균열 방향성 자료의 밀도 분석 및 자동 군집화 알고리즘 연구

초록

The orientation of fractures distributed within rock masses is an important factor that affects the behavior of rock masses. Therefore, the orientation data of the borehole images measured in the field are data that must be closely analyzed in engineering activities targeting rock masses. These fractures can be classified into clusters with similar orientations, and the representative characteristics of fractures in each cluster are used to reflect the influence of the fractures on the behavior of the rock mass in engineering analysis. In this study, the orientation data obtained from the borehole image were used to automatically cluster the fractures distributed in the target rock mass, and the results were visualized to present the distribution and statistical characteristics of the fractures in the rock mass. In this process, the Fisher mixture model method was developed as an automatic clustering algorithm based on Fisher distribution, and several alternatives for rational and efficient setting of initial para- meters were presented.

키워드