宝鸡黄土区滑坡遥感调查中遥感数据尺度问题探讨

遥感数据的尺度效应决定了可识别内容与识别精度，同-区域内的遥感数据尺度不同，同样的遥感处理模型或者方法将得到不同的处理结果。以宝鸡黄土区滑坡遥感调查为基础，对不同遥感数据源的滑坡体的最小可识别面积、图斑面积精度测算、最佳及最大成图比例尺、遥感地质灾害解译对比性分析、滑坡遥感解译精度评价等等与遥感尺度有密切关系的问题进行了探讨。研究结果表明，在宝鸡黄土区，调查大、中型以上滑坡的信息，可以采用SPOT-5（2.5 m）融合图像数据，比例尺为1:25000或1:50000；调查中型、小型滑坡及较大滑坡体内部结构定量信息，可以采用QuickBird（0.61 m）融合图像数据，比例尺为1:5000。

DISCUSSION ON THE SCALE PROBLEM IN REMOTE SENSING DATA FOR LOESS LANDSLIDE SURVEY IN BAOJI, SHAANXI PROVINCE, CHIAN

The content and accuracy are determined by the scale of remote sensing(RS) data, the RS Processing molds or methods vary with Remote sensing data scale in the same region. Based on the landslide RS survey in Baoji loess area, we discuss the problems that have great relationship with RS scale, such as Mininum regional area of landslide mass in different RS data source, precision measurement of spot area, the best and maxinum mapping scale, comparative analysis of interpretation of RS geological hazards, accuracy evaluation of landslide RS interpretation and so on. The result reveals that when investgating large and medium-sized landslide in Baoji loess area, SPOT-5(2.5 m)Fused image data can be used, and the scale is 1:25000 or 1:50000. Meanwhile, QuickBird(0.61m)Fused image data is used for medium-small landslide and quantitative information calculation on the internal structure of larger landslide, and the scale is 1:5000.