本論文目的是協助 AIS 航路歸納、高風險區分析功能應用試 驗，以檢討現行航路之合理性及安全性，可輔助船舶航路規劃、勤務 派遣以及船舶的路徑與追蹤。 利用Douglas-Peucker 演算法作航跡Generalization(概略化)。 利 用DBSCAN 演算法找出船隻節點的密集區域及航跡走向。再以格子 基礎密度演算法(Grid-Based Density Algorithm)繪出各個航道中心線 提供各船舶航行路線參考。將台灣附近海域形成一幅海上航路網，不 管是國內船舶在港口往來亦或是國外航線往來，利用分析出的航路網 判斷其船隻預計到達時間、可能航行路線。 This thesis aims to extract the most taken routes and high-risk areas in Taiwan waters in order to review current routes by efficiency and safety, assist the planning of ships’ routing, service dispatch, and the tracking of ships. Various spatial analysis techniques are used to explore the database of vessel dynamic reports received from a coastal network of automatic identification system (AIS) for application testing. Track generalization is implemented based on Douglas-Peucker Algorithm. DBSCAN Algorithm was used to find out the clustering of waypoints and crossovers of vessel tracks as well as directions of tracks. Grid-based line density algorithm was used to extract the center line of the most take route. Results from these analyses form a network of sea routes around Taiwan waters. The time of ship’s arrival and the routes that ships may sail can then be estimated with the aid of such information of the route network, no matter domestic routes or international routes.