| Autor: |
Nuaimi, Maitha A. A., Salem, Eman H., Ibrahim, Hala F. S., Ali, Asma Abdulla A. Al, Abdulmuttalib, Hussein M., Biswas, Raja, Ha, Le Hai, Banerjee, Sandip, Mittal, Abhay S. |
| Předmět: |
|
| Zdroj: |
ISPRS Annals of Photogrammetry, Remote Sensing & Spatial Information Sciences; 2024, Vol. 10 Issue 4/W4, p137-143, 7p |
| Abstrakt: |
Accurate detection and monitoring of urban changes are crucial for sustainable urban planning and management, sustaining economic growth, as well as advancing smart city initiatives. Due to the unprecedented urbanization and rapid population growth in the emirates of Dubai over the last few decades, it is essential to closely monitor and detect changes in urban land cover. While traditional classifiers using low-medium resolution open-source satellite images have shown success in the broad-level classification of land use and land cover concerning urbanization, they are time-consuming, labour-intensive, and limited to minute-level detailed urban change detection. Thus, the present study significantly focuses on precise change detection of various urban classes and associated vegetation cover during 2021–2022 using high-resolution KhalifaSAT images through the application of a deep learning algorithm. The study implemented an advanced machine learning tool (available in the Geo-AI platform, Eofactory) to generate a super-resolution image. Also, after performing various pre-processing of the image, a deep learning-based U2-Net model was implemented for the change detection. The model has efficiently detected changes in different urban features such as newly constructed buildings, rooftop changes, demolished buildings, and building extensions as well as changes in associated vegetation with an overall accuracy of 0.82 and, a kappa coefficient of 0.76. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
