. Drawing sample

The traditional Chinese painting has extremely high collection value and artistic value. However, museums do not study the color preference of traditional Chinese paintings. The results of oil painting are not suitable for traditional Chinese painting which uses special methods to draw on fragile paper or silk. Traditional Chinese paintings can be classified into two types, i.e., calligraphy and painting, according to the expression technique and skill style. The main difference between the two drawing methods is that the calligraphy is fine and smooth, and the freehand drawing uses simple strokes to draw the landscape. Therefore, two representative paintings are selected as experimental samples, the first original work is loquat mountain bird picture, created by Nansong Lin Toona, and existing in Beijing Imperial palace Bozhou. The study used printed samples with a picture size of 27 × 27cm, mainly yellow and brown as the dominant colors, and also green, white and black. The second picture is a true drawing drawn by a folk painter, and is a freehand drawing and a picture color measurement site thereof. The picture subject is flower and bird, the picture size is 35 × 35cm, mainly takes the yellow of the base material as the main color, and also comprises red, green, black and white, and the concrete sample picture is shown in figure 2.

5. Is tested

5 subjects participated in the experiment with a male to female ratio of 18: 7. The subjects were from 19 to 31 years of age, with a mean age of 23.8 years and a standard deviation of 2.9 years. They all passed the Ishihara color blindness test and written informed consent. Fig. 4 shows an experimental scene of evaluating the pictorial work by a subject, and a chair used in an experiment is placed in front of an observation lamp box to ensure that the visual angle of the experiment is fixed. The picture is placed in the lamp box, the distance between the picture and the tested eye is about 59 cm, and the tested eye cannot see the light source.

6. Procedure of experiment

The experimental flow chart is shown in fig. 3.

The number of comparisons generated using pairwise comparisons may be expressed as:

wherein J is the comparison frequency of the experiment, R is the repetition frequency of the experiment, and N is the number of samples required to be evaluated.

The experiment had four illumination sets and a repeat illumination set, with 7 CCTs in each illumination set, so each picture was tested for 105 comparisons. First, each participant was given 1 minute of time to adapt to the experimental environment, and then after learning the entire experimental process, three pairs of training lighting conditions were randomly selected for judgment.

To avoid order bias, each pair of each group is played in a random order, and the order of occurrence of the two lighting conditions in each pair is also random. The duration of each lighting condition was 4 seconds. The experimenter can inform that the former working condition is X and the latter is Y, only letters of preference working conditions need to be answered orally when the experimenter judges, and the result is recorded by the experimenter. The working condition to be tested can be played for multiple times until the tested object can make a positive judgment. Two minutes of rest between each illumination group. In order to check the validity of the tested data, one illumination group is randomly selected to carry out repeated experiments after four illumination groups are judged, and the misjudgment frequency of each tested data is not more than three times. During data statistics, two working conditions of X and Y are assumed in each pair of working condition groups, if X is more than Y (more prefers X), A is marked with 2 points, and B is marked with 0 point; if X < Y (more preferably Y) is the opposite; if X is Y (X and Y have the same preference), each is given a score of 1. After all comparisons were completed, the experimenter played randomly the lighting conditions for each group and asked the participants to select the lighting conditions deemed to have a neutral (neither cold nor warm) CCT, and the experimental scene of the subject evaluating the pictorial representation is shown in fig. 4.

In the subjective questionnaire, subjects scored the color preference for each lighting condition. In this study, "color preference" represents the preference of a subject for a painted-appearance color; after the pair-wise comparison is finished, the test is rested for five minutes, and then the test needs to sequentially score 28 randomly ordered working conditions by using seven grades of evaluation, wherein the scores are-3 to 3 (including 0). The 28 working conditions are divided into four groups in random sequence, the groups are used as units for scoring, fatigue of the tested object is prevented, and the rest is carried out for two minutes between each group. Before each group is scored, the whole group of working conditions are played twice in turn to make the test familiar.

7. Data reliability analysis

Misjudgment or abnormal values often occur in subjective evaluation experiments, and the evaluation result should be checked firstly to ensure the reliability of data. The consistency coefficient proposed by Kendall is widely applied to test experimental data, particularly a pair comparison experiment, so that the data reliability of an evaluator is judged according to a Kendall consistency coefficient method before experimental result statistics, and a calculation formula is shown in 4-2.

Wherein W is a consistency coefficient (the larger the W value is, the stronger the data consistency is); n is the number of the evaluated objects; k is the number of the scorers or the standard number according to which the scorers score; s is the sum R of the rating of each evaluated object_iAnd the average of all these sums

Is a sum of squared deviations, i.e.

Where m is_iThe number of the duplicate ranks, n, in the evaluation result of the i-th evaluator_ijThe number of the same levels of the jth repetition level in the evaluation result of the ith evaluator. The calculation results are shown in tables 4-2 and 4-3.

Since the number of the scorers K is 25, the number of the evaluated objects N is 7 and exceeds the range of the Kendal consistency coefficient significance critical value table, the W value is converted into chi²Value, then X²Checking and calculating formula X²K (N-1) W. If the degree of freedom d of the experimental data is equal to N-1 equal to 6 and the significance is 0.01 by checking a chi-square test critical value table, if the experimental data x is²The values are all larger than 16.81, which indicates that the W values all reach the extremely significant level, and tables 4-2 and 4-3 show that the tested work condition evaluation in the research has stronger consistency.

TABLE 4-2 values and values of data obtained by pairwise comparison

Tables 4-3 values and values of the subjectively evaluated data

3. Chrominance preference evaluation model

The support vector machine can solve the classification problem existing in a high-dimensional space by constructing a hyperplane. The method has strong generalization capability and is suitable for establishing a small sample model. The present study employed psychophysical experiments to obtain data that were small and non-linear. For such data, the support vector machine converts it from a low dimensional space to a high dimensional space and constructs a hyperplane for classification. Therefore, the support vector machine is suitable for classifying data in the research and establishing a chromaticity preference level map for guiding museum lighting design.

The arithmetic mean of the color preferences of two Chinese traditional paintings evaluated by the test was calculated, the calculated model included three preference levels, which were respectively "dislike", "acceptable", and "like", and fig. 5(a) is a data scatter plot. The C-SVC model (C-Support Vector Classification) is built in the research by using Matlab2017a software and an LIBSVM toolkit. In the training mode, the illumination and the CCT are used as characteristic variables, and the processed preference level is used as a target variable. The optimal C value and gamma value are obtained after the parameter optimization, and the optimization result is shown in fig. 5 (b). The maximum model accuracy under these parameters is 85.7143%. And (3) predicting the test set by using the model, wherein the classification accuracy is 82.1%, and the reliability of the model is proved to be strong. A chromaticity preference level map obtained by the preference evaluation model is shown in fig. 6, which shows the variation of preference levels for different combinations of CCT and illuminance.

In the chromaticity preference model, two indexes of CCT and illumination are used as input quantities, the chromaticity preference level is used as an output quantity, and the chromaticity preference degree of any spectrum when the museum environment watches the traditional Chinese painting can be visually expressed.

In FIG. 6, the two curves separating the preferred regions plotted with the support vector machine are the classification hyperplanes. These two curves are fitted according to the scatter points identified by the model, as shown in fig. 7. The resulting equations 5-1 and 5-2 divide these three regions of preference, respectively.

y＝80.7-0.337x¹+3.83593×10^-4x²-1.17815×10^-7x³-3.50934×10^-12x⁴ +7.72954×10^-15x⁵-1.32446×10^-18x⁶+7.10801×10^-23x⁷ (5-1)

Where y denotes an illuminance, and x denotes a correlated color temperature.

Equation (5-1) is a polynomial function describing the classification boundaries of "dislike" and "acceptable" regions, with the lowest luminance values belonging to the "acceptable" regions gradually decreasing as the correlated color temperature increases, down to about 80lx and going steady. Equation (5-2) is an ellipse equation describing the classification boundaries of "acceptable" and "favorite" regions, with the ellipse centered at (4047.04339K,177.31952 lx). The approximate range of the "favorite" region is 3000K-5000K, 120lx or more.

A. Light source color display preference and perception model of Chinese traditional painting

1. Experimental light source

The experiment is carried out in a museum laboratory of the sustainable construction and environment research center of the institute of architecture of Tianjin university, and the environment of the museum laboratory is built as close to the environment of the museum as possible. The room area is 9 x 9m, the space height is 4.2m, and the wall is a light partition wall. 10 showcases are placed in the room, and the showcases are of three types, wherein the length, the width and the height are 600 x 2200mm, 1200 x 600 x 2400 mm and 2250 x 600 x 1250mm respectively. The light source used was a 60W LEDCOB down lamp (diameter 282mm, height 130mm, power 60W). A CL500A spectroradiometer is adopted to measure the luminous environment in a museum laboratory, the average illumination of the ambient illumination is 9.9lx, the illumination uniformity is 0.87, the color temperature range is 2714 +/-32K, and R is_aIs 83, D_uvThe range is 0.0015. + -.3. The evaluation showcase size is 600 x 2200mm, the base is 800mm, the middle is 1100mm high four-sided glass, the top is placed with an 11-channel spectrally tunable light source (tholsite led tube, manufactured by qianming smart lighting technologies ltd., usa), and the light emitting surface size is 270 x 270 mm. The illumination of exhibits for traditional chinese painting has special provisions, according to CIE 157: 2004. in the museum lighting design specification GBT _23863-_uvThe | < 0.005, CCT 3000 + -44K, illumination 50 + -0.8 lx. The target Rf-Rg combinations for 24 conditions are 65: 80. 90, 100, 110, 120, 75: 90. 100, 110, 85: 90. 100, 110, 95: 100, each combination is provided with two gamut shapes, one for enhanced red saturation or for diminished red saturation. The overall environment and geometry of the museum laboratory is shown in fig. 8.

Table 5-224 working conditions of photometric and colourmetric information

2. Drawing sample

Fig. 2 shows two strictly selected experimental painting samples with the sizes of 35 × 35cm and 43.5 × 60.5cm, wherein the subjects of the two paintings are flowers, birds and figures, and are two main painting subjects of the traditional Chinese painting. The first painting is a real painting drawn by modern folk painters, the color of the painting is mainly yellow, and the painting also comprises green, red, black and white, and the theme of the painting is dragonfly and lotus. The second drawing was originally from the Tang Dynasty, probably more than a thousand years old today, and existed in museums in Liaoning province. The subject is a human picture, two noble women in the picture enjoy the flowers, the color of the picture is mainly red, and the picture also comprises a large amount of yellow and brown colors and a small amount of black and white colors. The painting is placed in a showcase, and because the painting is small, a gray light absorption cloth is laid on the bottom of the painting to prevent interference. The detailed illustration is shown in fig. 2.

3. Is tested

37 subjects participated in the experiment, and the ratio of male to female was 14:23, and they were all researchers at Tianjin university (all Chinese people). The subjects were from 20 to 32 years of age, with a mean age of 24.8 years and a standard deviation of 2.3 years. They all passed the Ishihara color blindness test and filled out an informed consent. In the experiment process, the tested person stands in the fixed area in front of the showcase, the distance from the showcase is about 300-. The experimental light source can not be seen in the experimental process.

4. Procedure of experiment

The experimental time was 1:00-6:00 every afternoon to exclude the interference of biorhythms. After arriving at the museum set up, the experimenters stated the experimental process and the cautions during the five minutes of the test to adapt to the low-light environment and wear the black experimental coat. To avoid sequence bias, a unique scoring sequence was established for each test. Before the experiment begins, all working conditions are played out of order to enable the tested person to be familiar with, and three working conditions are randomly selected for practice, so that the tested person can understand the scoring requirement. The subjects orally scored the color appearance in terms of preference, naturalness, and two aspects using five-level ratings (-2, -1, 0, 1, 2), and the experimenters scored. In this study, "preference" represents how liked the subject was for a pictorial appearance; "naturalness" describes the degree of naturalness of the painting appearance; in addition, under the condition of not informing the tested object, four working conditions are randomly extracted and repeatedly scored, so that the total number of evaluation is 28 times. And after half of the working condition evaluation is finished, the rest is carried out for two minutes to prevent the test from generating fatigue. The test requires the same procedure for both pictures, and the whole procedure takes about 30 minutes.

5. Data reliability analysis

The in-test variability was evaluated by calculating the mean of the normalized sum of squared residuals (STRES) between 8 repeat scores and the original score for each test for two pictures, and the in-test variability was evaluated by calculating the mean of the STRES between the mean of 37 test samples and the overall sample. The STRESS values ranged from 0 to 100, with 0 indicating that the two sets of data were identical. The calculation method of STRESS is as follows:

wherein,

where F is the adjustment factor, Δ V for the test of the in-test reliability_iAnd Δ E_iRespectively judging the result of each test for the ith stimulation twice; for the inter-test reliability test, Δ V_iAnd Δ E_iThe average evaluation value of a certain test and the average evaluation value of the whole test are shown. In addition, the standard deviation of these values can be used to represent the variability between the subjects, so the standard deviation is also calculated in this chapter.

The mean and variance of the in-and inter-test variability for the preference and naturalness are shown in Table 5-2. The intra-subject variability is smaller than the inter-subject variability. The maximum value of the differences among the tested flowers and birds is the preference degree of the flower and bird pictures, and the value is 38.1; the maximum value of the difference in the test is the vividness, and the value is 26.0; the minimum value is naturalness and takes 24.0. The differences of the two indexes of the preference degree and the naturalness are very similar, and the differences between the tested data of the two pictures are very consistent, so that the tested data of each bit is considered to be effective and is used for data analysis.

TABLE 5-2 mean and variance of the in-and-out-of-test variability of the preference and naturalness of two pictures

Mean represents Mean and SD represents standard deviation.

6. Color development perception model

And selecting a proper color development quality index for mathematical modeling, and establishing a best fitting regression model and an influence relation formula of the color development quality index and the color development quality index to achieve a good prediction effect on subjective evaluation. The predictor is in R_f、R_g、R_cs,h1、R_cs,h16And selecting fourteen secondary items and pairwise matching items thereof, and respectively establishing preference degree and naturalness prediction models. In the regression fitting process, the adjusted R is used²To avoid the artificial high-fitting models, which usually have a higher R²Values, but only because they have a large number of predictors. Therefore, the best prediction model should be the simplest model, i.e., the least fitting terms are used, while the largest decision coefficient (higher R) is reached²And adjusted R²) In this subsection, 2017 aamatlab software was used.

(1) Preference model

First, a single predictor is preliminarily analyzed, R_fIs an important linear predictor (r) of preference rating²0.64, after adjustment²0.63), and R_gIs less predictive (r)²0.15, after adjustment²＝0.11)，R_cs,h1、R_cs,h16Minimal effect (r)²0.07, after adjustment²＝0.03；r²After adjustment, r is 0.10²＝0.06)。According to 14 prediction factor parameters obtained by a stepwise regression method, sequentially adding R into the model_f、R_g ²、R_f*R_gThe visualization model is shown in fig. 9.

In the preference model, R_f、R_gThe two color indexes are used as input quantity, the preference degree is used as output quantity, and the preference degree of any spectrum when the museum environment watches the Chinese traditional painting can be visually expressed.

Tables 5-9 preference model predictor parameters

Bolded indicates that the significance level is at the level of 0.01, and the correlation coefficient of the marker is considered significant.

(2) Nature (Natural) model

R_fIs an average fidelity (fidelity), i.e., the degree of color reduction, i.e., with R_fThe fidelity will also increase. In the modeling process, a stepwise regression method is adopted, firstly, R is added into the model_fLater, the model has achieved better prediction capability (r)²0.74, after adjustment²0.72). This confirms the previous analysis that a light source with higher average fidelity would be considered more natural. Model addition R_gAfter that, the model achieves the expected fitting effect (r)²0.87, after adjustment²0.86), the remaining predictor was added and the model fit did not change significantly (r)²R is 0.89 after adjustment²0.86). In the selection process of the prediction factors, 14 prediction factors are selectedIn order to ensure the simplicity of the model and prevent overfitting, the nature model Natural is as follows, and the model visualization graph is shown in fig. 10:

Natural＝-3.12+0.0299R_f+0.0118R_g (5-4)

in the naturalness model, two color indexes of Rf and Rg are used as input quantity, naturalness is used as output quantity, the naturalness of any spectrum in the museum environment for watching Chinese traditional paintings can be visually expressed, and the visual model is shown in figure 10.

TABLE 5-10 Nature model predictor parameters

To verify the prediction model, first 24 spectra are calculated to obtain the preference and naturalness calculated by the prediction model, and then compared with the average preference and naturalness obtained by the psychophysical experiment, as shown in fig. 11. The closer the scatter is to the fitted line, the higher the model prediction. Through the graph, it can be known that the prediction model proposed by the research can better represent the perception quantity of subjective evaluation indexes such as preference, naturalness and the like for spectrums with different fidelity, color gamut indexes and color gamut areas.

Taking traditional Chinese painting with high photosensitivity and high value as an example, obtaining the color preference characteristics of typical fragile Chinese cultural relics under the working conditions of different illumination and correlated color temperature combinations through psychophysical experiments and a support vector machine, establishing a chromaticity preference evaluation model and drawing a chromaticity preference grade graph, thereby providing a research basis for the color preference of special light environments of museums. An explanation is provided for the best mode.

1. Painting sample specimen

Two representative paintings are selected as experimental samples, the first original act is loquat mountain bird picture, created by Nansong forest Chinese toon, and existing in Beijing old palace Bozhou. The study used printed samples with a picture size of 27 × 27cm, mainly yellow and brown as the dominant colors, and also green, white and black. The second picture is a true drawing drawn by a folk painter, and is a freehand drawing and a picture color measurement site thereof. The picture subject is flower and bird, the picture size is 35 × 35cm, mainly takes the yellow of the base material as the main color, and also comprises red, green, black and white, and the concrete sample picture is shown in figure 2.

2. Selection of experimental light sources

The light source adopts a light box with 11 channels and tunable spectrum, which is THOUSLITE LEDCube to generate 28 light sources used in the experiment. The size of the light box is 38cm by 50cm by 55 cm. The gray light-absorbing cloth is adhered to the inner wall of the lamp box, so that the influence of glare on the irradiation effect is effectively avoided. CIE 157: 2004, the illumination values for the museum environment are specified to be in the range of 50-200lx, and therefore four illumination values of 50, 100, 150 and 200lx are selected as experimental variables in equal ratios in the range. Seven CCT variables of 2700, 3000, 3500, 4000, 4500, 5000 and 5500K are respectively set at each illumination value. These 28(4 x 7) operating conditions | D_uv|≤0.0051， CRI R_aIs 97 +/-1, R_fIs 96 +/-2 and R_gIs 100 +/-3. The parameters of the experimental light source are shown in Table 8-1, and the spectral power distribution plots for the 28 operating conditions are shown in FIG. 12. The experimental light source was the only illuminant in the evaluation.

Photometric and colourmetric information of table 8-128 working conditions

3. Is tested

25 subjects participated in the experiment with a male to female ratio of 18: 7. The subjects were from 19 to 31 years of age, with a mean age of 23.8 years and a standard deviation of 2.9 years. They all passed the Ishihara color blindness test and written informed consent. Fig. 4 shows an experimental scene of evaluation of pictorial representations by a subject, and a chair used in an experiment is placed in front of an observation lamp box to ensure that the visual angle of the experiment is fixed. The picture is placed in the lamp box, the distance between the picture and the tested eye is about 59 cm, and the tested eye cannot see the light source.

4. Procedure of experiment

Specific experimental operational flow referring to fig. 3, the number of comparisons generated using pairwise comparisons can be expressed as:

To avoid order bias, each pair of each group is played in a random order, and the order of occurrence of the two lighting conditions in each pair is also random. The duration of each lighting condition was 4 seconds. The experimenter can inform that the former working condition is X and the latter is Y, only letters of preference working conditions need to be answered orally when the experimenter judges, and the result is recorded by the experimenter. The working condition to be tested can be played for multiple times until the tested object can make a positive judgment. Two minutes of rest between each illumination group. In order to check the validity of the tested data, one illumination group is randomly selected to carry out repeated experiments after four illumination groups are judged, and the misjudgment frequency of each tested data is not more than three times. During data statistics, two working conditions of X and Y are assumed in each pair of working condition groups, if X is more than Y (more prefers X), A is marked with 2 points, and B is marked with 0 point; if X < Y (more preferably Y) is the opposite; if X is Y (X and Y have the same preference), each is given a score of 1. After all comparisons are completed, the experimenter plays randomly the lighting conditions for each group and asks the participants to select lighting conditions that are considered to have a neutral (neither cold nor warm) CCT.

In the subjective questionnaire, subjects scored the color preference for each lighting condition. After the pair-wise comparison is finished, the test is rested for five minutes, and then the test needs to sequentially score 28 randomly ordered working conditions by using seven grades of evaluation, wherein the scores are-3 to 3 (including 0). The 28 working conditions are divided into four groups in random sequence, the groups are used as units for scoring, fatigue of the tested object is prevented, and the rest is carried out for two minutes between each group. Before each group is scored, the whole group of working conditions are played twice in turn to make the test familiar. After the scoring is finished, the tested object needs to answer which two colors are dominant in the evaluation process for judgment and is sequenced. The same flow is needed to be carried out on two pictures to be tested, the whole process needs about 2 hours, and in order to prevent visual fatigue, the experiment is divided into two days, wherein one hour is carried out each time.

5. Data reliability analysis

Is a sum of squared deviations, i.e.

Since the number of the scorers K is 25, the number of the evaluated objects N is 7 and exceeds the range of the Kendal consistency coefficient significance critical value table, the W value is converted into chi²Value, then X²Checking and calculating formula X²K (N-1) W. The freedom degree d-N-1-6 of experimental data, checking chi-square and checking critical value table^[89]When the significance is 0.01, if the experimental data x²The values are all larger than 16.81, which indicates that the W values all reach the extremely significant level, and tables 4-2 and 4-3 show that the tested work condition evaluation in the research has stronger consistency.

TABLE 4-2 values and values of data obtained by pairwise comparison

Tables 4-3 values and values of the subjectively evaluated data

6. Chrominance preference evaluation model

Where y denotes an illuminance, and x denotes a correlated color temperature.

In the research, a preference evaluation model is established by using a support vector machine, and a chromaticity preference level graph is drawn by using the preference evaluation model. The map can describe the corresponding relation between the color preference and the illumination intensity and CCT, and provides reference for museum lighting design. The "preferred" area in the chromaticity preference level diagram is a CCT range of approximately 3500K to 4500K and an illuminance of approximately 120lx to 200lx, as shown in fig. 4 to 11. For the "acceptable" area, 80lx is the minimum illumination. For a museum to obtain a better visual effect, a higher illumination value is needed at a lower CCT, which should be as close to 4000K as possible.

The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.

Claims

1. An illumination color quality evaluation method suitable for Chinese traditional painting is characterized in that psychophysical experiments are adopted to obtain data for scoring the color preference of a subject on each illumination condition, a support vector machine is adopted to convert the data from a low-dimensional space to a high-dimensional space, a hyperplane is constructed for classification, and a chromaticity preference grade chart is established for guiding the illumination design of a museum.

2. The method according to claim 1, wherein an arithmetic mean of color preferences of two Chinese traditional paintings to be evaluated is calculated, the calculated model comprises three preference levels, namely "dislike", "acceptable", and "like", and a C-SVC model (C-Support Vector Classification) is constructed by using Matlab2017a software and LIBSVM toolkit, and in the training mode, parameter optimization is performed by using the illumination and CCT as characteristic variables and the processed preference level as a target variable;

3. The method for evaluating the quality of the lighting color suitable for the traditional Chinese painting as claimed in claim 2, wherein the detailed steps are as follows:

1. light source

photometric and colourmetric information for the conditions in tables 5-128

1. Drawing sample

2. is tested

3. process for producing a metal oxide

4. The method for evaluating the quality of the lighting color suitable for the traditional Chinese painting as claimed in claim 1, wherein the detailed steps are as follows: the method also comprises the following data reliability analysis steps:

before statistics of scoring data results, judging the data reliability of an evaluator according to a Kendall consistency coefficient method, wherein a calculation formula is shown in 4-2:

Is a sum of squared deviations, i.e.

Where m is_iThe number of the duplicate ranks, n, in the evaluation result of the i-th evaluator_ijIs the ith evaluatorThe number of the same levels of the jth repetition level in the evaluation result of (1); the calculation results are shown in tables 4-2 and 4-3;

TABLE 4-2 values and values of data obtained by pairwise comparison

Tables 4-3 values and values of the subjectively evaluated data