应用累积和分析算法的地形起伏度最佳统计单元确定

Extracting optimum statistical unit for relief degree of land surface with CUSUM algorithm

地形起伏度是开展区域滑坡敏感性和危险性评价的基础数据,其提取精度直接影响区域滑坡危险性评价和风险评估的正确性。本文以四川省低山丘陵区为研究区,基于邻域统计分析算法提取不同窗口半径下的平均地形起伏度,表明平均地形起伏度与分析窗口尺度存在显著的对数变化关系;根据地貌发育理论,应用累积和分析算法(CUSUM)对"平均地形起伏度-分析窗口半径"曲线的突变情况进行分析,通过精度检验,确定该区域地形起伏度提取的最佳统计窗口半径为1.1km。研究结果对于提高相似区域的滑坡灾害危险性评价和风险评估的精度有参考价值。

Relief degree of land surface whose higher extraction precision is of the benefits to hazard assessment precision of regional geo-environment is basis data for developing susceptibility and hazard as- sessment of regional landslide hazard. This paper chose hilly area in Sichuan as study area and found that there was a remarkable logarithmic relationship between average relief degree of land surface and window scale after extracting average relief degree of land surface under different window radius based on neighbor- hood statistics analysis. According to the theory of physiognomy evolution, it analyzed the saltation of Av- erage Relief Degree of Land Surface-Window Radius curve with CUSUM （Cumulative Sum） method and came to conclusion that the best statistics window radius was 1.1kin for extraction of relief degree of land surface in this area, The study result would provide a reference for improving the precision of landslide sus- ceptibility and hazard assessment in similar areas.