一种基于Landsat 8的多波段组合水体指数模型

本文通过分析水体在Landsat 8数据中可见光波段和近红外波段的波谱差异,将Landsat 8数据中可见光波段作为一组,近红外波段和中红外波段作为另一组,构建了多波段组合水体指数(MBCWI)模型。基于Landsat 8数据在合肥、安康和康定地区共3景数据5种不同场景进行水体提取试验。结果表明,该模型不仅能够抑制云层、阴影、裸土、亮色地物和建筑物等对水体提取的影响,还能较好地提取出含有大量蓝藻的水体,且阈值稳定,Kappa系数优于0.968 5,总体精度高达99.69%,总体误差小于8.92%。相较于其他水体指数而言,提取精度显著提高。

A multi-band combination water index model based on Landsat 8

In this paper, by analyzing the spectral difference between the visible light band and the near-infrared band in the Landsat 8 data, the visible light band in the Landsat 8 data is taken as one group, and the near-infrared band and the mid-infrared band are taken as another group, and a multi-band combination water index (MBCWI) model is constructed. Based on Landsat 8 data, water extraction experiments were carried out in Hefei, Ankang and Kangding areas with a total of 3 scene data and 5 different scenarios. The results show that the model can not only suppress the influence of clouds, shadows, bare soil, bright-colored ground objects and buildings on water extraction, but also can extract water containing a large number of cyanobacteria well, and the threshold is stable, and the Kappa coefficient is better than 0.968 5, the overall accuracy is as high as 99.69%, and the overall error is less than 8.92%. Compared with other water indices, the extraction accuracy is significantly improved.