基于热点区域簇群的林区地图瓦片缓存策略

Forest Map Tile Cache Strategy Based on Hot Area Regional Cluster

为了提升林区瓦片地图的运行与服务效率,本文针对瓦片的特点以及用户操作习惯,提出一种基于区域簇群(Regional Cluster)的瓦片缓存策略。该策略从林区地图的区域性出发,根据瓦片记录信息,首先通过全局空间自相关分析确定聚类类型,再通过局部空间自相关分析出热点区域。策略结合瓦片地图的塔状结构特点,对传统的LFU算法进行改进,保留热点区域瓦片金字塔内的瓦片数据群,提高瓦片缓存命中率。实验表明,该策略能够提高瓦片缓存命中率,加快林区瓦片地图的访问速度。

To improve the operation and service efficiency of tile map,this paper proposes a tile caching strategy based on Regional Cluster according to the characteristics of tile and user operation habits.Based on the regional map of forest region and tile record information,this strategy firstly determines the clustering type through global spatial autocorrelation analysis,and then identifies hot spots through local spatial autocorrelation analysis.According to the tower structure characteristics of tile map,the traditional LFU algorithm is improved to preserve the tile data group in tile pyramid of hot spot area and improve the hit rate of tile cache.Experiments demonstrate that the strategy can improve the tile cache hit rate and accelerate the tile map access speed.