Deep Learning-Based Prediction of Whiting (Merlangius merlangus) Length, Weight, and Catch Amount Using Satellite and Meteorological Data Uydu ve Meteorolojik Verilerle Mezgit (Merlangius merlangus) Boyu, Aǧirliǧi ve Av Miktarinin Derin Öǧrenme Tabanli Tahmini
34th Signal Processing and Communications Applications Conference, SIU 2026, İstanbul, Türkiye, 7 - 10 Temmuz 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/siu71813.2026.11636626
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: Deep Learning, Meteorological Time Series, Multimodal Data Fusion, Remote Sensing, Whiting (Merlangius merlangus)
- Ankara Yıldırım Beyazıt Üniversitesi Adresli: Evet
Özet
Satellite and meteorological data are increasingly used resources in marine and fisheries management. Combined with deep learning methods, these data have been utilized to predict the length, weight, and catch amount of whiting, a heavily caught species in the Black Sea. Within the scope of the study, Landsat 8/9 and Copernicus Sentinel-2 satellite images along with meteorological time series such as sea surface temperature, wind, precipitation, and wind direction were integrated to construct a multidimensional dataset. This dataset was analyzed using an LSTM and a CNN+LSTM hybrid model. Results revealed that while the LSTM model showed limited performance in predicting length, weight, and catch amount, the CNN+LSTM model achieved higher accuracy in all three predictions when satellite imagery were included. The findings demonstrate that incorporating spatial information together with meteorological time series improves prediction performance in fisheries data modeling.