標題: Kriging方法於地質圖製作之應用
The Application of the Kriging Method on the Construction
作者: 黃明萬
Ming-Wan Huang
潘以文
Yii-Wen Pan
土木工程學系
關鍵字: Kriging;地質圖;地質調查;地層位態;Kriging;geological maps;digital terrain model
公開日期: 2000
摘要: 本文嘗試將Kriging法則導入地質圖製作過程,主要工作在於利用Kriging法則來處理地質調查成果所得的原始資料,並以電腦自動化完成地質圖的製作。 本研究以MATLAB為程式語言,引用DTM數值地形模型資料製作地形圖,並套繪地質調查所得資料點,選擇具地層空間特性的變異圖(variogram),將地層位態關係以每一資料點形成一空間上單斜面的方式加入Kriging預測機制中,對空間上的地質狀況作模擬,使用預測結果繪製地層分界線、地層構造等高線、地質剖面圖以及預測結果之誤差變異數等高線圖。完成之結果與其他地質圖自動化製作程式之結果作比對均相當吻合。 一個地層空間分佈預測結果好壞的判斷,必須綜合性的觀察所有結果,地層出露線、地層構造線及預測之誤差變異數分佈狀況都合理的狀況下,才是一個好的預測機制。影響Kriging法則預測結果的主要因素為變異圖模式的種類與變異圖模式的關連距離﹔基值(sill)的改變只對預測之誤差變異數產生影響,但此影響可藉正規化誤差變異數的方式加以消除。預測結果之誤差變異數可作為規劃補充地質調查計畫之輔助,Kriging法則也可使用來評估補充調查預期可以得到的可靠度改善程度。
This study aims to study the possibility to generate geological maps by the Kriging method. The construction of a geological model makes use of bedding information including orientations data from outcrops and rock-cores. From the DTM and the geological model, it is then possible to obtain the geologic boundaries (or contact lines), and to draw a geological map or profile. The presented work utilizes MATLAB for the necessary programming. A unique surface grid and contour are constructed from digital terrain model (DTM). Next, the Kriging method is utilized to obtain the best estimation of geologic contours. The geological map then can be obtained by overlaying the contour representing the DTM and all geological boundaries (bedding boundaries) on a same map. The presented method is found reasonable for modeling outcrop trace of geologic formations. Besides obtaining the outcrop trace, the Kriging method also provides the distribution of the variance of estimated error. The results of the estimation of geologic model by the Kriging method depend factors including the available data, geologic structures, the assumed variogram model, and so on. The major factors are the assumed variogram model and the range of correlation. The value of sill only affects the variance of estimated error; its influence may be removed by a normalization of the variance of estimated error. It is possible to make use of the distribution of the variance of estimated error to evaluate the effect and efficiency of some new data from further site investigation. This information will be very helpful for determining the focus of further investigation.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT890015035
http://hdl.handle.net/11536/66426
Appears in Collections:Thesis