Detection of Pest Degree of Phyllostachys Chinese With Hyperspectral Data

doi:10.3964/j.issn.1000-0593(2021)10-3200-08

Journals >Spectroscopy and Spectral Analysis >Volume 41 >Issue 10 >Page 3200 > Article

Spectroscopy and Spectral Analysis
Vol. 41, Issue 10, 3200 (2021)

Detection of Pest Degree of Phyllostachys Chinese With Hyperspectral Data

DOI: 10.3964/j.issn.1000-0593(2021)10-3200-08 Cite this Article

Copy Citation Text

show less

Fig. 1. Location map of test area

Download full size | View in the Article

Fig. 2. Overall structure diagram

Download full size | View in the Article

Fig. 3. Original spectra and the curves after removing the spectral envelopes of samples under different injury levels

Download full size | View in the Article

Fig. 4. Hyperspectral remote sensing image of UAV

Download full size | View in the Article

Fig. 5. Spatial distribution of pest injury levels in sampling area

Download full size | View in the Article

植被指数	名称	计算公式
SAVI	土壤调节植被指数	1.5(ρ₈₇₀-ρ₆₈₀)/(ρ₈₇₀+ρ₆₈₀+0.5)
PSNDa	归一化色素差值指数a	(ρ₈₀₀-ρ₆₉₀)/(R₈₀₀+ρ₆₉₀)
PSNDb	归一化色素差值指数b	(ρ₈₀₀-ρ₆₃₅)/(ρ₈₀₀+ρ₆₃₅)
mND₇₀₅	修正归一化差值指数	(ρ₇₅₀-ρ₇₀₅)/(ρ₇₅₀+ρ₇₀₅-2ρ₄₄₅)
VOGa	Vogelman指数a	ρ₇₄₀/ρ₇₅₀
mSR₇₀₅	改进比值植被指数	(ρ₇₅₀-ρ₄₄₅)/(ρ₇₀₅+ρ₄₄₅)
DD	双重差异指数	(ρ₇₅₀-ρ₇₂₀)-(ρ₇₀₀+ρ₆₇₀)
NDI	归一化差异指数	(ρ₇₅₀-ρ₇₀₅)/(ρ₇₅₀+ρ₇₀₅)
GNDVI	绿光归一化植被指数	(ρ₇₅₀-ρ₅₅₀)/(ρ₇₅₀+ρ₅₅₀)

Table 1. Vegetation index and its calculation formula

View in the Article

参数	名称	定义
Depth₅₆₀	绿峰吸收深度	在560 nm处的吸收深度
Area_G	绿峰吸收面积	在500~610 nm之间包络线与光谱反射率之间的面积
Area_R	红边吸收面积	在680~760 nm之间包络线与光谱反射率之间的面积
λ_R	红边位置	在680~760 nm之间反射率的一阶导数红边最大值位置
dλ_R	红边振幅	在680~760 nm之间反射率的一阶导数红边最大值
SD_R	红边面积	在680~760 nm之间反射率的一阶导数红边面积
SD_R-SD_B	面积差值	红边面积与蓝边面积的差值

Table 2. Spectral characteristic parameters and its definitions

View in the Article

函数	ρ₄₀₀	ρ₄₀₆	ρ₅₈₆	ρ₅₉₃	ρ₆₈₉	ρ₈₇₆	常数	方差/%
y₁	788.24	961.86	-382.94	422.87	-204.32	40.16	-5.71	89.3
y₂	3721.39	-3461.94	-101.38	-109.92	74.46	5.53	-0.20	9.7
y₃	-6028.94	6204.72	352.27	-646.97	222.93	4.72	-5.74	1.0

Table 3. Discriminant function based on the original spectrum

View in the Article

函数	ρ₄₀₃	ρ₄₀₆	ρ₄₀₉	ρ₄₁₂	ρ₅₀₅	ρ₅₁₅	ρ₇₃₅	ρ₇₄₉	常数	方差/%
y₁	-113.40	-18.91	157.03	-49.78	-51.92	14.14	194.80	-376.44	239.65	88.3
y₂	59.12	-262.44	82.09	99.94	-74.53	42.71	26.86	-5.17	18.74	9.1
y₃	44.92	-162.80	-194.94	203.22	120.66	-78.28	-23.33	86.46	24.46	2.7

Table 4. Discriminant function based on the spectrum after removing the envelope

View in the Article

函数	PSNDa	PSNDb	mND₇₀₅	mSR₇₀₅	NDI	SAVI	VOGa	DD	GNDVI	常数	方差/%
y₁	-32.64	-53.58	-764.78	-7.54	224.85	19.12	116.44	24.52	-1.15	-97.10	87.2
y₂	-64.11	13.36	136.33	20.61	-162.59	-5.76	132.71	35.06	21.21	-127.46	11.0
y₃	99.62	-71.51	121.12	9.22	-117.48	-34.92	140.24	-107.91	63.65	-189.80	1.8

Table 5. Discriminant function based on canopy vegetation indexes

View in the Article

函数	Depth₅₆₀	Area_R	Area_G	λ_R	dλ_R	SD_R	SD_R-SD_B	常数	方差/%
y₁	41.84	0.62	0.23	-68.53	0.03	228.59	-242.23	-82.95	84.0
y₂	60.65	-0.23	0.79	1829.17	-0.15	-264.07	174.68	53.37	14.2
y₃	156.17	0.10	1.84	-461.90	-0.08	324.33	-239.62	-122.10	1.8

Table 6. Discriminant function based on canopy spectral parameters

View in the Article

特征	检验精度/%					Kappa 系数	R²
特征	健康	轻度	中度	重度	总体	Kappa 系数	R²
原始光谱	92.0	73.3	82.4	84.8	84.4	0.79	0.89
去包络线	96.0	60.0	70.6	84.8	81.1	0.74	0.88
植被指数	96.0	73.3	70.6	78.8	79.7	0.74	0.88
光谱参数	92.0	73.3	64.7	84.8	78.7	0.76	0.85

Table 7. Different feature discriminant function accuracies

View in the Article

采样区域	健康		轻度虫害		中度虫害		重度虫害		总面积 /m²
采样区域	面积/m²	比例/%	面积/m²	比例/%	面积/m²	比例/%	面积/m²	比例/%	总面积 /m²
洋门A	222.16	2.32	1 395.09	14.56	2 530.45	26.41	5 432.30	56.70	9 580
洋门B	855.03	6.47	4 228.62	32.01	5 893.71	44.62	2 231.64	16.89	13 209
上湖A	4 815.13	43.89	2 004.44	18.27	3 410.13	31.09	740.30	6.75	10 970
上湖B	2 700.81	43.91	1 470.34	23.90	1 793.49	29.16	186.36	3.03	6 151

Table 8. Test results of Phyllostachys Edulis Pest

Download Citation

Save the article for my favorites

Paper Information