Direction-relation similarity model based on fuzzy close-degree

Directional similarity is widely used in the assessment of spatial scene similarity and spatial query. Based on different representation models, direction-relation similarity models should be partitioned into quantitative and qualitative models. The traditional qualitative direction-relation similarity models based on directional matrix model assess the directional similarity by computing the least cost for transforming one directional matrix into another. These models are limited to be used in the directional matrices whose elements sum up to 1. This paper proposes a qualitative directional similarity computation model to evaluate the directional similarity, which is based on the schema of fuzzy close-degree. The new model ignores the relations between the cardinal directions and has a wide range application in the similarity assessment.