基于知识表达和推理的河网自动选取

Automated river selection based on cartographic knowledge representation and reasoning

针对河系是地图上重要的要素,它的自动选取是地图综合的基本手段之一,把基于知识的人工智能方法引入地图自动综合是一种合理选择。把河系自动综合中的知识分为空间知识、属性知识及综合规则,研究了这三类知识的获取策略和形式化方法,建立了河系自动综合的产生式规则库;在规则库中依据规则驱动采用正向推理完成河系选取的推理过程,推理中全面考虑了河网知识和综合中各条规则的优先级和权重。通过实验验证了该方法对保持河系的整体结构特征的有效性。

River network is one of the most important features in maps. River selection is a key operation in automated map generalization. The knowledge in automated river selection in the paper includes structural river network data, river properties and selection rules. The methods of capturing and formalizing the knowledge are studied. Following the knowledge representation, it brings up a method to build up automated selection rules-base of river. Because the process affects the final result directly, the PRI and weight of using rules are the key points. Ordinal reasoning of production rules system is applied in the river selection, which is bases on the known verity and reasons to the conclusion by rules. Then repeat this process, until got a result is got whether the river is selected. At last, a sample is designed to explain how to use this method to select rivers of a complex river network.