Research On Automatic Color Design Method Of Thematic Map Based On Samples Learning

With the continuous development and innovation of geographic information tech-nology and computer technology,it is more convenient for people to obtain geographic data.Whether experienced cartographic experts or common users all can participate in map mapping,the era of public cartography is coming.Color as the graphic language of map and the important influencing factors in the process of mapping,can not only express qualitative and quantitative geographic information,but also reflect the artistic and aesthetic value of the map.Effective map color is benefit to improve the visual effect of the map,the information transmission capacity and content load of map are also improved.On the contrary,the invalid map color not only affects the map's view ability but also conveys the incorrect geographical information.Therefore,color design directly affects the readability and usability of the map.At present,a series of research on map color design are carried out by many map makers and considerable research results also obtained.But for the novice and common users who lack the mapping ex-perience,and the with influence of difficulty and ambiguity in the process of map mak-ing,some problems such as low automation and low efficiency are still existed.Many principles need to follow in map color design,such as the custom color principle,the identifiable principle,etc.These principles are all qualitatively described which lack of quantitative research.Secondly,the current map color design focuses on single-type(qualitative,sequential,diverging)color schemes generation,and mixed(two or more single-type color scheme combinations)color scheme studies on auto-matic generation of color schemes are lacked.Therefore,the automation method of map color design which based on map sam-ple learning is conducted in this paper.The main research contents and achievements are as follow:(1)With the respect to the current color quantitative researches are payed little attention on human satisfaction,rules of map color design like custom color are intro-duced,and some existing quantitative research of color are analyzed,such as histogram,kernel density estimation(KDE),etc.However,color probability values calculated by these methods(objectives)and the satisfaction of the colors selected by the reader(sub-jective)are lacking correlation.Therefore,based on the collected map samples,a method of linking kernel density estimation with human satisfaction is proposed to quantify color knowledge and guide map color design.(2)With the respect to the current color scheme research mainly focus on single color schemes,while the problems of mixed color schemes studies are lacked.After analyzing single-type and mixed color schemes,this paper combined with the current map color design which has low automation degree and low efficiency problems,the artificial bee colony algorithm(ABC)is proposed to solve the above problems.The features of ABC algorithm and the rules of mixed color schemes are combined for ini-tialization,fitness definition and local and global search to build the mixed color scheme model.The parameters of ABC algorithm are adjusted and defined in combi-nation with different color matching requirements,and the automation color designing of mixed color schemes are realized.(3)Experimental verification and usability analysis are realized.The evaluation of the generated color schemes was conducted by means of questionnaires,the efficiency,accuracy and satisfaction evaluation indicators were used.And after that color schemes which produced by our method are compared with the color schemes of authoritative experts or institutions.Based on the same reading task,25 readers performed a map reading experiment on three sets of comparison images.The results show that in terms of efficiency and correctness,the color scheme generated by the proposed method is generally not lower than the contrast color scheme,and the accuracy and efficiency of some contrast experiments are higher than the contrast color scheme.In terms of satis-faction,the color schemes participating in the comparison all achieved higher satisfac-tion;through the significant analysis,the satisfaction of the three groups of experiments was not significantly different,so the color scheme proposed by the proposed method is generally not worse than the comparison.The color scheme is even better than the contrast color scheme.