面轮廓线相似性和复杂性度量及在化简中的应用

本文对同名实体相似性和复杂性进行综合研究,通过抽取多个尺度下的内弦长、外弦长及弧到弦的平均投影长度构造弦特征矩阵,建立了多尺度面轮廓线相似性度量模型。依据边的振动频率及幅度,从凹口比例、凸壳周长相对增量及最小外包圆面积相对增量3个方面建立复杂性度量模型。在验证两类模型正确性和有效性的基础上,分析了不同复杂度目标在不同尺度下相似性变化规律,据此给出基于相似性约束的自动化简的一般流程。试验表明,本文方法可为化简算法设置合理阈值参数,增加化简算法的自动化程度。

Polygon contour similarity and complexity measurement and application in simplification

On a comprehensive study concerning polygon contour similarity and complexity, in which the inner chord length, outer chord length, and average arc-to-chord length corresponding to multi-scale entities are extracted to construct chord-feature matrices. A multi-scale polygon contour similarity measurement model is developed. According to the frequency and amplitude of the edges, a complexity measurement model is constructed considering the ratio of the notch, relative increment of the perimeter of the convex shell, and relative increment of the area of the minimum envelope. After verifying the validity and effectiveness of the two models, the similarity variation rules for entities of different complexity at different scales are analyzed, and a general automation simplification flow based on similarity constraints is obtained. Results of the experiments show that the proposed method can set reasonable threshold parameters for the simplification algorithm, and improve the level of automation.