A method for automatic migration of thematic map style

Abstract

The application discloses a kind of thematic map style automatic migration method, comprising the following steps: image space feature evaluation is carried out using image entropy, image classification is completed and image gamut is generated, and the color scheme of the color scheme of the source of thematic map is used as the source;According to the main rules related to the scientific nature of thematic map, the measurement model of the scientific nature of thematic map is established, and the scientific nature score of thematic map is used as the first optimization target;According to the main rules related to the artistic nature of thematic map, including map color emotional expression, color preference and color harmony, the measurement model reflecting the artistic beauty of thematic map is constructed, and the score of the artistic beauty of thematic map is used as the second optimization target;Thematic map style automatic migration is organized into a double-target multi-constraint optimization problem, and heuristic algorithm is used to search solution set, so as to obtain a series of thematic map style migration results.The present application can realize the fast generation of color matching scheme of different types of thematic map.

Description

Technical Field

[0001] This invention relates to the field of geographic information system technology, and in particular to a method for automatic migration of thematic map styles. Background Technology

[0002] Color plays a crucial role in creating personalized map designs that are both scientifically sound and aesthetically pleasing. Designing expressive color schemes from scratch is challenging; a common approach is to find inspiration, such as in a master painting or an impressive photograph, and then transfer the style of that inspiration to another image. While significant work has been done in computer graphics to transfer color from image to image, employing methods based on histograms, palettes, and neural networks, research on color transfer methods in map design is relatively limited. Furthermore, research on quantitative descriptions of thematic mapmaking rules that balance information transmission and artistic quality is lacking.

[0003] Currently, in the field of map color design, methods for transferring personalized and creative colors from one or more images to maps include those based on manual experience, neural networks, and map content. Methods based on manual experience involve cartographers manually transferring image colors to maps; the quality of this method depends on the cartographer's experience, knowledge, and understanding of the data, making it suitable for artists and experienced professional cartographers, but not for the general public. Neural network-based methods, such as pix2pix, CycleGAN, and GAN, transfer colors from images or paintings to maps. However, this method ignores the readability of map content, producing results that resemble paintings rather than maps. Methods based on map content quantify several design rules for map colors, such as semantic relationships and habitual colors, proposing a vector map style transfer method that balances map scientific accuracy and aesthetics. However, this method, designed for vector maps, does not consider the differences in content and cartographic rules for thematic maps, limiting its applicability. Currently, a universal, high-quality method for thematic map color design that considers map content has not yet been established.

[0004] To address the aforementioned issues, a new method is needed that can rapidly generate thematic maps while maintaining their scientific accuracy and artistic appeal, thus meeting the public's demand for personalized mapping. Chinese patent CN200910036810 (publication number CN101477559B), entitled "A Geographic Information Perception Method Based on Region Coloring," provides a method for geographic information perception based on region coloring. This method consists of two parts: region coloring and geographic information perception. Region coloring categorizes all regions on a map based on their colors; the region categories are adjusted as required; each region is assigned a unique RGB value according to its adjusted category; and each region is then colored according to its assigned RGB value. Geographic information perception involves obtaining the RGB value of a pixel on the map using a computer, finding the geographic location of that point using the RGB value, and simultaneously displaying the corresponding geographic information. This method can be applied to map coloring and lightweight geographic information systems, reducing programming complexity and saving resource consumption, demonstrating good and efficient practicality. Chinese patent CN202210299170 (application number CN114399607B), entitled "A Method for Adaptively Modifying Map Color Scheme to Reduce Map Energy Consumption," discloses a method for adaptively modifying map color scheme to reduce map energy consumption. This method uses the red, green, and blue component values ​​of the colors of various map elements as optimization variables for the color scheme, establishing a model to measure map energy consumption and a model to evaluate map information transmission quality. Based on map color rules, the basic conditions that each component value must meet after adjustment are determined. Addressing the issue of not being able to consider the transmission quality of electronic map information in relation to map energy consumption reduction, a heuristic search is used to evaluate each searched color scheme in the state space to find the globally optimal solution. The obtained Pareto front solution set is sampled and a series of optimal solutions are output, yielding the color scheme with the highest map information transmission quality under different energy consumption levels. While the patent titled "A Geographic Information Perception Method Based on Region Coloring" discloses a method for map region coloring and geographic information perception, this method offers a single color scheme for region coloring, lacking diversity. Furthermore, this method is only applicable to lightweight geographic information systems, limiting its applicability. The patent titled "A Method for Adaptively Modifying Map Colors to Reduce Map Energy Consumption" discloses a method that balances map information transmission quality and energy reduction. However, this method ignores the differences in thematic mapmaking rules and lacks a method for considering the map's artistic appeal, failing to simultaneously reduce energy consumption and maintain map aesthetics. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method for automatic migration of thematic map styles, which can realize the rapid generation of color schemes for different types of thematic maps.

[0006] To address the aforementioned technical problems, this invention provides a method for automatic style migration of thematic maps, comprising the following steps:

[0007] Step 1: Use image entropy to evaluate image spatial features, complete image classification, and generate image color gamut. Use this color gamut as the source of the color scheme for the thematic map.

[0008] Step 2: Based on the main rules for measuring the scientific validity of thematic maps, including visual contrast between map layers, semantic relationships between elements, and customary color rules, establish a measurement model for the scientific validity of thematic maps, and take the scientific validity score of thematic maps as the primary optimization objective.

[0009] Step 3: Based on the main rules for measuring the artistic value of thematic maps, including the emotional expression of map colors, color preferences, and color harmony, construct a measurement model that reflects the artistic beauty of thematic maps, and use the score of the artistic beauty of thematic maps as the second optimization objective.

[0010] Step 4: Organize the automatic thematic map style transfer into a bi-objective, multi-constraint optimization problem, and use a heuristic algorithm to search for the solution set, thereby obtaining a series of thematic map style transfer results.

[0011] Preferably, in step 1, image spatial feature evaluation is performed using image entropy to complete image classification and generate an image color gamut. This color gamut is then used as the source of the color scheme for the thematic map. Specifically, this includes the following steps:

[0012] Step 11: Utilize image entropy to reflect the average information content of the image source, representing the spatial characteristics of the image grayscale distribution, thereby classifying reference images to adapt to the migration of thematic maps with different spatial characteristics. For discrete two-dimensional images, p ij Given the spatial frequency of gray levels of adjacent pixels, image entropy is described as follows:

[0013]

[0014] The H-value reflects the complexity of image distribution, which in turn assesses the complexity of thematic map elements. Image classification enables style transfer for different thematic maps.

[0015] Step 12: Use the eigenvalues ​​of the Hessian matrix to detect changes in the reference image and individual regions of abrupt changes and gradations in the image; for regions of abrupt changes, separate points and lines by the direction of the change; for regions of gradations, separate surfaces and background by visual saliency. After extracting the four types of elements, use the appropriate color difference JNCD in CIELab as the color distance threshold to extract all colors and calculate the color proportion of each type of color. Divide the reference image into different regions by saliency measurement, namely point, line, surface and background colors, and use them as the color source for generating the color scheme of the thematic map.

[0016] Preferably, in step 2, based on the main rules for measuring the scientific validity of thematic maps, including visual contrast between map layers, semantic relationships between elements, and customary color rules, a measurement model for the scientific validity of thematic maps is established, and the scientific validity score of the thematic map is used as the primary optimization objective. Specifically, this includes the following steps:

[0017] Step 21: Quantitatively describe the visual contrast between points, lines, surfaces, and background layers. Visual contrast is mainly related to the spatial occupancy and color of symbols. The thematic map should exhibit significant color difference and brightness contrast. Let fv(ci,cj) represent the visual contrast between levels i and j. The distance between the two colors in the CIELab color space, where L is the brightness and δ is the distance between them. e and δ l These represent the color difference and brightness threshold between the two layers, respectively, where w is the weight of each element's visual level, i.e., the average spatial occupancy of the two layers. The visual contrast between thematic map layers is described as follows:

[0018]

[0019] Based on the visual contrast generated between pairs of layers, the overall layer visual contrast is described by calculating their average value, as follows:

[0020]

[0021] Step 22: Quantify the semantic relationships between features. Three semantic relationships between map objects are constrained by color difference and brightness between features: difference, association, and order. Let i and j represent the colors of two feature sets, and r... ij It represents the semantic relationship between two colors, where μ is the color distance threshold between differences and associations, and δ... l Let α be the brightness threshold for the difference, and α be the hue angle threshold for determining the order relationship. The hue difference between two colors, w d w a and w o c represents the weights of the three semantic relations. i c is the adjusted color of the i-th element set. j Let L be the adjusted color of the j-th element set, and L be the brightness of the color. For color saturation, the semantic score is described as follows:

[0022]

[0023] Based on the semantic relationship scores of the above pairwise elements, the overall semantic relationship score of the map is described as follows:

[0024]

[0025] Step 23: Quantifying customary color usage. Customary color usage rules affect the quality of map information transmission. i ' is its habitual color, γ is the color difference threshold for maintaining the habitual color, and the overall quality of the map's habitual color is described as follows:

[0026]

[0027] Step 24: Based on the above quantification of visual contrast between layers, semantic relationships between elements, and conventional color usage, the total score for measuring the scientific validity of the thematic map is obtained by multiplying the scores of the three factors and is used as the primary optimization objective, described as follows:

[0028] F(C) = f v (C)*f r (C)*f c (C).

[0029] Preferably, in step 3, based on the main rules for measuring the artistic value of thematic maps, including the emotional expression of map colors, color preferences, and color harmony, a measurement model reflecting the artistic beauty of thematic maps is constructed, and the score of the artistic beauty of thematic maps is used as the second optimization objective. Specifically, this includes the following steps:

[0030] Step 31: Measuring the similarity of color sentiment. The aesthetic appeal of a map must possess a relative logical and emotional resonance. This is achieved by quantifying color sentiment factors—activity, weight, and heat—to measure the color sentiment similarity between thematic maps and images, and normalizing the results as follows:

[0031]

[0032] Step 32: Measuring color preference. A number of factors influence color preference, such as hue difference, relative size, and spacing. Users can choose inspiration references based on their preferences, and then use the reference images as input to the model. The normalized mean absolute error (MAE) is used to measure similarity. map (c i ) represents the percentage of elements in the resulting map, P image (c i The following is a reference to the color proportions in the image:

[0033]

[0034] Step 33: Quantification of color harmony. A harmony model is introduced to quantify the overall color harmony of the map, and it is standardized as follows:

[0035] f h(C) = Harmony(C) / 5;

[0036] Step 34: Based on the above quantification of color emotion, color preference, and color harmony, the total score F for measuring the artistic beauty of the thematic map is obtained by multiplying the scores of the three. a (C) and take it as the second optimization objective, as shown below:

[0037] F a (C)=f e (C)*f s (C)*f h (C).

[0038] Preferably, in step 4, the automatic transfer of thematic map style is organized as a bi-objective, multi-constraint optimization problem, and a heuristic algorithm is used to search for the solution set, thereby obtaining a series of thematic map style transfer results. Specifically, this includes the following steps:

[0039] Step 41: Using the constraints set by the first and second optimization objectives established in Steps 2 and 3, organize the thematic map into a dual-objective, multi-constraint optimization problem, and use a heuristic algorithm to search for the global optimal solution to the problem;

[0040] Step 42: Taking into account the importance of both the scientific and artistic aspects of the thematic map, thematic maps with different tendencies and styles can be output according to the needs.

[0041] The beneficial effects of this invention are as follows: This invention can perform style matching for different types of thematic map content, and is applicable to vector or raster electronic maps of different scales and themes; it can take into account the content of thematic maps and the design rules of thematic maps, which greatly improves the production quality of thematic maps; and it improves the automation level of thematic map coloring by using the color migration method in computer graphics. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0043] Figure 2 These are different types of thematic maps for this invention.

[0044] Figure 3 This is a schematic diagram illustrating the style transfer results of different types of thematic maps according to the present invention. Detailed Implementation

[0045] This invention will consider the content of different topics, establish a scientific measurement model and an artistic beauty measurement model for thematic maps based on the color rules of thematic maps, classify and extract the colors of style elements based on the input reference image, and use a heuristic algorithm to find the optimal style transfer solution set.

[0046] According to such Figure 1 The process shown is as follows: Figure 3 The style transfer of thematic maps involves the following steps:

[0047] Step 1: Use image entropy to classify images and generate image color gamut, and use this color gamut as the source for generating the color scheme of the thematic map.

[0048] Step 2: Based on the main rules for measuring the scientific validity of thematic maps, including visual contrast between map layers, semantic relationships between elements, and customary color rules, establish a measurement model for the scientific validity of thematic maps, and take the scientific validity score of thematic maps as the primary optimization objective.

[0049] Step 3: Based on the main rules for measuring the artistic value of thematic maps, including the emotional expression of map colors, color preferences, and color harmony, construct a measurement model that reflects the artistic beauty of thematic maps, and use the score of the artistic beauty of thematic maps as the second optimization objective.

[0050] Step 4: Organize the automatic thematic map style transfer into a bi-objective, multi-constraint optimization problem, and use a heuristic algorithm to search for the solution set, thereby obtaining a series of thematic map style transfer results.

[0051] The more specific implementation steps are as follows:

[0052] Image classification and color gamut generation:

[0053] To characterize the spatial features of grayscale information, the pixel and its neighborhood information are introduced into the image, forming a new feature tuple denoted as (i,j); where i represents the grayscale value of the pixel (0≤i≤255), and j represents the average grayscale value of the neighborhood (0≤j≤255). Simultaneously, to reflect the comprehensive characteristics of the grayscale value at a given pixel location and the grayscale distribution of its surrounding pixels, f(i,j) is the frequency of occurrence of the feature tuple (i,j), W and H are the image dimensions (width and height), and the probability of f(i,j) occurring in the image is defined as P. ij The description is as follows:

[0054] Pij = f(i,j) / (WH)

[0055] Image entropy reflects the average information content of an image source and represents the spatial characteristics of the image's gray-level distribution. For discrete two-dimensional images, p ij Given the spatial frequency of gray levels of adjacent pixels, image entropy is described as follows:

[0056]

[0057] A large H-value indicates a complex image distribution, making it suitable for style transfer in thematic maps with many and complex elements, and vice versa.

[0058] Based on the visual perception process, the colors in the reference image are divided into four groups: points, lines, surfaces, and background to fit the map content. Specifically, the brightness changes in the image are first detected to identify abrupt and gradual transitions. For abrupt transitions, points and lines are separated by the direction of the transition; for gradual transitions, surfaces and background are separated by visual saliency. The extracted colors are used as the color source for generating the color scheme of the thematic map.

[0059] Establish a scientific measurement model for thematic maps:

[0060] First, thematic map design places greater emphasis on visual hierarchy, stressing the de-emphasis on the geographic base map and highlighting thematic elements. Visual contrast is a key factor in controlling element prominence and plays a crucial role in establishing perceptual differences. This method considers the visual contrast between four levels in thematic maps: points, lines, areas, and background. It quantitatively describes the visual contrast between pairs of elements using the spatial occupancy and color values ​​of their symbols. fv(ci,cj) represents the visual contrast between levels i and j. The distance between the two colors in the CIELab color space, where L is the brightness and δ is the distance between them. e and δ l These are the color difference and brightness thresholds between the two levels, respectively. w is the weight of each element's visual level, i.e., the average spatial occupancy of the two levels. For point elements, the size and number of points are considered; for line elements, the length and width of the line are considered; for surface and background elements, their area proportion is considered, with values ​​ranging from [0,1].

[0061]

[0062] The overall visual contrast of the layers is described by calculating the average value, as follows:

[0063]

[0064] Furthermore, semantic relationships between elements are an important aspect of using color to convey information. There are three types of semantic relationships: difference, order, and association. These three semantic relationships are widely present and encoded in thematic maps. Color distance is widely used to analyze the semantic relationships represented by color schemes. This invention is based on the CIELab color space and uses... The Euclidean distance between two colors is represented by L, where L represents brightness (0-100), a represents the distance from green (-) to red (+), and b represents the distance from blue (-) to yellow (+). The color distance is calculated as follows:

[0065]

[0066] Hue difference between colors The calculation is as follows:

[0067]

[0068] Color saturation difference The calculation is as follows:

[0069]

[0070] This invention employs a semantic relationship detection method based on color distance, where i and j represent the colors of two feature sets, and r ij It represents the semantic relationship between two colors, where μ is the color distance threshold between differences and associations, and δ... l The brightness threshold is the difference, α is the hue angle threshold that determines the order relationship, and w d w a and w o c represents the weights of the three semantic relations. i c is the adjusted color of the i-th element set. j Let L be the adjusted color of the j-th element set, and L be the brightness of the color. The encoding of each semantic relation is scored and described as follows:

[0071]

[0072] Then, the quality of all semantic relations is scored, and the results are normalized. The total semantic relation score f r (C) can be described as:

[0073]

[0074] Conventional color rules are used to evoke semantic resonance in colors and also affect the quality of map information transmission. The distance between two colors in the CIELab color space is expressed by f. c (C) Overall quality of map color usage, c i ' is its habitual color, and γ is the color difference threshold for maintaining the habitual color, described as:

[0075]

[0076] Finally, the values ​​of the three quantitative descriptions of the scientific validity of the thematic map are multiplied together to obtain the overall scientific validity of the thematic map, which is described by F(C):

[0077] F(C) = f v (C)*f r (C)*f c (C)

[0078] Establish a thematic map art aesthetic measurement model:

[0079] Aesthetic experience stems from the interaction between three nervous systems: sensation-motivation, emotion-appraisal, and meaning-perception. In this sense, aesthetics influences a user's perception of a map, their emotional connection to it, and the extent to which they extract information from it. This method establishes a metric model for evaluating the artistic beauty of thematic maps, based on three aspects: color emotion, color preference, and color harmony, derived from map aesthetics and cognitive science.

[0080] First, the aesthetic appeal of a map must be in harmony with its logic and emotion. This method considers three color-related emotional factors: activity, weight, and heat.

[0081] Factor 1: Activity includes active-passive, fresh-old, clean-dirty, and modern-classic.

[0082] Factor 2: Weight includes hard-soft, positive-negative, and heavy-light.

[0083] Factor 3: Heat is labeled as chromothermic and defined by warm-cold.

[0084] The formulas for calculating the three emotion factors based on color appearance attributes such as hue, brightness, and saturation are as follows:

[0085]

[0086] weight=-1.8+0.04*(100-L)+0.45cos(h ab * -100°)

[0087] heat = -0.5 + 0.02 * (ch) ab * ) 1.07 *cos(h ab * -50°)

[0088] All variables follow the definitions in the formula above.

[0089] The sentiment matrix E for a single color consists of three factors, namely...

[0090] E=[activity, weight, heat]

[0091] There is an additive relationship between color combination sentiment and single-color sentiment. Based on this relationship, color combination sentiment can be predicted using the arithmetic mean of single-color sentiment, after normalization:

[0092]

[0093] Where E max =[activity max weight max heat max ]

[0094] Measuring the color sentiment similarity between thematic maps and images f e (C), :

[0095]

[0096] Where E map (c) represents the color emotion of the map, E image (c) represents the color mood of the image.

[0097] In addition, a number of factors influence color preferences, such as tonal differences, relative size, and spacing. Users can choose inspirational references based on their preferences and then use these reference images as input to the model. This method quantifies the similarity between the generated map colors and the reference image colors, and then uses the normalized mean absolute error (MAE) to measure the similarity f. s (C), P map (c i ) represents the percentage of elements in the resulting map, P image (c i () refers to the color proportions in the reference image, described as follows:

[0098]

[0099] Calculate color harmony. The overall color harmony of the map is quantified by introducing a harmony model. The harmony rate in the model ranges from 0 to 5, corresponding to a five-point rating system, and is standardized as follows:

[0100] f h (C) = Harmony(C) / 5

[0101] Based on the above quantification of color emotion, color preference, and color harmony, the total score F for measuring the artistic beauty of the thematic map is obtained by multiplying the scores of these three factors. a (C), as shown below:

[0102] F a (C)=f e (C)*f s (C)*f h (C)

[0103] The scientific and artistic scores of the color schemes obtained in the thematic map style transfer are calculated by the two models mentioned above. Compared with the general thematic map color schemes, this method can better highlight the thematic characteristics, maintain semantic relationships and habitual colors in the process of generating color schemes, and improve the readability of thematic maps.

[0104] A heuristic algorithm is used to optimize and solve the thematic map style transfer model:

[0105] First, the discrete point cloud in the RGB color space is transformed into a discrete point cloud in the CIELab color space through color space conversion, facilitating the measurement of color difference and the search for optimized color schemes. Based on the image classification results from step 1, preliminary style matching between the image and the thematic map is achieved. Then, the thematic map style transfer is organized into a bi-objective, multi-constraint optimization problem, including both scientific (hierarchical layering, encoding semantic relationships, and conventional color rules) and artistic aesthetics (color emotional similarity, color harmony and similarity) objectives. Finally, a heuristic algorithm is used to evaluate each searched color scheme in the CIELab color space to find the global optimal solution to the problem, outputting a series of stylized thematic maps, such as... Figure 3 As shown.

[0106] This invention first divides the image using image entropy and extracts the image color gamut, using it as the source for the color scheme of thematic maps. Then, based on the color rules of thematic maps, it constructs quantitative indices to measure the scientific and artistic beauty of thematic maps by using the visual contrast between layers, semantic relationships between elements, habitual color usage, color sentiment, color harmony, and color preference rules as variables. A heuristic algorithm is then used to explore the optimal solution for both objectives, yielding a series of different style transfer results. This invention provides a style transfer method based on image spatial features and thematic map content, applicable to the automatic style transfer of vector or raster thematic maps of different scales and themes.

Claims

1. A method for automatic thematic map style transfer, characterized in that, The method comprises the following steps: Step 1, image space feature evaluation is carried out by using image entropy, image classification is completed, and an image color gamut is generated, and the color gamut is taken as a source of a thematic map color scheme; Step 2, a scientificity measurement model of the thematic map is established according to main rules related to scientificity of the thematic map, including visual contrast between map layers, semantic relationship between elements and habitual color rules, and a scientificity score of the thematic map is taken as a first optimization target; Step 3, a measurement model reflecting artistic beauty of the thematic map is constructed according to main rules related to artisticity of the thematic map, including color emotion expression, color preference and color harmony of the map, and an artistic beauty score of the thematic map is taken as a second optimization target; Step 4, the automatic migration of the thematic map style is organized into a double-target multi-constraint optimization problem, a heuristic algorithm is used to search a solution set, and a series of thematic map style migration results are obtained.

2. The method for automatic thematic map style migration as claimed in claim 1, wherein, In step 1, image space feature evaluation is carried out by using image entropy, image classification is completed, and an image color gamut is generated, and the color gamut is taken as a source of a thematic map color scheme, and the method comprises the following steps: Step 11, using image entropy to reflect the average information quantity of image source, indicating the spatial characteristics of image gray scale distribution, so as to classify the reference image to adapt to the migration of thematic map with different spatial characteristics. For discrete two-dimensional image, p ij is the spatial frequency of adjacent pixel gray scale, and the description of image entropy is as follows: The distribution complexity of the image is reflected by the H value, and then the complexity of the thematic map elements is considered, and the style migration of different thematic maps is realized through image classification; Step 12, the Hessian matrix is used to detect changes in the reference image and separate areas of abrupt change and gradual change in the image; for the areas of abrupt change, points and lines are separated by the abrupt change direction; for the areas of gradual change, surfaces and backgrounds are separated by visual saliency, after extracting the four types of elements, an appropriate color difference JNCD in CIELab is used as a color distance threshold to extract all colors and calculate the color proportion of each type of color, the reference image is divided into different areas by saliency measurement, which are points, lines, surfaces and background colors, and they are taken as color sources for generating a thematic map color scheme.

3. The method for automatic thematic map style migration as claimed in claim 1, wherein, In step 2, a scientificity measurement model of the thematic map is established according to main rules related to scientificity of the thematic map, including visual contrast between map layers, semantic relationship between elements and habitual color rules, and a scientificity score of the thematic map is taken as a first optimization target, and the method comprises the following steps: Step 21, quantitatively describe the visual contrast between points, lines, surfaces and background layers, which is mainly related to the spatial occupancy and color of symbols. Significant color difference and brightness contrast should be reflected in the color of thematic maps. f v (c i ,c j ) represents the visual contrast between i and j two levels, is the distance between two colors in CIELab color space, L is the brightness, δ e and δ l are the color difference and brightness threshold between two levels, w ij is the weight of the visual level of each element, that is, the spatial average occupancy of two levels. The visual contrast between the layers of the thematic map is described as: Based on the visual contrast between two layers, the average value is calculated to describe the overall layer visual contrast, which is described as: Step 22, Quantification of inter-element semantic relations, the three semantic relations between objects, difference, association and order, are quantified by inter-element color difference and luminance constraint map, r ij is the semantic relation between two colors, μ is the color distance threshold for difference and association, and α is the hue angle threshold for determining the order relation, is the hue difference between two colors, w d , w a and w o are the weights of the three semantic relations, c i is the adjusted color of the i-th element set, c j is the adjusted color of the j-th element set, L is the luminance of the color, is the saturation of the color, and the semantic score is described as: Based on the semantic relationship scores of the above two elements, the overall semantic relationship score of the map is described as: Step 23, the color habit quantification, the color habit rule affects the quality of map information transmission, c i is its color habit, γ is the color difference threshold value for maintaining the color habit, and the overall quality of the map color habit is described as: Step 24, based on the quantification of the above-mentioned visual contrast between layers, semantic relationship between elements and habitual color, the total score measuring the scientificity of the thematic map is obtained by multiplying the scores of the three, and the total score is taken as the first optimization target, which is described as: F(C) = f v (C)*f r (C)*f c (C).

4. The method for automatic topic map style migration as claimed in claim 1 wherein, In step 3, a measurement model reflecting artistic beauty of the thematic map is constructed according to main rules related to artisticity of the thematic map, including color emotion expression, color preference and color harmony of the map, and an artistic beauty score of the thematic map is taken as a second optimization target, and the method comprises the following steps: Step 31, Similarity measure of color emotion, thematic map must have relative logical and emotional fit with map form beauty, through quantitative color emotion factors: activity, weight and heat, to measure the color emotion similarity between thematic map and image, and normalize as follows: wherein E map (c) is a color emotion of the map, E image (c) is a color emotion of the image; Step 32, the measure of color preference, a series of factors will affect the color preference, the user selects the inspiration reference according to his own preference, then the reference image is used as the input of the model, the normalized mean absolute error MAE is used to measure the similarity, P map (c i ) is the proportion of elements of the result map, P image (c i ) is the proportion of colors in the reference image, as follows: Step 33, Quantification of color harmony, introduce harmony model to quantify the overall color harmony degree of map, and standardize as follows: f h (C) = Harmony(C) / 5; Step 34, based on the above-mentioned quantification of color emotion, color preference and color harmony, the total score F measuring the artistic beauty of thematic map is obtained by multiplying the scores of the three a (C) and taking it as the second optimization goal, as follows: F a (C) = f e (C)*f s (C)*f h (C).

5. The method for automatic topic map style migration as claimed in claim 1 wherein, In step 4, the automatic migration of thematic map style is organized into a double target multi-constraint optimization problem, and a heuristic algorithm is used to search the solution set, so as to obtain a series of thematic map style migration results, which include the following steps: Step 41, the constraints set by the first optimization target and the second optimization target established in steps 2 and 3 organize the thematic map into a double target multi-constraint optimization problem, and a heuristic algorithm is used to search for the global optimal solution of the problem; Step 42, considering the importance of the two indexes of scientific nature of thematic map and artistic beauty of thematic map, different styles of thematic map with different tendencies are output according to the demand.

Images