提出了一种基于关联规则分类算法进行道路网网格模式识别的方法。首先基于空间认知理论,从人类视觉认知Gestalt原理出发,采用几何与拓扑关系特征量化计算网眼形状参量,从网眼自身形状指标、相邻网眼排列一致性、相邻网眼形状相似性3方面描述网眼模式。然后利用关联规则分类算法对道路网网格描述特征进行识别分类。实验表明,本文方法有效可行,具有一定的鲁棒性。
The association rule based classification approach was proposed to grid pattern recognition in road net- works. On the basis of spatial cognition and Gestalt principles, road meshes were characterized by geometric and topological relationship on aspects of shape criterion, consistent arrangement and neighbor mesh similarity; then the parameters were used to classify the road meshes with the method of classification based on association to recog- nize grid pattern. Experimental result showed that this approach was effective, feasible and practicable for grid pattern recognition.