Terrain Artistic Expression Of Ink Style Based On DEM And Style Transfer

The artistic expression of terrain is based on the scientific cognition and abstraction of the geographical environment,using artistic expressions to depict and express the terrain and landforms,not only can convey terrain information,but also spread art and culture.Existing research focuses on the expression and reproduction of the high accuracy of terrain and the fidelity of terrain form,but despise the aesthetic feeling of terrain expression.The artistic expression of terrain is one of the important contents of terrain expression.So far,terrain expression that effectively integrates science and artistry is relatively rare.In particular,the artistic expression of terrain in ink style requires in-depth exploration.Therefore,this research is based on DEM data and deep convolutional neural network as the basic model architecture.It introduces ink painting expression techniques with Chinese national characteristics,constructs a terrain ink style rendering model based on style transfer,and realizes the terrain ink style artistic expression.The main research contents and conclusions of the thesis are as follows:1.Comprehensive analysis of ink art style oriented to topographical expressionIn the expression of topographical ink style,the artistic style of ink is the artistic presentation method of topographical expression.The thesis analyzes the reasons for the differences in the artistic styles of ink painting,classifies ink styles according to painting techniques,and selects the typical painting technique of depicting terrain and ground objects in ink paintings—the style presented in the wrinkle method,as the ink style for rendering the terrain,and based on the deep convolutional neural network realizes the extraction of the texture features of ink art style.In addition,terrain is the carrier of the artistic style of ink,and the abstraction of terrain features is the scientific basis for artistic expression of terrain.Therefore,this research is based on DEM data and constructs a skeleton of terrain features by extracting terrain feature elements,so that the transfer of terrain ink style is not separated real terrain structure.2.Construction and realization of topographical Ink Style Transfer ModelThe performance of the artistic characteristics of ink painting is more subjective and the forms are complex and diverse.Using existing image art style learning algorithms to achieve ink style rendering has certain limitations.Based on the pretrained deep neural network model,the paper designs a deep convolutional neural network for distinguishing and extracting ink style texture features.A sample feature data set containing different ink styles was produced.Using this data set to train the network,combined with the style transfer algorithm idea,a topographic ink style transfer model was constructed based on the deep neural network,which realized the expression of multiple topographic ink styles.Based on the DEM data,taking Lushan Mountain and Taibai Mountain as examples,the ink style terrain expression in different scene ranges and different perspectives was realized.The experimental results show that the model constructed in this paper can achieve the artistic expression of multiple ink styles of terrain,breaking the limitation that the previous methods can only achieve a single style rendering.It is not only suitable for large-scale terrain ink style rendering,but also achieves interactive and multiple perspectives artistic expression of ink style of terrain.3.Analysis and Evaluation of Topographical Ink Style ExpressionThis article evaluates the results of the topographical ink style transfer model from both qualitative and quantitative perspectives.Qualitative evaluation was conducted in the form of a questionnaire survey.The results of the survey showed that: the results of the model generation can take into account the scientific nature of the terrain expression(maintenance of the terrain feature structure)and the artistry(the aesthetic reflection of ink style);the model parameters were determined through the survey results,as it provides a reference for improving the effect of model generation;using the image texture analysis method for quantitative evaluation,the results show that the model can accurately realize the transfer of texture features of various ink styles,but the transfer degree of different texture features is slightly different,from high to bottom.Sorting,followed by obviousness>complexity>similarity>thickness.This research integrates the expression form of painting art into the topographical expression,increases the aesthetics of the topographical expression,and further enriches the content of the artistic expression of the topographical expression;based on the DEM-based topographical ink style expression technology and method,it expands the topographical ink style painting computer drawing method.The research results have positive significance both in terms of enriching the diversity of topographical expressions and in the inheritance of Chinese traditional culture.