基于Landsat 8影像的黄土干旱区地表水提取——以宝鸡市北部地区为例

Extraction of surface water information based on Landsat 8 image:a case study in the loess-covered area of northern Baoji city

基于遥感影像的水体信息提取对于陆表水资源调查和监测具有重要意义。以Landsat 8 OLI影像为数据源,秦岭北麓某黄土覆盖区为研究区,探讨了反映水体像元值的数理、空间统计分布特征。结果表明:OLI B4/B7影像可有效突出各类地表水体信息;研究区水体异常服从多重分形分布,根据无标度区走势可划分出3种不同性质的自然水体,即以大型水库为代表的深水水体、浅水水体(包括浅滩和湿地)和小规模汇水区;假异常较之于水体异常具有不稳定性和空间离散性,可分别通过GIS叠置分析、热点分析予以剔除。野外踏勘证实本研究水体解译精度接近100%。

Extraction of surface water information by the remote sensing image is of great significance to the surveying and monitoring of regional water resources.In this article,Landsat 8 OLI image is used as the data source,and a loess coverage area in the northern foot of the Qinling Mountains is taken as the study area.The study focuses on the characteristics of the mathematical and spatial statistics of the pixel values in the enhanced image that can expose the water information.The results showed that(1)The B4/B7 ratioing image of the OLI can effectively highlight the water information of different properties.(2)Pixels reflecting the water are proved to be fractally distributed.Three different properties of the natural water can be clarified as deep water like some large reservoirs,shallow water(including shoals and wetlands),and small-scale catchment areas.(3)False anomalies are more unstable and dispersive than the water anomalies.These punctate false anomalies can reasonably be removed by GIS overlay analysis and hotspot analysis.The accuracy of this interpretation turns out to be close to 100%.