利用Sentinel-2光谱指数和改进的单类随机森林的塑料大棚提取方法

利用Sentinel-2光谱指数和改进的单类随机森林的塑料大棚提取方法

肖明珠, 李培军

A method for plastic greenhouse extraction integrating Sentinel-2 spectral indices and an improved one-class random forest

XIAO Mingzhu, LI PeiJjun

表3 本研究使用的塑料大棚指数公式

Tab.3 The plastic greenhouse indices and the equations used in this study

指数名	公式	变量含义
V_I	$V_{I} = \frac{R_{s w i r 1} - R_{n i r}}{R_{s w i r 1} + R_{n i r}} \times \frac{R_{n i r} - R_{r e d}}{R_{n i r} + R_{r e d}}$	$R_{c o a s t a l}$ , $R_{b l u e}$ , $R_{g r e e n}$ , $R_{r e d}$ , $R_{n i r}$ , $R_{s w i r 1}$ , $R_{s w i r 2}$ 分别为海岸气溶胶、蓝光、绿光、红光、近红外、短波红外1、短波红外2波段的光谱反射率; $λ_{L P}$ 与 $λ_{R P}$ 分别为上述一组波段中最短和最长的波长; $R_{i}$ 为波长i对应波段的反射率
PMLI	$P M L I = \frac{R_{s w i r 1} - R_{r e d}}{R_{s w i r 1} + R_{r e d}}$
MDI	$M D I = \overset{λ_{L P}}{\sum_{i = λ_{R P}}} \sqrt[]{R_{i}^{2} + (λ_{R P} {- i)}^{2}} - \overset{λ_{R P}}{\sum_{i = λ_{L P}}} \sqrt[]{R_{i}^{2} + (i - λ_{L P})^{2}}$
GDI	$G D I = \frac{M D I}{3} - \frac{R_{b l u e} - (\frac{R_{s w i r 1} + R_{s w i r 2}}{2})}{R_{b l u e} + (\frac{R_{s w i r 1} + R_{s w i r 2}}{2})}$
PGI	$P G I = \{\begin{array}{l} 0, N D V I > 0.73 \\ 100 \times \frac{R_{b l u e} \times (R_{n i r} - R_{r e d})}{1 - \frac{R_{b l u e} + R_{g r e e n} + R_{n i r}}{3}} \\ 0, N D B I > 0.005 \end{array}, 其他$
APGI	$A P G I = 100 \times R_{c o a s t a l} \times R_{r e d} \times \frac{2 \times R_{n i r} - R_{r e d} - R_{s w i r 2}}{2 \times R_{n i r} + R_{r e d} + R_{s w i r 2}}$