测绘地理信息

船舶室内安全问题受到越来越多的关注，在遇到突发事件时，GIS的最短路径分析可模拟最短逃生路径。鉴于船舶拓扑结构的复杂性，使用经典Dijkstra算法计算最短路径过于缓慢。首先，以一条9层甲板的船舶CAD数据为例，提出了一种基于浏览器/服务器（browser/server,B/S）结构的分层分区最短路径优化算法；然后，采用Java本地接口（Java native interface,JNI）编程方法调用底层C/C++编写的算法进行计算；最后，利用Web图形库（Webgraphics library,WebGL）技术实现船舶三维场景和最短逃生路径的可视化。结果表明，分层分区算法可有效削减计算量10%～80%，跨层数越多、跨区域数越少时，优化效果越明显；采用控制变量法对比分析可知，该算法的计算效率较未优化算法提高了50%以上，且在B/S架构下的效率较客户端/服务器（client/server,C/S）架构提升了25～50倍。

More and more attention has been paid to the indoor safety of ships. In case of emergencies, the shortest path analysis in GIS can simulate the shortest escape path. Given the complexity of the ship's topology, it is too slow to calculate the shortest path using the classical Dijkstra algorithm.Firstly, taking the CAD data of a 9-deck ship as an example, a shortest path optimization algorithm based on B/S architecture is proposed. Then, JNI programming method has been used to call the algorithm written in C/C++ language to calculate the shortest escape path between cabins. Finally, WebGL technology is used to visualize the ship's 3D scene and the shortest escape path. The experimental results show that the hierarchical partitioning algorithm can effectively reduce the calculation amount by 10%～80%. The optimization effect will be more obvious when the number of cross-layers is more and the number of cross-regions is less. As shown by the comparative analysis results using the control variables method, the efficiency of the algorithm is more than 50% higher than that of the unoptimized algorithm, and the efficiency of the algorithm under B/S architecture is about 25 to 50 times higher than that under C/S architecture. The proposed algorithm can significantly reduce the calculation time due to its high efficiency of hierarchical partitioning method and cross-platform characteristic of the B/S architecture. Thus, it can provide valuable decision support for the ship's escape and rescue work.