隨著感測器快速的發展,資料量急速增長的年代,如何把瞬息萬變的資料即時的轉換成能有幫助的資訊是相當重要的,在傳統地理空間資訊的架構中,難以即時對空間資料進行即時的運算,讓資料能夠即時轉換成有幫助的空間資訊,本研究提出了一個局部更新的模式,在感測器資料獲取時就先進行局部空間計算,減少每次空間分析的運算量。在本研究試驗中,50公尺網格的空間內插,局部更新的模式比傳統的模式減少了12秒的時間,在20公尺網格的空間內插中則減少了1分15秒,10公尺網格的空間內插中減少了4分30秒的時間。可以得出在精度越高的情況下減少運算時間的成效更顯著。在決定局部運算的區域範圍的方法中,以10公里的環域分析(Buffer)、K最近鄰分類演算法(KNN)及15公里的K近鄰演算法(KNN)法,較能表達出即時的空間上的測站變動狀況。希望在未來資料量越來越多的情況下,能夠減少更多的運算時間,達成即時的空間分析。
With the rapid development of sensors and data, processing data into useful information in real-time becomes important. Under the traditional structure of GIS, real-time processing and transferring useful information are difficult. The study proposed a partial update model that processes partial spatial information at the time of sensor data retrieval in order to reduce analysis computation. The results showed that partial update model was 12 seconds faster than traditional model in 50-m grid spatial analysis, 1 minute and 15 seconds faster in 20-m grid spatial analysis, and 4 minutes 30 seconds faster in 10-m grid spatial analysis. The model was significantly faster at higher precision. Among the methods determining the scope of partial processing, 10-km Buffer and KNN, and 15-km KNN methods were better at reflecting the change in the spatial status of observation stations in real-time. It is expected to reduce more processing time with increasing data to achieve real-time spatial analysis.