EUVIMEDEUVIMED entdecken
Crossref

Advanced Tropical Cyclone Intensity Estimation Using EfficientNetV2S with Spatiotemporal Geospatial Fusion and Transfer Learning

Jaya Vardhan Reddy Koppula, Janjanam Lavanya, Peserlanka Devendra Babu

International Journal of Innovative Science and Research Technology · 09.04.2026 · 3698

DOI: 10.38124/ijisrt/26mar2136 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Innovative Science and Research Technology, online erschienen am 09.04.2026. Verfasst von Jaya Vardhan Reddy Koppula und 2 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Tropical cyclones are the most destructive of natural catastrophes, since it has cost money and lives of many people all around the globe. The cyclone intensity parameters such as the maximum sustained wind speed (Vmax) as well as the minimum sea-level pressure (MSLP) play a very critical role in disaster preparedness and mitigation. The multi-model deep learning system proposed in this paper uses satellite and meteorological metadata to predict the strength of the cyclone. The model utilizes an already trained EfficientNetV2S model as a spatial feature extractor of multi-channel satellite image and a fully connected metadata branch that consists of geographical and time information. It is performed on the fused representation with multi-output regression (Vmax and MSLP both). Tropical Cyclone Intensity Regression (TCIR) data has been used to train and test the model. The capability to predict Vmax and MSLP with an error of 2.43 knots and 2.87 hPa has been shown to be good and the correlation coefficients of Vmax and MSLP have been shown to be good (0.992 and 0.977 respectively). The proposed methodology can make more predictions and shows the success of using visual and contextual information on the prediction of cyclone intensity union.

Zitieren

APA

Koppula, J. V. R., Lavanya, J., & Babu, P. D. (2026). Advanced Tropical Cyclone Intensity Estimation Using EfficientNetV2S with Spatiotemporal Geospatial Fusion and Transfer Learning. International Journal of Innovative Science and Research Technology, 3698. https://doi.org/10.38124/ijisrt/26mar2136

Vancouver

Koppula JVR, Lavanya J, Babu PD. Advanced Tropical Cyclone Intensity Estimation Using EfficientNetV2S with Spatiotemporal Geospatial Fusion and Transfer Learning. International Journal of Innovative Science and Research Technology. 2026:3698. doi:10.38124/ijisrt/26mar2136

BibTeX
@article{Koppula2026,
  author = {Koppula, Jaya Vardhan Reddy and Lavanya, Janjanam and Babu, Peserlanka Devendra},
  title = {Advanced Tropical Cyclone Intensity Estimation Using EfficientNetV2S with Spatiotemporal Geospatial Fusion and Transfer Learning},
  journal = {International Journal of Innovative Science and Research Technology},
  year = {2026},
  pages = {3698},
  doi = {10.38124/ijisrt/26mar2136}
}
Kostenlos starten →