Share Email Print
cover

Proceedings Paper

Identification of Phragmites australis and Spartina alterniflora in the Yangtze Estuary between Bayes and BP neural network using hyper-spectral data
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The aim of this work was to identify the coastal wetland plants between Bayes and BP neural network using hyperspectral data in order to optimize the classification method. For this purpose, we chose two dominant plants (invasive S. alterniflora and native P. australis) in the Yangtze Estuary, the leaf spectral reflectance of P. australis and S. alterniflora were measured by ASD field spectral machine. We tested the Bayes method and BP neural network for the identification of these two species. Results showed that three different bands (i.e., 555 nm,711 nm and 920 nm) could be identified as the sensitive bands for the input parameters for the two methods. Bayes method and BP neural network prediction model both performed well (Bayes prediction for 88.57% accuracy, BP neural network model prediction for about 80% accuracy), but Bayes theorem method could give higher accuracy and stability.

Paper Details

Date Published: 19 September 2016
PDF: 8 pages
Proc. SPIE 9975, Remote Sensing and Modeling of Ecosystems for Sustainability XIII, 99750H (19 September 2016); doi: 10.1117/12.2236772
Show Author Affiliations
Pudong Liu, East China Normal Univ. (China)
Colorado State Univ. (United States)
Jiayuan Zhou, East China Normal Univ. (China)
Runhe Shi, East China Normal Univ. (China)
Colorado State Univ. (United States)
Chao Zhang, East China Normal Univ. (China)
Chaoshun Liu, East China Normal Univ. (China)
Colorado State Univ. (United States)
Zhibin Sun, Colorado State Univ. (United States)
Wei Gao, East China Normal Univ. (China)
Colorado State Univ. (United States)


Published in SPIE Proceedings Vol. 9975:
Remote Sensing and Modeling of Ecosystems for Sustainability XIII
Wei Gao; Ni-Bin Chang, Editor(s)

© SPIE. Terms of Use
Back to Top