基于信息熵调优的交错瓦片地图符号设计优化方法

An Optimization Method for Interleaved Tile Map Symbol Design Based on Information Entropy Tuning

摘要: 随着矢量瓦片渲染技术的广泛应用，交错瓦片地图的符号设计面临多比例尺层级交错场景下的视觉一致性挑战。传统方法依赖制图师经验，缺乏定量评估手段，在不同比例尺地图数据共存的视图中易产生符号突变与视觉冲突。本文提出一种基于信息熵调优的交错瓦片地图符号设计优化方法。该方法以瓦片为基本计算单元，构建要素概率分布计算模型与样式熵计算模型，建立综合瓦片信息熵模型，通过迭代调优瓦片样式信息熵，指导地图符号库的持续优化。实验以交错瓦片为例，按1:1000至1:1000万共9个显示比例尺区间划分测试场景，配置对应数据比例尺组合并开展渐进加载性能测试。结果表明，应用本方法后瓦片整体信息熵从0.72降低至0.16，降幅达77.8%，显著提升了交错场景下地图符号显示的连续性与稳定性；可支持跨数据精度1:1000至1:1000万范围内的符号显示，以S4 (显示1:10万)场景为例，视口叠加图幅数N = 20时，平均帧率降至29 fps为性能拐点，视口内最多支持约20幅数据叠加显示，有效解决了交错瓦片地图多比例尺符号设计的定量优化难题。

Abstract: With the widespread application of vector tile rendering technology, the design of map symbols for interleaved tile maps faces significant visual consistency challenges in multi-scale interleaved scenarios. Traditional methods rely on cartographers’ experience and lack quantitative evaluation mechanisms, resulting in abrupt symbol transitions and visual conflicts when map data of different scales coexist in the same view. This paper proposes an optimization method for interleaved tile map symbol design based on information entropy tuning. Taking tiles as the basic computational unit, the method constructs a feature probability distribution model and a style entropy calculation model, establishes a comprehensive tile information entropy model, and guides the continuous optimization of the map symbol library through iterative tuning of tile style information entropy. Experiments use staggered tiles vector nautical charts as test data, dividing nine display-scale intervals from 1:1,000 to 1:10,000,000 with corresponding data-scale configurations and progressive loading performance tests. Results demonstrate that after applying this method, the overall tile information entropy decreases from 0.72 to 0.16, a reduction of 77.8%, significantly improving the display continuity and stability of map symbols in interleaved scenarios. The method supports cross-scale symbol display from 1:1,000 to 1:10,000,000; in the S4 scenario (display scale 1:100,000), the performance knee occurs at N = 20 overlapping map sheets with an average frame rate of 29 fps, effectively addressing the challenge of quantitative optimization in multi-scale symbol design for interleaved tile maps.