基于 Landsat_8 影像的泾河峡谷段高含沙河流信息提取方法.

This paper proposed an improved method for automatic extraction of multi-band spectral relationship of narrow and deep rivers in mountains and valleys with high sediment concentration from Landsat 8 OLI images, which overcame the mixed pixel effect of mountain shad ow and exposed bedrock in the downstream section of the reservoir. Taking Jingeun to Dongzhuang Reservoir section in the middle reaches of Jinghe River on the Loess Plateau as an example, it separated the water body and mountain shadow, bare mountain and bedrock beach by considering the characteristics of the multi-band spectral, and then constructed a water body extraction model suitable for narrow and deep rivers with high sediment concentration in valleys. Compared with other water extraction models, the method proposed in this paper can over come the influence of shadow close to river channel and bedrock downstream of reservoir on water extraction, and improve the extraction accuracy of narrow and deep rivers with high sediment concentration in mountain canyon. [ABSTRACT FROM AUTHOR]