| Autor: |
Lestari, Anugrah Indah, Kushardono, Dony, Vetrita, Yenni, Santoso, Imam, Kartika, Tatik, Prasasti, Indah |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2023, Vol. 2941 Issue 1, p1-10, 10p |
| Abstrakt: |
Devegetation monitoring is important to maintain the hydrological cycle and ensure the availability of clean water. Rapid monitoring needs to be carried out since land cover can change quickly. Remote sensing can be used as a technology to monitor devegetation since it has multitemporal data and can cover a large area. This study intends to investigate a Synthetic Aperture Radar (SAR) remote sensing data feature extraction method for devegetation detection using a one-dimensional Convolutional Neural Network (CNN) located on Sumba Island. This study examines two schemes: the first one uses sigma (σ0) and gamma nought (γ0) on VH and VV polarizations with speckle filtering and the other uses Gray Level Co-occurrence Matrix (GLCM) of sigma (σ0) and gamma nought (γ0) on VH and VV polarizations with no speckle filtering as an input. With the Satellite pour l'Observation de la Terre (SPOT) images as reference data, the result shows that Sentinel-1 SAR multitemporal data detected devegetation with the highest overall accuracy of 81.57%, using 11x11 GLCM texture features on a combination of VH and VV polarization serving as input. Misclassification is mainly found in steep terrain or hilly areas. However, for large-scale monitoring, this rapid method has been promising. This approach can also support the National Research and Innovation Agency in developing a national radar satellite program. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
