• Spectroscopy and Spectral Analysis
  • Vol. 42, Issue 3, 678 (2022)
Tong JI1、1; 2;, Bo WANG1、1; 2;, Jun-ying YANG1、1; 2;, Qiang LI1、1; 2;, Guo-xing HE1、1; 2;, Dong-rong PAN3、3;, and Xiao-ni LIU1、1; 2; *;
Author Affiliations
  • 11. College of Pratacultural Science, Gansu Agricultural University, Lanzhou 730070, China
  • 33. Grassland Technique Extension Station of Gansu Province, Lanzhou 730070, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2022)03-0678-08 Cite this Article
    Tong JI, Bo WANG, Jun-ying YANG, Qiang LI, Guo-xing HE, Dong-rong PAN, Xiao-ni LIU. Spectral Characteristic Analysis and Spectral Identification of Desert Plants in Yanchi, Ningxia[J]. Spectroscopy and Spectral Analysis, 2022, 42(3): 678 Copy Citation Text show less
    The original reflectance spectra of desert grassland plants
    Fig. 1. The original reflectance spectra of desert grassland plants
    Obfuscation matrix of random forest model Note: the off-diagonal bubbles represent misjudgment, and the size of the bubbles represents the number of judgment. The larger the number of samples, the larger the bubble
    Fig. 2. Obfuscation matrix of random forest model Note: the off-diagonal bubbles represent misjudgment, and the size of the bubbles represents the number of judgment. The larger the number of samples, the larger the bubble
    Variale importance of random forest classification model Note: The blue column in the figure is the importance of the variable. The larger the value of the variable, the stronger the importance of the variable. The yellow column is the Gini coefficient. The higher the coefficient, the better the classification cut
    Fig. 3. Variale importance of random forest classification model Note: The blue column in the figure is the importance of the variable. The larger the value of the variable, the stronger the importance of the variable. The yellow column is the Gini coefficient. The higher the coefficient, the better the classification cut
    Obfuscation matrix of SVM model
    Fig. 4. Obfuscation matrix of SVM model
    KNN model error graph
    Fig. 5. KNN model error graph
    Obfuscation matrix of KNN model
    Fig. 6. Obfuscation matrix of KNN model
    Original reflectance spectra
    Fig. 7. Original reflectance spectra
    试验地海拔/m地理坐标草原类型优势种
    二步坑1 45538.080°N, 102.765°E温性荒漠草原甘草
    冯记沟1 40137.699°N, 106.878°E温性荒漠草原针茅
    高沙窝1 45838.047°N, 107.065°E温性荒漠草原蒙古冰草
    麻黄山1 84837.270°N, 107.023°E温性荒漠草原猪毛蒿
    Table 1. Types and geographical positions of the grassland communities
    仪器名称光谱范围/nm采集时间天气要求光谱采集时间/ms测定要求
    FieldSpec® 4 Hi-ResASD300~2 50010:00—14:00干燥、 无风、 晴朗无云或少云100测量后及时进行白板校正
    Table 2. Instrument parameters and requirements
    序号植物种拉丁名序号植物种拉丁名
    1二色补血草白花丹科Limonium bicolor Kuntze17猪毛蒿菊科Artemisia scoparia Waldst.
    2沙葱百合科Allium mongolicum Regel18丝叶小苦荬菊科Ixeridium graminifolium Tzvel.
    3百里香唇形科Thymus mongolicus Ronn.19黑沙蒿菊科Artemisia ordosica Krasch.
    4乳浆大戟大戟科Euphorbia esula Linn.20风毛菊菊科Saussurea japonica DC.
    5地构叶大戟科Speranskia uberculate Baill.21白莲蒿菊科Artemisia sacrorum Ledeb.
    6沙打旺豆科Astragalus adsurgens Pall.22阿尔泰狗娃花菊科Heteropappus altaicus Novopokr.
    7甘草豆科Glycyrrhiza uralensis Fisch.23虫实藜科Corispermum mongolicum Iljin
    8达乌里胡枝子豆科Lespedeza davurica Schindl.24北方獐牙菜龙胆科Swertia diluta Benth.
    9草木樨状黄芪豆科Astragalus melilotoides Pall.25老瓜头萝藦科Cynanchum komarovii Al.
    10隐子草禾本科Cleistogenes squarrosa Keng26地锦葡萄科Parthenocissus tricuspidata Planch.
    11无芒稗禾本科Echinochloa crusgalli var. mitis27二裂委陵菜蔷薇科Potentilla bifurca Linn.
    12蒙古冰草禾本科Agropyron mongolicum Keng28鹅绒委陵菜蔷薇科Potentilla anserina L.
    13虎尾草禾本科Chloris virgata Sw.29泡泡草茄科Physalis alkekengi L.
    14狗尾草禾本科Setaria viridis Beauv.30沙茴香伞形科Ferula bungeana Kitagawa
    15大针茅禾本科Stipa grandis P. Smirn.31牵牛花旋花科Pharbitis nil Ching
    16蒺藜蒺藜科Tribulus terrester Linn.32北芸香芸香科Evodia lepta Merr.
    Table 3. Plant directory
    植被指数名称计算公式
    NDVI705[8]归一化植被指数705(R750-R705)/(R750+R705)
    GNDVI[8]绿通道植被指数(R800-R550)/(R800+R550)
    PRI[8]光化学植被指数(R531-R570)/(R531+R570)
    OSAVI[8]土壤调节植被指数1.16((R800-R670)/(R800+R670+0.16))
    VARI[9]可视化气压阻抗指数(R555-R680)/(R555+R680-R480)
    PSRI[9]植被衰减指数(R680-R500)/R750
    NDWI[9]归一化水指数(R857-R1 241)/(R857+R1 241)
    Table 4. Spectral Index
    gammacosterrordispersion
    1×10-610
    100
    0.877 8
    0.877 8
    0.052 4
    0.052 4
    1×10-510
    100
    0.877 8
    0.820 3
    0.0524
    0.0679
    1×10-410
    100
    0.677 5
    0.184 0
    0.066 5
    0.072 1
    1×10-310
    100
    0.877 8
    0.877 8
    0.052 4
    0.052 4
    1×10-210
    100
    0.820 3
    0.701 0
    0.071 1
    0.067 9
    1×10-110
    100
    0.287 9
    0.046 4
    0.086 1
    0.039 7
    Table 5. Gamma and cost
    Tong JI, Bo WANG, Jun-ying YANG, Qiang LI, Guo-xing HE, Dong-rong PAN, Xiao-ni LIU. Spectral Characteristic Analysis and Spectral Identification of Desert Plants in Yanchi, Ningxia[J]. Spectroscopy and Spectral Analysis, 2022, 42(3): 678
    Download Citation