Classification of major species in the sericite–Artemisia desert grassland using hyperspectral images and spectral feature identification

• DOI: 10.7717/peerj.17663
• Publication Date: 2024-07-18T07:28:01Z
• Authors: Wenhao Liu; Wanqiang Han; Guili Jin; Ke Gong; Jian Ma
• Keywords: China; QH301-705.5; R; Discriminant Analysis; Hyperspectral Imaging; Grassland; Identification parameters; Artemisia; Grassland plants; Remote Sensing Technology; Screening; Spectral characteristics; Medicine; Fisher discriminant; Seasons; Biology (General); Desert Climate; Agricultural Science

Background The species composition of and changes in grassland communities are important indices for inferring the number, quality community succession grasslands, accurate monitoring is foundation evaluating, protecting, utilizing resources. Remote sensing technology provides a reliable powerful approach measuring regional terrain information, identification by remote will improve effectiveness monitoring. Methods Ground hyperspectral images sericite – Artemisia desert different seasons were obtained with Soc710 VP imaging spectrometer. First-order differential processing was used to calculate characteristic parameters. Analysis variance extract main species, namely, Seriphidium transiliense (Poljak), Ceratocarpus arenarius L., Petrosimonia sibirica (Pall), bare land spectral parameters vegetation seasons. On this basis, Fisher discriminant analysis divide samples into training set test at ratio 7:3. identify three plants land. Results selection significant differences ( P < 0.05) between recognition objects effectively distinguished features, also differed due growth period species. overall accuracy model established index decreased following order: June (98.87%) > September (91.53%) April (90.37%). feature (89.77%) (88.48%) (85.98%). Conclusions models based on months superior those parameters, accuracies ranging from 1.76% 9.40% higher. Based image data, use as can enable sericite–Artemisia grassland, providing basis further quantitative classification images.