摘要
在分析格网模式下城市道路网几何形态及空间分布特点的基础上,运用矩形度、平行度、主方向一致度、质心方向一致度、形状相似度等指标描述道路网眼特征,构建道路网模式识别输入参数组合,然后应用LightGBM方法对道路网眼进行分类,根据输出结果识别道路网格网模式。实验结果表明,该方法可以比较准确地识别城市道路网格网模式。
Based on the analysis of the geometric appearance and spatial arrangement characteristics of ur-ban road networks in grid pattern, the features of the road mesh are described by using the indi-cators of rectangularity, parallelism, main direction similarity, centroid direction similarity and shape similarity, formed a combination of parameters for model input of pattern recognition of road networks, then, use LightGBM to classify road mesh and identify the grid pattern of road networks according to the output. The experimental results show that the method can accurately identify the grid pattern of urban road networks.
出处
《测绘科学技术》
2020年第1期1-8,共8页
Geomatics Science and Technology
基金
国家重点研发计划(2017YFB0503601、2017YFB0503502)
国家自然科学基金(41671448)
四川省重点研发计划(19ZDYF0839)。
