| Summary: | 碩士 === 國立成功大學 === 土木工程學系 === 104 === In recent year, three-dimensional laser scanner technology with high precision and high efficiency constantly updated. It’s using to modeling the building, heritage conservation, monitor bridge deformation, post-earthquake reconnaissance, coastal terrain retreat monitoring, large-scale terrain change monitoring and etc. It has gradually replaced the traditional measurement technology. However, currently technologies in the large number of scattered point cloud processing applications are concentrated in specific areas such as building models or point cloud data simulation. And the point cloud data will contain a lot of noise to the need for further processing. So the application of point cloud data is in the research and development stage, needs sustainable development. This study is divided into two directions, namely, automatic object recognition and post-earthquake reconnaissance. The first theme focuses on the automatic identification of specific objects from the scattered cloud data, the original point cloud data operated according to the set of calculation strategy. Using of region grow method will be scattered point cloud data classification. And then use all kinds of algorithms, such as boundary extraction method to extract the characteristics of clustering group. The final cross-match point group feature to recognize particular object. The latter theme mainly by investigation the disaster caused by the earthquake. Study the structural damage analysis caused by seismic force, Such as structural deformation and displacement, surface damage, records and Extraction of the earthquake induced large landslide. And put forward specific damage assessment, quantitative indicators. The results show that the feasibility of automatic identification of specific objects, and several quantitative assessment results can be used to assess structural damage.
|
|---|
