617. Deep Learning-Based Satellite Image Classification for Accurate Deforestation Detection in Romania
Invited abstract in session TE-36: Analytics for Forest Fires Prevention and Risk Analysis, cluster Use of Analytics in Forest Fires Management.
Tuesday, 16:15-17:45Room: FENH309
Authors (first author is the speaker)
| 1. | Darie Moldovan
|
| Babes-Bolyai University | |
| 2. | Marian Lucian Cotolan
|
| Babes-Bolyai University | |
| 3. | Adina Tilea
|
| Babes-Bolyai University | |
| 4. | Bianca Nicoleta Marian
|
| Babes-Bolyai University |
Abstract
This research proposes a deep learning framework for monitoring and detecting deforestation in Romania, utilizing two distinct datasets consisting of satellite images - one featuring the Amazon basin and the other featuring four key regions of the Carpathian Mountains, collected by the Sentinel-2 satellite. To achieve accurate classification, the images were pre-processed and resized to serve as input for two classification models: a Convolutional Neural Network (CNN) model implemented using SAS Viya and a fine-tuned ResNet50 model developed in Python. The objective of this research is to predict whether deforestation has taken place, contributing to the preservation and protection of Romania's forests. The findings of this study can have significant implications for the development of effective strategies to combat deforestation and promote sustainable land use practices.
Keywords
- Deep Learning
- Applications, Climate Change
- Data Science
Status: accepted
Back to the list of papers