基于无人机高光谱影像和机器学习的红树林树种精细分类

姜玉峰,齐建国,陈博伟,闫敏,黄龙吉,张丽

Classification of Mangrove Species with UAV Hyperspectral Imagery and Machine Learning Methods

Yufeng Jiang,Jianguo Qi,Bowei Chen,Min Yan,Longji Huang,Li Zhang

表2 高光谱影像提取特征

Table 2 Hyperspectral image extraction features

特征类型	特征名称	描述或公式	重要度排序
光谱特征	NDVI ₇₀₅	$\frac{ρ_{750} - ρ_{705}}{ρ_{750} + ρ_{705}}$	1
	SR	$\frac{ρ_{N I R}}{ρ_{R E D}}$	2
	NDVI	$\frac{ρ_{N I R} - ρ_{R E D}}{ρ_{N I R} + ρ_{R E D}}$	3
	VOG₁	$\frac{ρ_{740}}{ρ_{720}}$	4
	CRI₂	$\frac{1}{ρ_{510}} - \frac{1}{ρ_{700}}$	5
	CRI₁	$\frac{1}{ρ_{510}} - \frac{1}{ρ_{550}}$	7
	RG	$\frac{ρ_{\overset{ˉ}{R E D}}}{ρ_{\overset{ˉ}{G R E E N}}}$	8
	PRI	$\frac{ρ_{531} - ρ_{570}}{ρ_{531} + ρ_{570}}$	9
纹理特征	Mean_3*3_B29	$\frac{\sum_{i, j = 0}^{N - 1} P_{i j}}{N^{2}}$	6
纹理特征	Entropy_5*5_mean	$\frac{\sum_{B = 1}^{45} \sum_{k = 0}^{N - 1} V_{k} (- L n V_{k})}{45}$	10