A semantic relationship-based energy-saving color matching method and device for electronic maps

Abstract

This invention discloses an energy-saving color matching method and apparatus for electronic maps based on semantic relationships. The method includes: spatializing the colors of the original map; modeling the map colors as discrete point clouds in the CIELab color space according to perceptual uniformity rules; analyzing the discrete point clouds using a clustering algorithm based on color difference rules to extract semantic relationships between map colors, including differences, associations, and orders; establishing a color semantic relationship evaluation model based on the extracted semantic relationships; reorganizing the colors in the RGB color space and establishing an energy consumption estimation model based on the red, green, and blue component values ​​of each color; establishing the map energy-saving problem as a multi-objective, multi-constraint optimization problem; using a heuristic search algorithm to find energy-saving colors; and obtaining the optimal energy-saving color matching scheme by analyzing the Pareto front solution set. This invention is applicable to vector and raster electronic maps of different scales and themes, automatically extracting semantic relationships to reduce map energy consumption.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of geographic information system (GIS) and computer-assisted cartography (CAC), and particularly relates to an electronic map energy-saving color matching method and device based on semantic relationship. BACKGROUND

[0002] As a digital tool, electronic map can provide visualization and spatialization support for sustainable development status, strategy and progress, but its design and operation need to consume a large amount of energy. With the wide application of electronic map in various industries, the power consumption and greenhouse gas emissions will increase rapidly, and the insufficient endurance caused by high energy consumption of electronic map display on mobile devices may also pose a major safety hazard in extreme cases. In recent years, researchers at home and abroad have begun to realize the energy consumption problem of electronic map, and consider energy consumption as a constraint condition for electronic map design, together with form, content, interaction, user experience, etc.

[0003] The current mainstream map energy-saving display technology includes color dimming technology and color adjustment technology. The color dimming technology saves display energy by reducing the backlight intensity, but it does not consider the content and form characteristics of the map, which may cause color distortion, deformation and difficult-to-identify results, thereby affecting the readability of the electronic map. The color adjustment technology replaces high-energy-consuming colors by changing the hue, brightness and saturation of colors to reduce display energy consumption, providing a new method for energy-saving electronic map design. Emerging organic light-emitting diode (OLED) is rapidly replacing traditional liquid crystal display (LCD), as each pixel is self-luminous, the content of the map display can be adjusted at the pixel level to achieve energy-saving effect, so the design of the map content can affect the energy consumption of the map itself. The current main color adjustment method generally takes energy consumption as the objective function and constructs a map energy-saving model with map color setting rules as the constraint condition. As an important map color setting rule, color semantic relationship rule is increasingly applied in map design. When adjusting the color to save energy, the semantic relationship should be maintained to reduce ambiguity in the map energy-saving process. However, there is no automatic extraction of semantic relationship and application in the current adaptive energy-saving color matching method of the map. SUMMARY

[0004] The present application aims to solve the problem of low efficiency of semantic relationship setting in the process of adaptive energy-saving of the map, and provides an electronic map energy-saving color matching method and device based on semantic relationship, which realizes the rapid extraction of map color semantic relationship and adaptive energy-saving display of the electronic map.

[0005] Technical solution: According to the first aspect of the application, a semantic relation-based electronic map energy-saving color matching method comprises the following steps:

[0006] The original map colors are spatialized, and the colors are modeled as a discrete point cloud in the CIELab color space according to the perceived uniformity rule;

[0007] The color points are clustered and analyzed according to the color difference rule, and the semantic relationship between the color points is obtained based on the clustering results, wherein when the distance between two color points is less than the first neighborhood radius, they are divided into the same color cluster, and when the distance between two color points in a color cluster is less than the second neighborhood radius, they are divided into the same color sub-cluster; the relationship between different color clusters is difference; when a color cluster contains only two color points, the color points in the cluster form a correlation relationship; when the saturation distance between the color points in a color sub-cluster is greater than the saturation distance threshold and the brightness distance is greater than the brightness distance threshold, it is determined that the color points in the sub-cluster have an order relationship, otherwise it is determined to have a correlation relationship;

[0008] An evaluation model of the difference relationship, the order relationship and the correlation relationship is established, and the semantic relationship evaluation model of the entire map is obtained by combination;

[0009] The original map colors are reorganized in the RGB color space, and an energy consumption estimation model suitable for different map types is established according to the red, green and blue component values of each color;

[0010] A multi-objective and multi-constrained optimization problem is established according to the semantic relationship evaluation model and the energy consumption estimation model, a heuristic search algorithm is used to find energy-saving alternative colors, and the optimal energy-saving color matching scheme is obtained by analyzing the Pareto frontier solution set.

[0011] Further, the semantic relationship evaluation model is:

[0012]

[0013] Where i and j represent the numbers of any two element classes, c i and c j represent the colors of the i th element class and the j th element class, n represents the number of element classes, r ij represents the semantic relationship between element classes, r ij is D when the element classes represent a difference relationship, r ij is A when the element classes represent a correlation relationship, and r ij is O when the element classes represent an order relationship.

[0014] Wherein, the score f r (c i ,c j ,D) is calculated according to the following formula:

[0015]

[0016] wherein μ denotes a distance threshold value for dividing the difference relation and the association relation, denotes a distance between two colors in the CIELab color space, and θ denotes a hue angle difference threshold value, denotes a hue angle difference between two colors, wherein denote a component a and a component b in the CIELab color space of a color of the i-th element class and the j-th element class, respectively.

[0017] a score f of the association relation r (c i ,c j , A) is calculated according to the following formula:

[0018]

[0019] wherein μ denotes a distance threshold value for dividing the difference relation and the association relation, denotes a distance between two colors in the CIELab color space.

[0020] a score f of the sequential relation r (c i ,c j , O) is calculated according to the following formula:

[0021]

[0022] wherein γ denotes a distance threshold value for dividing the association relation and the sequential relation, α denotes a saturation distance threshold value in the sequential relation, and β denotes a lightness distance threshold value in the sequential relation, denotes a distance between two colors in the CIELab color space, and L i and L j denote lightness of the i-th element class and the j-th element class, and denote saturation of the i-th element class and the j-th element class, wherein a * and b * denote a component a and a component b in the CIELab color space, respectively.

[0023] Further, the energy consumption estimation model is expressed as:

[0024]

[0025] wherein E(C) denotes an overall energy consumption of the map color scheme, n denotes a number of element classes, E iEnergy consumption of the i-th color, W i Proportion of the i-th color, For vector map, p i Area of the i-th feature class, for raster map, p i Pixel number of the i-th feature class, R i Red channel component value, G i Green channel component value, B i Blue channel component value.

[0026] Further, the optimization problem introduces color consistency rules and recognizable rules as constraint conditions, which are expressed as:

[0027]

[0028] In the formula, i and j represent the numbers of any two feature classes, c i and c j Color of the i-th feature class and the j-th feature class, c′ i Adjusted color of the i-th feature class, n represents the number of feature classes, Color distance in CIELab color space, ε represents color recognizable distance threshold, and ∈ represents color distance threshold before and after adjustment;

[0029] The optimization problem reorganizes and searches colors in the HSV color space in the solving process, and obtains a Pareto front solution set through multiple iterations, as a series of energy-saving color matching schemes, wherein the Pareto front is a continuous curved surface, and there is one or more inflection points on the curved surface, that is, the first derivative of the curved surface function suddenly changes, and the first inflection point is used as a satisfactory solution of the energy-saving map color matching scheme; when the inflection point is not clear, the center point of the Pareto front surface is used as a satisfactory solution, that is, the solution that takes into account the energy-saving degree and the semantic relationship.

[0030] According to a second aspect of the present application, an electronic map energy-saving color matching device based on semantic relationship comprises:

[0031] A color space processing module is configured to perform spatialization processing on original map colors, and model the colors as discrete point clouds in CIELab color space according to the perceptual uniformity rule;

[0032] The semantic relation extraction module is configured to perform clustering analysis on the color points according to a color difference rule, and obtain semantic relations between the color points based on a clustering result, wherein two color points are divided into a same color cluster when a distance between the two color points is less than a first neighborhood radius, and two color points in a color cluster are divided into a same color sub-cluster when a distance between the two color points is less than a second neighborhood radius; the color clusters have a difference relation; when a color cluster contains only two color points, the color points in the cluster have an association relation; when the color points in a color sub-cluster have a saturation distance greater than a saturation distance threshold and a brightness distance greater than a brightness distance threshold, the color points in the color sub-cluster have an order relation, otherwise, the color points have an association relation;

[0033] The semantic relation evaluation model construction module is configured to establish evaluation models of the difference relation, the order relation and the association relation, and combine the evaluation models to obtain a semantic relation evaluation model of the entire map.

[0034] The energy consumption estimation model construction module is configured to reorganize colors of the original map in an RGB color space, and establish an energy consumption estimation model suitable for different map types according to red, green and blue component values of the colors.

[0035] The optimization problem construction and solution module is configured to establish a multi-objective and multi-constraint optimization problem according to the semantic relation evaluation model and the energy consumption estimation model, find an energy-saving color by using a heuristic search algorithm, and obtain an optimal energy-saving color matching scheme by analyzing a Pareto front solution set.

[0036] According to a third aspect of the present application, a computer device is provided, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs, when executed by the processors, implement steps of the semantic relation-based electronic map energy-saving color matching method according to the first aspect of the present application.

[0037] According to a fourth aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the computer program, when executed by a processor, implements steps of the semantic relation-based electronic map energy-saving color matching method according to the first aspect of the present application.

[0038] Beneficial effects: the application models the color of the original map as a discrete point cloud, automatically extracts the semantic relationship between the color points by using a clustering algorithm, maintains the semantic relationship between the map colors in the process of adjusting the color matching to reduce energy consumption, compared with the color dimming method, the method can avoid the loss of color semantic relationship, improve the readability of the map; compared with the color adjustment method, the method introduces multiple map coloring rules, improves the effect of map energy-saving display. Compared with the existing color adjustment method considering semantic relationship, the application realizes automatic extraction of semantic relationship, further improves the adaptive energy-saving efficiency of the map. The application optimizes the map energy-saving process, improves the energy-saving efficiency, and is suitable for different scales, different thematic vector or raster electronic maps. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a flowchart of the method of the application;

[0040] Figure 2 is a semantic relationship extraction result graph, wherein (a) is a clustering result, and (b) is semantic extraction;

[0041] Figure 3 is a semantic relationship example graph, wherein (a) is an original map, (b) is a semantic relationship, and (c) is a semantic relationship matrix;

[0042] Figure 4 is an optimal energy-saving map example. DETAILED DESCRIPTION

[0043] The technical solutions of the application will be further described below with reference to the drawings.

[0044] Color, as an important map visual language, not only affects the information transmission function of the map, but also affects the design of the map content and further affects the overall energy consumption of the electronic map. In map making and visualization, color is used to encode the semantic relationship between map features, including difference, order and association, forming a kind of cohesive aesthetic style. In an electronic map, the difference relationship represents different thematic and type of ground feature elements, such as vegetation, water body, building, etc., which are described by colors with large color difference, and the colors have obvious contrast; the association relationship represents the same thematic and different feature types of ground feature elements, such as rivers, lakes and oceans, which are described by colors with small color difference; the order relationship represents ground feature elements with certain level or size, such as different road levels, etc., which are described by level type color matching scheme according to the numerical value of the element class, and the brightness and saturation of the color change correspond to the change of the element data value.

[0045] The application discloses a kind of energy-saving color matching method of electronic map based on semantic relation.First, the color of original map is spatialized, and according to the rule of perceived uniformity, the color of map is modeled as discrete point cloud in CIELab color space, according to the rule of color difference, the distance between colors is measured by Euclidean distance, and the semantic relation between colors is automatically extracted, including difference, association and order, by clustering algorithm analysis of discrete point cloud.Then, a color semantic relation evaluation model is established according to the extracted semantic relation; colors are reorganized in RGB color space, and an energy consumption estimation model suitable for different map types is established according to the red, green and blue component values of each color;Finally, the map energy-saving problem is established as a multi-objective, multi-constrained optimization problem by introducing recognizable rules and color consistency rules, and a heuristic search algorithm is used to find energy-saving colors, and the optimal energy-saving color matching scheme is obtained by analyzing the Pareto frontier solution set.The application is suitable for different scales, different thematic vector and raster electronic maps, and can realize automatic extraction of semantic relation and reduction of map energy consumption.

[0046] Reference Figure 1 The method of the application specifically comprises the following steps:

[0047] Step 1, spatialize the color of original map, and model the color as discrete point cloud in CIELab color space according to the rule of perceived uniformity.

[0048] Commonly used color spaces include RGB, HSV and CIELab color spaces.RGB color space uses additive color method to represent any color, uses a three-dimensional coordinate system to represent three orthogonal components in color space, and the coordinates represent red, green and blue channel values, respectively, with each channel value ranging from 0 to 255.The color coordinates are related to the device, but do not conform to the dimensions of human color perception.HSV color space describes color from three dimensions: hue H, lightness V and saturation S, which can provide an approximate value of color perception dimension and facilitate color search within the color space range, but the three components are not independent.CIELab color space is a uniform color space obtained by nonlinear transformation of CIEXYZ, which is more consistent with human visual perception, and is widely used to measure color difference.L represents lightness (0-100);a * represents from green (-) to red (+), b * represents from blue (-) to yellow (+), a * , b * range is usually -128 to +127.Therefore, RGB color space is convenient for measuring color energy consumption, HSV color space is convenient for searching energy-saving colors within the color space range, and CIELab color space is convenient for measuring color difference.

[0049] Considering the perceptual uniformity rule, the original map color data is spatialized in CIELab color space, organized and described as a three-element As Figure 2 To achieve the spatialization of map color data, as shown in (a), the a value and b value of the color are taken as the X and Y axes in the Cartesian coordinate system, and the color is mapped to the three-dimensional Cartesian coordinate system, modeled as a discrete point cloud in the CIELab color space, where the a value and b value reflect the hue and saturation of the color of each element of the map, and the L value reflects the brightness of the color of each element of the map. When all the colors of the map are modeled in CIELab color space to form a discrete point cloud, the color of each element class corresponds to a point in the discrete point cloud. Thus, the differences, orders and association relationships in the semantic relationship can be extracted according to the distances between the color points. In this paper, an element class represents an element in the map, such as forest, river, etc. in the urban land use map, which respectively represent an element class. Each element class has a color, corresponding to a color point.

[0050] Step 2, according to the color difference rule, the color points are clustered and analyzed, and the semantic relationship between the color points is obtained based on the clustering results.

[0051] The color of the map needs to consider the color difference rule, and with the improvement of the color setting requirements of the map, the color difference formula is also developing, generally including three methods based on MacAdam ellipse, based on Munsell new standard, and linear transformation according to CIE colorimetric system. The CIE1976L*a*b*color difference formula published in 1976 uses Euclidean distance to represent the distance between colors Can be described as:

[0052]

[0053] Where i and j represent the i-th and j-th element classes in the map, c i and c j represent the color of the i-th element class and the j-th element class, L i , L j , respectively represent the brightness value, component a and component b of the i-th element class and the j-th element class color in CIELab color space.

[0054] Based on the distance between colors, the application uses a clustering algorithm to analyze the discrete point cloud, automatically extracts the semantic relationship between the colors of the map, including difference, association and order. Among them, when the distance between two color points is less than the first neighborhood radius, they are divided into the same color cluster, and when the distance between two color points in a color cluster is less than the second neighborhood radius, they are divided into the same color sub-cluster; the difference between each color cluster is the difference; when a color cluster contains only two color points, the color points form an association relationship; when the saturation distance between each color point in a color sub-cluster is greater than the saturation distance threshold and the brightness distance is greater than the brightness distance threshold, it is determined that the color points in the sub-cluster have an order relationship, otherwise it is determined to be an association relationship.

[0055] Specifically, clustering analysis is a main method of data mining, which can group different objects and divide different groups according to the similarity between objects. In order to mine the potential map color knowledge rules in the map data, a clustering algorithm is used to analyze the original map colors, which can divide different element colors according to the set conditions, so as to extract the color semantic relationship between map objects. The difference relationship is the most basic semantic relationship, according to which the colors can be divided into different groups, and the association relationship and the order relationship are further analysis of different color groups. Therefore, first, the map colors are clustered to determine the difference relationship between different element classes, a clustering algorithm is used to analyze the discrete point cloud, and μ represents the neighborhood radius for determining the difference relationship between colors. When the distance between color c i and color c j is less than μ, it means that the two colors are divided into the same color cluster, otherwise they are divided into different color clusters. As shown in (a) of FIG. 1, all the discrete point clouds in the CIELab color space are clustered and analyzed, and the color points are divided into 6 color clusters. As shown in (a) of FIG. 2, the difference relationship is formed between the colors in each cluster, and if a color cluster contains only two color points, the color points in the cluster form an association relationship. Figure 2 Figure 2 As shown in (a) of FIG. 2, the difference relationship is formed between the colors in each cluster, and if a color cluster contains only two color points, the color points in the cluster form an association relationship.

[0056] Since the order relationship colors present gradient changes in saturation and brightness, and the association relationship does not have regular and uniform gradient changes. Therefore, the application further clusters and analyzes the clusters containing multiple color points in the CIELab color space, and γ represents the distance threshold for dividing the association relationship and the order relationship in the same color cluster. When the distance between two colors is less than γ, they are divided into the same sub-cluster. The colors in the sub-cluster are sorted according to the saturation and brightness, and α represents the saturation distance threshold for dividing the association relationship and the order relationship, and β represents the brightness distance threshold for dividing the association relationship and the order relationship. When the saturation distance between each color in the sub-cluster is greater than α and the brightness distance is greater than β, it is determined that the colors in the sub-cluster have an order relationship, otherwise it is determined to be an association relationship. As shown in (a) of FIG. 3, the colors in the sub-cluster are further clustered and analyzed, and the colors in the sub-cluster are divided into two sub-clusters, and the colors in each sub-cluster form an order relationship. Figure 3 ​The D group color points form a sequential relationship through further clustering analysis, and the E group color points form a two-group correlation relationship through further clustering analysis.

[0057] According to the above method, the colors of each feature class of the map can be divided into different color clusters, and the differences, sequential relationships and correlation relationships between the colors are formed. For example, Figure 3 The original map in (a) is a city land use map, and the following semantic relationships can be obtained Figure 3 The semantic relationship in (b), and Figure 3 The semantic relationship matrix shown in (c). In the process of adjusting the color to reduce energy consumption, the semantic relationship between the colors needs to be maintained to improve the readability of the energy-saving map and the quality of information transmission. Compared with the general semantic relationship setting method, this method automatically extracts the semantic relationship, the result is more accurate, optimizes the energy-saving process, and improves the energy-saving efficiency.

[0058] Step 3, establish an evaluation model of the difference relationship, the sequential relationship and the correlation relationship, and combine them to obtain a semantic relationship evaluation model of the whole map.

[0059] In the process of map energy saving, in addition to considering the energy saving effect, the maintenance of color semantic relationship is also needed to quantify the map quality and avoid color distortion caused by color gamut compression to black during the energy saving process. In addition, the color semantic relationship not only conveys the intention of the cartographer, but also contains the understanding of the map reader. In the process of making an energy-saving map, the semantic relationship between the colors of the original map should be maintained as much as possible to convey the same map information before and after energy saving. In order to more clearly represent the semantic relationship and group division between the feature classes of the map, as shown in Figure 4 (c) is described by using a semantic relationship matrix. In the matrix, D represents that the colors of two feature classes are determined to be a difference relationship, A represents that the colors of two feature classes are determined to be a correlation relationship, and O represents that the colors of two feature classes are determined to be a sequential relationship.

[0060] For the colors determined to be a difference relationship D, a certain contrast needs to be maintained during the energy saving process, the distance between the two colors should always be greater than the neighborhood radius μ for dividing the difference relationship and the correlation relationship, the hue angle difference between the two colors should always be greater than the hue angle threshold θ, and the score f r (c i ,c j ,D) can be represented as:

[0061]

[0062] In the formula, i and j represent the numbers of any two feature classes, c i and c jci and c denote the color of the i-th element class and the j-th element class, μ denotes the distance threshold value for dividing the difference relation and the association relation, denotes the distance between two colors in the CIELab color space, θ denotes the hue angle difference threshold value, denotes the hue angle difference between two colors. Use denote the component a and the component b of the i-th element class and the j-th element class color in the CIELab color space, respectively, and the hue angle difference can be expressed as:

[0063]

[0064] For the colors determined to be the association relation, the degree of proximity between the colors needs to be maintained in the energy saving process, the distance between the two colors should always be less than the neighborhood radius μ for dividing the difference relation and the association relation, and the score f r (c i ,c j ,A) can be expressed as:

[0065]

[0066] In the formula, i and j denote the numbers of any two element classes, c i and c j denote the color of the i-th element class and the j-th element class, μ denotes the distance threshold value for dividing the difference relation and the association relation, denotes the distance between two colors in the CIELab color space.

[0067] For the colors determined to be the sequential relation, the saturation and the brightness thereof need to maintain a certain gradient in the energy saving process, the saturation difference between the colors is greater than α, the brightness difference is greater than β, and the color difference between the colors is less than the color distance threshold value γ for dividing the association relation and the sequential relation, and the score f r (c i ,c j ,O) can be expressed as:

[0068]

[0069] In the formula, i and j denote the numbers of any two element classes, c i and c j denote the color of the i-th element class and the j-th element class, γ denotes the distance threshold value for dividing the association relation and the sequential relation, α denotes the saturation distance threshold value in the sequential relation, and β denotes the brightness distance threshold value in the sequential relation, denotes the distance between two colors in the CIELab color space, L i and L jLij represents the lightness of the ith element class and the jth element class, and Sij represents the saturation of the ith element class and the jth element class, a * and b * respectively represent the component a and the component b in the CIELab color space, saturation can be described as:

[0070]

[0071] The semantic relationship between the colors of the element classes should be maintained in the process of adjusting the colors to reduce the energy consumption, and the semantic relationship score is embodied through the semantic relationship. According to the semantic relationship between the ith element class and the jth element class, a corresponding method is selected to score the semantic relationship quality in the energy-saving process, and the result value range is 0 to 1. The evaluation model F(C) of the semantic relationship can be described as:

[0072]

[0073] In the formula, n represents the number of element classes, r ij represents the semantic relationship between the element classes. When the element classes are embodied as a difference relationship, r ij is D; when the element classes are embodied as an association relationship, r ij is A; and when the element classes are embodied as an order relationship, r ij is O.

[0074] The semantic relationship between the original color matching scheme of the map and all the color matching schemes searched in the energy-saving process can be calculated by the above evaluation model. Compared with general color adjustment methods, the method considers the distance threshold values of the division difference, the order and the association relationship, and the saturation distance and the lightness distance threshold values in the order relationship in the search process, can better quantitatively quantify the semantic relationship between the colors of the map, and improve the information transmission quality of the map.

[0075] Step 4: Reorganize the original map colors in the RGB color space, and establish an energy consumption estimation model suitable for different map types according to the red, green and blue component values of each color.

[0076] The RGB color space uses different degrees of superposition of R (Red), G (Green) and B (Blue) components to produce various colors, and is a device-independent color space. The colors of the original map are organized and described as a triple <R i ,G i ,B i > according to the RGB color space, and the map color matching scheme is described as a color set Si={s1,s2,…,si}.

[0077] To ensure the color adjustment process adapts to the characteristics of different maps, the color proportion of each feature class is extracted in different ways. For a vector map, the color proportion is obtained by calculating the area ratio of each feature class to the whole map; for a raster map, the color proportion is obtained by calculating the pixel number ratio of each feature class to the whole map. The color proportion of the i-th feature class is denoted as W i It can be described as:

[0078]

[0079] For a vector map, p i represents the area of the i-th feature class, and for a raster map, p i represents the pixel number of the i-th feature class, and n represents the number of feature classes in the electronic map.

[0080] OLED display screen can adjust the content of map display at pixel level to achieve energy saving effect because each pixel is self-luminous. The energy consumption of OLED display screen can be calculated according to the red, green and blue channel values of each color. Let R i represent the red channel component value, G i represent the green channel component value, and B i represent the blue channel component value. Combined with the color proportion of each feature class, the energy consumption of the whole electronic map can be estimated. Let E(C) represent the overall energy consumption of a certain color scheme of the map, n represent the number of feature classes, E i represent the energy consumption of the i-th color, W i represent the proportion of the i-th color. The map energy consumption model can be described as:

[0081]

[0082] The energy consumption of the original color scheme of the map and the color scheme searched in all energy-saving processes can be calculated by the above energy consumption model. Compared with general color adjustment methods, this method takes energy consumption into color adjustment, and can adapt to the characteristics of different electronic maps, improving the accuracy of energy consumption estimation in the energy-saving process.

[0083] Step 5, according to the semantic relationship evaluation model and the energy consumption estimation model, a multi-objective and multi-constrained optimization problem is established, a heuristic search algorithm is used to find the energy-saving color, and the optimal energy-saving color scheme is obtained by analyzing the Pareto frontier solution set.

[0084] Adaptive energy saving of electronic map needs to consider both map energy consumption and semantic relationship of map, according to the established map energy consumption estimation model and semantic relationship evaluation model. Since the map will compress the color gamut of the original map color to black in the energy saving process, which destroys the semantic relationship between the original map colors, in most cases the energy saving of the map and the semantic relationship cannot be satisfied at the same time. In order to solve the above problems, the map energy saving problem can be defined as a multi-objective and multi-constraint optimization problem, and the map energy consumption and the semantic relationship of map color are quantitatively analyzed and the objective function is constructed, which can obtain the map energy saving model considering both map energy consumption and color semantic relationship:

[0085]

[0086] In order to facilitate global search, the energy consumption result is converted into the adjusted map energy reduction percentage, and the map energy reduction is constructed as a MAX-MAX problem, E n represents the energy consumption of the adjusted map, E o represents the energy consumption of the map before adjustment:

[0087]

[0088] The visual effect of map design will affect the user's map cognitive result, and the color design involves many color setting rules. Taking the color setting rules as the constraint conditions of optimization variables can control the color change within a certain range and improve the design quality of the map. In the map energy saving model, in order to ensure that the energy consumption is reduced while the adjusted color is relatively close to the original color, the distance between any two colors should be greater than the threshold of perceptible color difference (JNCD) to maintain the distinguishable rules between feature classes, and the color consistency rule and distinguishable rule are introduced as constraint conditions, wherein i and j represent the numbers of any two feature classes, c i and c j represent the colors of the i th feature class and the j th feature class, c′ i represents the adjusted color of the i th feature class, n represents the number of feature classes, represents the distance of color in CIELab color space, ε represents the color distinguishable distance threshold, ∈ represents the distance threshold of color before and after adjustment, the color consistency rule and distinguishable rule can be described as:

[0089]

[0090] In the color adjustment process, all results are constrained by the above two conditions. Compared with the general color dimming and color adjustment method, this method controls the color change within a certain range and maintains the distinguishability between colors, avoids color distortion, and improves the readability of the map.

[0091] According to the established double-target energy-saving model and constraint conditions related to map coloring rules, the map energy-saving problem is organized into a MAX-MAX double-target multi-constraint optimization problem, and a heuristic search is used to search for energy-saving color matching schemes within a certain range. Since the HSV color space describes colors through hue, brightness, and saturation, it can provide approximate values of color perception dimensions, facilitating color search within the color space range and reducing computational cost. Therefore, colors are reorganized and searched in the HSV color space during the energy-saving process, and a Pareto front solution set, i.e., a series of energy-saving color matching schemes, is obtained through multiple iterations. Assuming that the Pareto front surface is a continuous surface and there is one or more inflection points on the surface, i.e., the first derivative of the surface function changes abruptly, the first inflection point is used as a satisfactory solution of the energy-saving map color matching scheme. When the inflection point is not clear, the center point of the Pareto front surface, i.e., a solution that takes into account the energy-saving degree and semantic relationship, is used as a satisfactory solution, as shown in Figure 1 which can obtain a map color matching scheme that maximally reduces energy consumption with slight changes to the appearance of the map.

[0092] Based on the same technical concept as the method embodiment, the application also provides an electronic map energy-saving color matching device based on semantic relationships, comprising:

[0093] A color space processing module is configured to perform spatialization processing on original map colors, and model the colors as a discrete point cloud in a CIELab color space according to a perceptual uniformity rule;

[0094] A semantic relationship extraction module is configured to perform clustering analysis on the color points according to a color difference rule, and obtain semantic relationships between the color points based on the clustering results. When the distance between two color points is less than a first neighborhood radius, they are divided into the same color cluster. When the distance between two color points in a color cluster is less than a second neighborhood radius, they are divided into the same color sub-cluster. The color clusters are in a difference relationship. When a color cluster contains only two color points, the color points in the cluster form a correlation relationship. When the saturation distance between the color points in a color sub-cluster is greater than a saturation distance threshold and the brightness distance is greater than a brightness distance threshold, the color points in the sub-cluster are in a sequential relationship, otherwise they are in a correlation relationship.

[0095] A semantic relationship evaluation model construction module is configured to establish evaluation models of the difference relationship, the sequential relationship, and the correlation relationship, and combine them to obtain a semantic relationship evaluation model of the entire map;

[0096] An energy consumption estimation model construction module is configured to reorganize the original map colors in an RGB color space, and establish an energy consumption estimation model that is suitable for different map types according to the red, green, and blue component values of the colors;

[0097] The optimization problem construction and solution module is used for establishing a multi-objective and multi-constraint optimization problem according to the semantic relation evaluation model and the energy consumption estimation model, searching for energy-saving colors by using a heuristic search algorithm, and obtaining an optimal energy-saving color matching scheme by analyzing a Pareto frontier solution set.

[0098] It should be understood that the electronic map energy-saving color matching device based on semantic relations in the embodiments of the present application can realize all the technical solutions in the method embodiments described above, and the functions of each functional module can be realized according to the methods in the method embodiments described above, and the specific implementation process can be referred to the related description in the above embodiments, which will not be described here in detail.

[0099] The present application also provides a computer device, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the program is executed by the processor to realize the steps of the electronic map energy-saving color matching method based on semantic relations as described above.

[0100] The present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by the processor to realize the steps of the electronic map energy-saving color matching method based on semantic relations as described above.

[0101] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0102] The present application is described with reference to flowcharts according to the method, device (system), and computer program product of the embodiments of the present application. It should be understood that each flow in the flowcharts and the combination of the flows in the flowcharts can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices produce a means for implementing the functions specified in the flowcharts or multiple flows. Figure 1 The means for implementing the functions specified in the flowcharts or multiple flows.

[0103] These computer program instructions can also be stored in a computer-readable memory capable of guiding the computer or other programmable data processing devices to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction means, which implements the functions specified in the flowFigure 1 the function specified in the flow or flows.

[0104] These computer program instructions can also be loaded into a computer or other programmable data processing devices, so that a series of operational steps are performed on the computer or other programmable data processing devices to generate a computer-implemented process, thus the instructions executed on the computer or other programmable data processing devices provide a process for implementing the function specified in the flow or flows. ​ the function specified in the flow or flows.

Claims

1. A method for energy-saving color matching of an electronic map based on semantic relations, characterized in that, The method comprises the following steps: spatializing the original map colors, and modeling the colors as a discrete point cloud in a CIELab color space according to a perceived uniformity rule; performing clustering analysis on the color points according to a color difference rule, and obtaining semantic relationships between the color points based on the clustering results, wherein two color points are divided into the same color cluster when the distance between the two color points is less than a first neighborhood radius, and two color points in a color cluster are divided into the same color sub-cluster when the distance between the two color points is less than a second neighborhood radius; the semantic relationship between color clusters is a difference relationship; when a color cluster contains only two color points, the color points in the cluster form an association relationship; when the saturation distance between the color points in a color sub-cluster is greater than a saturation distance threshold and the lightness distance between the color points is greater than a lightness distance threshold, the color points in the sub-cluster are determined to have a sequential relationship, otherwise, the color points are determined to have an association relationship; establishing evaluation models of the difference relationship, the sequential relationship and the association relationship, and combining the evaluation models to obtain a semantic relationship evaluation model of the entire map; reorganizing the original map colors in an RGB color space, and establishing an energy consumption estimation model that is adaptive to different map types according to red, green and blue component values of the colors; establishing a multi-objective and multi-constraint optimization problem according to the semantic relationship evaluation model and the energy consumption estimation model, searching for energy-saving alternative colors by using a heuristic search algorithm, and obtaining an optimal energy-saving color matching scheme by analyzing a Pareto frontier solution set; the optimization problem introduces a color consistency rule and a distinguishability rule as constraint conditions, and the color consistency rule and the distinguishability rule are represented as: where i and j represent the index of any two element classes, c i and c j represent the color of the ith element class and the jth element class, c′ i represent the adjusted color of the ith element class, n represents the number of element classes, represents the distance of colors in CIELab color space, ε represents the color distinguishable distance threshold, and ∈ represents the color distance threshold before and after adjustment. the optimization problem reorganizes the colors in an HSV color space and searches for the colors in the solving process, and a Pareto frontier solution set is obtained through multiple iterations, which is a series of energy-saving color matching schemes; the Pareto frontier surface is a continuous curved surface, and there is one or more inflection points on the curved surface, that is, the first derivative of the curved surface function suddenly changes; the first inflection point is used as a satisfactory solution of the energy-saving map color matching scheme; when the inflection point is not clear, the center point of the Pareto frontier surface is used as a satisfactory solution that takes into account the energy-saving degree and the semantic relationship.

2. The method of claim 1, wherein, The semantic relationship evaluation model is represented as: where i and j represent the number of any two element classes, c i and c j represent the color of the ith element class and the jth element class, n represents the number of element classes, r ij represents the semantic relationship between each element class, when the relationship between element classes is embodied as a difference relationship, r ij is D; when the relationship between element classes is embodied as an association relationship, r ij is A; when the relationship between element classes is embodied as a sequential relationship, r ij is O.

3. The method of claim 2, wherein, Score f of the differential relationship r (c i ,c j D) is calculated according to the following formula: wherein μ denotes a distance threshold value for dividing the difference relation and the association relation, denotes a distance between two colors in the CIELab color space, and θ denotes a hue angle difference threshold value, denotes a hue angle difference between two colors, wherein denote components a and b in the CIELab color space of the color of the i-th element class and the j-th element class, respectively.

4. The method of claim 2, wherein, Score f of the association r (c i ,c j A) is calculated according to the following formula: wherein μ denotes a distance threshold value for dividing the difference relation and the association relation, denotes the distance of two colors in the CIELab color space.

5. The method of claim 2, wherein, Score f of the sequential relationship r (c i ,c j O) is calculated according to the formula: where γ represents a distance threshold value that divides the association and the sequential relationship, α represents a saturation distance threshold value in the sequential relationship, and β represents a brightness distance threshold value in the sequential relationship, represents a distance between two colors in a CIELab color space, L i and L j represents brightness of an i-th element class and a j-th element class, and represents saturation of an i-th element class and a j-th element class, where a * and b * respectively represent a component a and a component b in the CIELab color space.

6. The method of claim 1, wherein, The energy consumption estimation model is represented as: where E(C) represents the overall energy consumption of the map color scheme, n represents the number of feature classes, E i represents the energy consumption of the i-th color, W i represents the proportion of the i-th color, For a vector map, p i represents the area of the i-th feature class, for a raster map, p i represents the number of pixels of the i-th feature class, R i represents the red channel component value, G i represents the green channel component value, B i represents the blue channel component value.

7. A device for energy-saving color matching of an electronic map based on semantic relations, characterized in that The method comprises the following steps: a color space processing module, configured to spatialize the original map colors, and model the colors as a discrete point cloud in a CIELab color space according to a perceived uniformity rule; a semantic relationship extraction module, configured to perform clustering analysis on the color points according to a color difference rule, and obtain semantic relationships between the color points based on the clustering results, wherein two color points are divided into the same color cluster when the distance between the two color points is less than a first neighborhood radius, and two color points in a color cluster are divided into the same color sub-cluster when the distance between the two color points is less than a second neighborhood radius; the semantic relationship between color clusters is a difference relationship; when a color cluster contains only two color points, the color points in the cluster form an association relationship; when the saturation distance between the color points in a color sub-cluster is greater than a saturation distance threshold and the lightness distance between the color points is greater than a lightness distance threshold, the color points in the sub-cluster are determined to have a sequential relationship, otherwise, the color points are determined to have an association relationship; The semantic relation evaluation model construction module is configured to establish evaluation models of difference relations, sequence relations and association relations, and to combine the evaluation models to obtain a semantic relation evaluation model of the entire map; The energy consumption estimation model construction module reorganizes the colors of the original map in the RGB color space, and establishes an energy consumption estimation model that is suitable for different map types according to the red, green and blue component values of each color; The optimization problem construction and solution module is configured to establish a multi-objective and multi-constraint optimization problem according to the semantic relation evaluation model and the energy consumption estimation model, to search for energy-saving colors by using a heuristic search algorithm, and to obtain an optimal energy-saving color matching scheme by analyzing a Pareto frontier solution set. The optimization problem introduces color consistency rules and distinguishable rules as constraint conditions, and the color consistency rules and distinguishable rules are expressed as: where i and j represent the index of any two element classes, c i and c j represent the color of the ith element class and the jth element class, c′ i represent the adjusted color of the ith element class, n represents the number of element classes, represents the distance of colors in CIELab color space, ε represents the color distinguishable distance threshold, and ∈ represents the color distance threshold before and after adjustment. The optimization problem reorganizes colors in the HSV color space and searches for colors in the HSV color space during a solution process, and obtains a Pareto frontier solution set as a series of energy-saving color matching schemes through multiple iterations. The Pareto frontier surface is a continuous curved surface, and there is one or more inflection points, i.e., first derivative mutation points of the curved surface function, on the curved surface. The first inflection point is used as a satisfactory solution of the energy-saving map color matching scheme. When the inflection point is not clear, the center point of the Pareto frontier surface, i.e., a solution that takes into account the energy-saving degree and the semantic relation, is used as a satisfactory solution.

8. A computer device, comprising: comprise: one or more processors; memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs, when executed by the processors, implement the steps of the semantic relation-based electronic map energy-saving color matching method according to any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer programs, when executed by the processors, implement the steps of the semantic relation-based electronic map energy-saving color matching method according to any one of claims 1-6.

Images