DEM起伏纹理的视觉认知内涵与多变量消隐方法

地形是在内外营力的共同作用之下综合形成的，不同类型的地形具有不同的起伏形态特征，从全局或宏观的观测视角，可以视觉感知到这种地形起伏及其特征规律。纹理是众多物体表面呈现的有规律的线条组合，DEM作为地形的数据表达，不同地形类型区域的起伏形态具有纹理的显著表征。目前，对于DEM纹理研究主要是从DEM影像认知的角度出发，利用图像纹理的计算方法进行地形纹理对象描述和指标的提取，并基于获取的纹理对象进行区域地形特征分析和地貌类型的分类，在视觉上的反映存在不足。DEM起伏纹理作为一种基于某种观测视角和尺度的地形要素组合对象，首先需要对其进行系统的视觉认知，明确基于视觉认知的DEM起伏纹理构成要素和不同视觉变量控制的纹理要素组合结构，通过实现DEM起伏纹理的多维可视化建模，达到DEM数字地形建模的抽象与简化。基于此，本文以DEM地形起伏纹理为研究对象，对DEM起伏纹理进行视觉认知，明确DEM起伏纹理的涵义并建立提取方法，使用建立的方法实现了DEM起伏纹理的提取，取得了较好的提取效果。从视觉感知的机制出发，确定了决定视觉感知结果的视觉变量组合。确定了包含景深、长度和邻域等用于视觉综合的参数，设计建立了多变量的DEM起伏纹理消隐方法。最后设置视觉变量并使用该方法实现了DEM起伏纹理的消隐，消隐后DEM起伏纹理的数量分别降低了45.2%和53.2%，视觉上也取得了较理想的结果，具有一定的应用价值。

Visual Cognitive Connotation and Multivariate Blanking Method of DEM Relief Texture

Terrain is formed under the combined action of internal and external forces. Different types of terrain have different morphological characteristics. From the perspective of global or macro-observation, the topographic relief of local terrain and its characteristics can be visually perceived. Texture is a combination of regular lines on the surface of objects. DEM is the digital expression of terrain, and the morphology of different terrain types has a unique representation of texture. At present, the research on DEM texture is mainly from the perspective of DEM image cognition, and corresponding calculation methods of image texture are used to describe the terrain texture objects and extract indexes. Based on the obtained texture objects, the analysis of regional terrain characteristics and the classification of landform types are then carried out. However, DEM relief texture is a combination of terrain elements based on a certain view angle and scale. Firstly, it is necessary to make a systematic visual cognition of DEM relief texture and clarify the composition structure of DEM relief texture based on visual cognition and texture elements controlled by different visual variables. Through the realization of multidimensional visual modeling of DEM relief texture, the abstraction and simplification of DEM digital terrain modeling are achieved. Based on this, this paper takes DEM relief texture as the research target, carries out visual cognition on DEM relief texture, clarifies the meaning of DEM relief texture, and establishes the extraction method. The established method is used to realize the extraction of DEM relief texture, and good extraction results are achieved. Starting from the mechanism of visual perception, the combination of visual variables that determine the results of visual perception is determined. The parameters for visual synthesis including depth of field, length, and neighborhood are determined, and a multivariate DEM relief texture blanking method is designed and established. Finally, the visual variables are set and this method is used to realize the blanking of DEM relief texture. After blanking, the number of DEM relief textures is reduced by 45.2 % and 53.2 %, respectively. The visual results are also ideal, which indicates a certain application value of our methods.