在地理資訊系統中，常需利用圖形中之物件進行查詢及分析。在此論文中， 我們提出一個新的資料結構，稱為CBS字串(Coordinate-based symbolic string)，做為圖形之符號表示。我們用物件投影在X軸和Y軸上之起始邊界和結束邊界的真實座標來表示物件，並利用兩個決策樹來推論物件間在一維空間和二維空間上的空間關係。因此在推論空間關係時並不需要將影像重建，較有效率。同時我們考慮物件大小和物件間距離，除了能正確的重建影像外，並能有效率的處理需要這些尺寸資訊的查詢及分析。 In geographical information systems, it is often to perform query and analysis based on the objects in images. In this paper, we propose a new data structure, called the CBS string (Coordinate-Based Symbolic String), as the symbolic representation of images. We use the real coordinates of the projection of the objects along X-axis and Y-axis to represent objects. Also, two decision trees are used to infer the spatial relationships between objects in the one-dimensional and two-dimensional space. Hence, the image reconstruction is not required in the process of spatial inference and the process will be more efficient. In the mean time, we consider the size of the object and the distance between objects. In addition to the accurate image reconstruction, the process of query and analysis requiring those metric information can be handled efficiently.