Calculating Leaf Area Index Using Neural Network and WorldView 3 Multispectral Imagery

Leaf Area Index (LAI) holds significant importance as a specific characteristic of Leaf Areas in the field of smart agriculture. This study explores the estimation of LAI using a multi-spectral image from WorldView 3 satellite. The image combines 8 VNIR bands and has a spatial resolution of 1.24m. T...

Full description

Saved in:
Bibliographic Details
Published in2024 59th International Scientific Conference on Information, Communication and Energy Systems and Technologies (ICEST) pp. 1 - 4
Main Authors Polimenov, Ventsislav, Ivanova, Krassimira, Tsvetkova, Mihaela, Anastasova, Elena, Dimitrova, Katya
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.07.2024
Subjects
image processing
Indexes
machine learning
Neural networks
remote sensing
Satellite images
Satellites
Smart agriculture
Spatial resolution
Training
Online AccessGet full text

Cover

More Information
Summary:Leaf Area Index (LAI) holds significant importance as a specific characteristic of Leaf Areas in the field of smart agriculture. This study explores the estimation of LAI using a multi-spectral image from WorldView 3 satellite. The image combines 8 VNIR bands and has a spatial resolution of 1.24m. To overcome the limited amount of available data, the image was split into smaller subsets called paxels, resulting in 500 paxels for training and testing. For enhancing machine learning models' performance, the standardisation of a dataset is made, after that, a Multilayer Perceptron with a specific architecture aimed to predict LAI from the multiple bands is trained. The achieved results showed promising performance in LAI prediction. Overall, the study demonstrates the potential of using satellite imagery and machine learning algorithms to improve our understanding of crop health and productivity.
ISSN:2603-3267
DOI:10.1109/ICEST62335.2024.10639753