| Summary: | 碩士 === 正修科技大學 === 營建工程研究所 === 101 === In southern Taiwan rainfall occurs in the plum rain season from May to June and in the typhoon season from July to September. Furthermore, seasonal rainfall and frequent earthquakes combined the weak geology often cause slope failure. Hence, the traffic safety and the protection of human life and property will be endangered.
In this research, the Tai-20 highway in Jiasian area was selected as experimental sites to extract potential associated factors of slope by the application of Remote Sensing and Geographic Information Systems. By referencing previous publications, seven factors were chosen to evaluate the slope failure potential: slope, aspect, elevation, Normalized Difference Vegetation Index (NDVI), geology, distance from fault lines, distance from rivers. The prediction model was built up by the following steps. First, we used 5m×5m grid to build 17889 data points of analysis factor. And we constructed a cliff face potential analysis model akin to a neural network. Last, we performed regression analysis by digitizing aerial photographs of the failure area layers and Central Weather Bureau rainfall data.
According to slope failure data, rainfall characteristics and potential values, we conducted the validation of failure prediction model, and set damage threshold above 0.9. In addition, by GIS analysis and the comparison of forecast and actual failure area, The result showed that small failure type (100m2~200m2) and large failure type (500m2 or more) are more closer to the fact. Another simulation analysis for a torrential rain. The result displayed 29 slopes destroyed.
|
|---|
