TY - GEN
T1 - Band selection in sentinel-2 satellite for agriculture applications
AU - Zhang, Tianxiang
AU - Su, Jinya
AU - Liu, Cunjia
AU - Chen, Wen Hua
AU - Liu, Hui
AU - Liu, Guohai
N1 - Publisher Copyright:
© 2017 Chinese Automation and Computing Society in the UK - CACSUK.
PY - 2017/10/23
Y1 - 2017/10/23
N2 - Various indices are used for assessing vegetation and soil properties in satellite remote sensing applications. Some indices, such as NDVI and NDWI, are defined based on the sensitivity and significance of specific bands. Nowadays, remote sensing capability with a good number of bands and high spatial resolution is available. Instead of classification based on indices, this paper explores direct classification using selected bands. Recently launched Sentinel-2A is adopted as a case study. Three methods are compared, where the first approach utilizes traditional indices and the latter two approaches adopt specific bands (Red, NIR, and SWIR) and full bands of on-board sensors, respectively. It is shown that a better classification performance can be achieved by directly using the three selected bands compared with the one using indices, while the use of all 13 bands can further improve the performance. Therefore, it is recommended the new approach can be applied for Sentinel-2A image analysis and other wide applications.
AB - Various indices are used for assessing vegetation and soil properties in satellite remote sensing applications. Some indices, such as NDVI and NDWI, are defined based on the sensitivity and significance of specific bands. Nowadays, remote sensing capability with a good number of bands and high spatial resolution is available. Instead of classification based on indices, this paper explores direct classification using selected bands. Recently launched Sentinel-2A is adopted as a case study. Three methods are compared, where the first approach utilizes traditional indices and the latter two approaches adopt specific bands (Red, NIR, and SWIR) and full bands of on-board sensors, respectively. It is shown that a better classification performance can be achieved by directly using the three selected bands compared with the one using indices, while the use of all 13 bands can further improve the performance. Therefore, it is recommended the new approach can be applied for Sentinel-2A image analysis and other wide applications.
KW - Agriculture
KW - Machine learning
KW - Remote sensing
KW - Sentinel-2A
KW - Supervised classification
UR - http://www.scopus.com/inward/record.url?scp=85039980246&partnerID=8YFLogxK
U2 - 10.23919/IConAC.2017.8081990
DO - 10.23919/IConAC.2017.8081990
M3 - Conference article published in proceeding or book
AN - SCOPUS:85039980246
T3 - ICAC 2017 - 2017 23rd IEEE International Conference on Automation and Computing: Addressing Global Challenges through Automation and Computing
BT - ICAC 2017 - 2017 23rd IEEE International Conference on Automation and Computing
A2 - Zhang, Jie
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 23rd IEEE International Conference on Automation and Computing, ICAC 2017
Y2 - 7 September 2017 through 8 September 2017
ER -