【摘要】 多尺度分割方法划分斜坡单元，存在分割尺度选择主观性强、试验耗时以及划分结果适宜评价指标单一等问题。以福建武平十方镇为研究区，固定分割尺度，试验确定出多尺度分割方法中最佳同质性参数的组合，利用同质性局部方差变化率曲线厘定可能的分割尺度，不同尺度结合同质性参数组合进行斜坡单元划分。从斜坡单元与历史滑坡几何联系、斜坡单元划分原理、斜坡单元现实应用三个方面，初步建立了评价斜坡单元划分结果是否适宜的3项指标：形状指数、坡向全局莫兰指数和地形起伏度。研究结果表明，对多尺度分割优化方法的优化降低了分割尺度选择的主观性并显著减少了处理时间，适宜单元评价指标确定出的斜坡单元更加切合实际地形地貌。更多还原

【Abstract】 The application of multi-scale segmentation method to divide slope units is confronted with issues such as subjective selection of segmentation scales, time-consuming experiments and limited suitability evaluation indicators for the results. This paper focuses on Shifang Town, Wuping County, Fujian Province as the study area, fixes the segmentation scale, and determines the optimal combination of homogeneity parameters through experimental methods in the multi-scale segmentation approach. By utilizing the homogeneity local variance change rate curve, potential segmentation scales are determined, and slope units are delineated by combining different scales with homogeneity parameter combinations. Initially, three indicators are established for evaluating the suitability of slope unit delineation results from three perspectives: geometric relationship between slope units and historical landslides, principles of slope unit delineation, and practical applications of slope units. These indicators include shape index, global Moran’s I for aspect, and relief degree of landform. Case studies demonstrate that the optimization of the multi-scale segmentation method reduces the subjectivity in selecting segmentation scales and significantly decreases processing time. Additionally, the slope units determined by the suitable unit evaluation indicators are more aligned with the actual landforms and terrain features.更多还原