地图认知计算是现代地图学与地理信息科学领域的一个重要研究课题。当前相关研究主要是基于自然场景的视觉注意计算，考虑了空间目标颜色、密度、方向等信息，但没有考虑地图空间目标颜色、符号、注记等组合的几何形态复杂性与目标图形结构差异对用户视觉注意的刺激，以及任务与地图知识对用户视觉注意的引导。为此，本项目基于人脑视觉注意机制与眼动跟踪技术，将复杂的地图认知计算问题分为自底而上刺激驱动与自顶而下任务引导两个层次的视觉注意过程，并分别研究这两个层次的定量计算模型与方法，主要包括：（1）地图视觉显著性计算模型；（2）用户任务与地图知识语义度量方法；（3）眼动实验方案设计与眼动特征参数提取；（4）地图认知综合计算模型与评价方法。本项目一方面将建立更准确、合理的地图认知计算模型与方法，揭示地图空间认知规律；另一方面，提出的模型与方法可延伸应用于人机交互地图适人化设计、表达与智能决策等方面。

The computation of map cognition is one of the most important topics of Cartography and Geographical Information Science. Related works mainly focus on natural scene's visual attention process integrating color, orientation, size and orientation while ignoring visual information of graphic structure and geometric complexity of map shape, symbol, annotation and their combination. Influence of user task and map knowledge of guiding visual attention is also not been taken into consideration in relevant models. Based on visual attention mechanism and eye tracking technology, this project divides map cognition computation into two parts: the bottom-up (stimulus-driven) and top-down (task-guiding) visual information processes and investigate the computation models of the two processes. The research content mainly include: (1) computational modeling map visual saliency; (2) semantic measurements of user task and map knowledge; (3) design of eye tracking experiments and extraction of eye movement features; (4) comprehensive computation modeling of map cognition. The goal of this project is to develop a more accurate and reasonable computational model of map cognition to support humanization visualization of interactive maps, automated understanding and transferring of geographical information and intelligent interaction and decision making.