Araştırma Makalesi
BibTex RIS Kaynak Göster

Investigation of the Effect of Training Dataset Size on the Detection of Mucilage Formation in the Marmara Sea Using Convolutional Neural Networks

Yıl 2025, Sayı: 8 ÇEVRE, 121 - 130, 05.02.2025

Öz

Developing monitoring and observation systems plays a critical role in identifying measures to prevent adverse changes in marine environments, which hold great importance for ecosystems. In this context, the use of remotely sensed images, a product of advancing sensor technology, offers significant advantages for analyzing environmental changes. Additionally, employing deep learning techniques enhances the performance of analyses conducted on these complex and large-scale remotely sensed datasets. This study aims to detect and map mucilage formations observed in the Gulf of İzmit, located in the eastern Marmara Sea, in 2021 using satellite imagery and deep learning models. Atmospherically corrected Sentinel-2 (Level-2) images were used as the primary data source, and classifications were performed using pixel-based deep learning models based on Convolutional Neural Network (CNN) architectures.

In this context, the classification performance of CNN models was thoroughly evaluated in relation to the size of the training dataset used in model development. Following the resampling of Sentinel-2 satellite images and the masking of land pixels, sample pixels representing clean water surfaces and mucilage formations within the study area were collected. The dataset size, composed of these sample pixels, was determined based on the sampling ratios of ground truth datasets to examine the impact of dataset size on mucilage detection during the training and testing stages of the deep learning models.

The results demonstrated that increasing the number of samples used in model training improved the predictive performance of the models by approximately 7%. Moreover, extensive mucilage formations covering an area of 4 km² in the study area were identified on May 19, 2021. Quantitative evaluations indicated that larger datasets contribute positively to the classification accuracy of deep learning models.

Kaynakça

  • Acar, U., Yılmaz, O. S., Çelen, M., Ateş, A. M., Gülgen, F., & Şanlı, F. B. (2021). Determination of Mucilage in the Sea Of Marmara Using Remote Sensing Techniques with Google Earth Engine. International Journal of Environment and Geoinformatics, 8(4), 423-434.
  • Colkesen, I., Kavzoglu, T., Sefercik, U. G., & Ozturk, M. Y. (2023). Automated Mucilage Extraction İndex (AMEI): A Novel Spectral Water Index for Identifying Marine Mucilage Formations From Sentinel-2 Imagery. International Journal of Remote Sensing, 44(1), 105-141.
  • Colkesen, I., Ozturk, M. Y., & Altuntas, O. Y. (2024). Comparative Evaluation of Performances of Algae Indices, Pixel-And Object- Based Machine Learning Algorithms in Mapping Floating Algal Blooms Using Sentinel-2 Imagery. Stochastic Environmental Research and Risk Assessment, 38(4), 1613-1634.
  • Cozzi, S., Ivancic, I., Catalano, G., Djakovac, T., & Degobbis, D. (2004). Dynamics of the Oceanography Properties during Mucilage Appearance in the Northern Adriatic Sea: Analysis of the 1977 event in comparison to earlier events. Journal of Marine Systems, 50, 223–241. https://doi.org/10.1016/j.jmarsys.2004.01.007
  • Esi, Ç., Ertürk, A., & Erten, E. (2024). Nonnegative Matrix Factorization-Based Environmental Monitoring of Marine Mucilage. International Journal of Remote Sensing, 45(11), 3764-3788.
  • Hinton, G. E., and R. R. Salakhutdinov. (2006). Reducing the Dimensionality of Data with Neural Networks. Science 313 (5786): 504–507. doi:10.1126/science.1127647.
  • Hou, Y., Liu, Z., Zhang, T., & Li, Y. (2021). C-UNet: Complement UNet for Remote Sensing Road Extraction. Sensors, 21(6), 2153.
  • Hu, K., Zhang, D., & Xia, M. (2021). CDUNet: Cloud Detection UNet for Remote Sensing Imagery. Remote Sensing, 13(22), 4533.
  • Kattenborn, T., Leitloff, J., Schiefer, F., & Hinz, S. (2021). Review on Convolutional Neural Networks (CNN) in Vegetation Remote Sensing. ISPRS Journal of Photogrammetry and Remote Sensing, 173, 24-49.
  • Kaushal, A., Gupta, A. K., & Sehgal, V. K. (2024). A Semantic Segmentation Framework with UNet-Pyramid for Landslide Prediction using Remote Sensing Data. Scientific Reports, 14(1), 1-23.
  • Kavzoglu, T., & Goral, M. (2022). Google Earth Engine for Monitoring Marine Mucilage: Izmit Bay in Spring 2021. Hydrology, 9(8), 135.
  • Kavzoğlu, T., Tonbul, H., Çölkesen, İ., & Sefercik, U. G. (2021). The Use of Object-Based Image Analysis for Monitoring 2021 Marine Mucilage Bloom in the Sea of Marmara. International Journal of Environment and Geoinformatics, 8(4), 529-536.
  • Mishra, V. K., & Mishra, A. K. (2024). Mapping Phytoplankton and Algal Blooms with a Novel Multi Sensor Water Index (MSWI). IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.
  • Özalp, H. B. (2021). First massive mucilage event observed in deep waters of Çanakkale Strait (Dardanelles), Turkey. Journal of the Black Sea/mediterranean Environment, 27(1), 49–66.
  • Schiaparelli, S., Castellano, M., Povero, P., Sartoni, G., Cattaneo- Vietti, R. (2007). A Benthic Mucilage event in North- Western Mediterranean Sea and its Possible Relationships with the Summer 2003 European Heatwave: Short Term Effects on Littoral Rocky Assemblages. Marine Ecology, 28(3), 341– 353.
  • Sefercik, U. G., Colkesen, I., Kavzoglu, T., Ozdogan, N., & Ozturk, M. Y. (2024). Assessing the Physical and Chemical Characteristics of Marine Mucilage Utilizing In-Situ and Remote Sensing Data (Sentinel-1,-2,-3). PFG–Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 92(4), 415-430.
  • Sun, D., Gao, G., Huang, L., Liu, Y., & Liu, D. (2024). Extraction of Water Bodies from High-Resolution Remote Sensing Imagery Based on a Deep Semantic Segmentation Network. Scientific Reports, 14(1), 14604.
  • Taş, S., Kuş, D., & Yılmaz, I. N. (2020). Temporal variations in Phytoplankton Composition in the North-Eastern Sea of Marmara: Potentially Toxic Species and Mucilage Event. Mediterranean Marine Science, 21(3), 668–683. https://doi.org/10.12681/mms.22562
  • Wang, Q., Zhang, X., Chen, G., Dai, F., Gong, Y., & Zhu, K. (2018). Change Detection Based on Faster R-CNN for High-Resolution Remote Sensing Images. Remote sensing letters, 9(10), 923-932.
  • Wasehun, E. T., Hashemi Beni, L., & Di Vittorio, C. A. (2024). UAV and Satellite Remote Sensing for Inland Water Quality Assessments: A Literature Review. Environmental Monitoring and Assessment, 196(3), 277.
  • Xue, H., Liu, K., Wang, Y., Chen, Y., Huang, C., Wang, P., & Li, L. (2024). MAD-UNet: A Multi-Region UAV Remote Sensing Network for Rural Building Extraction. Sensors, 24(8), 2393.
  • Yagci, A. L., Colkesen, I., Kavzoglu, T., & Sefercik, U. G. (2022). Daily Monitoring of Marine Mucilage Using the MODIS Products: A Case Study of 2021 Mucilage Bloom in the Sea of Marmara, Turkey. Environmental Monitoring and Assessment, 194(3), 170.
  • Yilmaz, E. O., Tonbul, H., & Kavzoglu, T. (2024). Marine Mucilage Mapping with Explained Deep Learning Model Using Water- Related Spectral Indices: A Case Study of Dardanelles Strait, Turkey. Stochastic Environmental Research and Risk Assessment, 38(1), 51-68.
  • Zhang, W., Tang, P., & Zhao, L. (2019). Remote Sensing Image Scene Classification Using CNN-CapsNet. Remote Sensing, 11(5), 494.
  • Xu, H. (2006). Modification of Normalised Difference Water Index (NDWI) to Enhance Open Water Features In Remotely Sensed Imagery. International Journal of Remote Sensing, 27(14), 3025-3033.
  • TÜİK (Türkiye İstatistik Kurumu). (2023). 2023 Adrese Dayalı Nüfus Kayıt Sistemi Sonuçları. https://data.tuik.gov.tr/

Evrişimsel Sinir Ağlarıyla Marmara Denizi'ndeki Müsilaj Oluşumlarının Tespitinde Eğitim Veri Seti Boyutunun Etkisinin İncelenmesi

Yıl 2025, Sayı: 8 ÇEVRE, 121 - 130, 05.02.2025

Öz

Ekosistem için büyük önem arz eden denizel ortamlarda meydana gelen olumsuz değişimlerin önüne geçilebilmesi için alınacak önemlerin belirlenmesinde takip ve izleme sistemlerinin geliştirilmesi kritik bir rol oynamaktadır. Bu noktada her geçen gün ilerleyen sensör teknolojisinin ürünü olan uzaktan algılanmış görüntülerin kullanılması çevresel değişim analizlerde önemli avantajlar sunmaktadır. Bunun yanı sıra, derin öğrenme tekniklerinin kullanılması da bu karmaşık ve büyük hacimli uzaktan algılanmış veri setlerinden daha yüksek performans alınmasını sağlamaktadır. Bu çalışmada, 2021 yılında Marmara Denizi’nin doğusunda yer alan Kocaeli Körfezi’nde gözlemlenen müsilaj oluşumlarının uydu görüntüleri ve derin öğrenme modelleri ile tespiti ve haritalandırılması hedeflenmiştir. Çalışmada atmosferik olarak düzeltilmiş Sentinel-2 (Seviye-2) görüntüleri temel veri kaynağı olarak kullanılmış ve piksel tabanlı Evrişimsel Sinir Ağları (ESA) mimarisine dayalı derin öğrenme modelleriyle sınıflandırma yapılmıştır. Bu kapsamda, ESA modellerinin sınıflandırma performansı, model oluşumunda kullanılan eğitim veri setinin boyutuyla ilişkili olarak detaylı bir şekilde ele alınmıştır. Sentinel-2 uydu görüntülerinin yeniden örneklenmesi ve kara piksellerinin maskelenmesinin ardından çalışma alanında yer alan temiz su yüzeyi ve müsilaj oluşumlarına yönelik örnek pikseller toplanmıştır. Derin öğrenme modellerinin eğitim ve test aşamasında kullanılmak üzere elde edilen örnek piksellerden oluşan veri seti boyutunun müsilaj oluşumlarının tespitine etkisinin incelenmesi amacı ile yer doğrulama veri seti örnekleme oranları dikkate alınarak belirlenmiştir. Çalışma sonuçları, model oluşumunda kullanılan örnek sayısının artması ile birlikte tahmin modelinin performansında %7 seviyelerinde artış olduğunu göstermektedir. Bununla birlikte 19 Mayıs 2021 tarihinde çalışma alanında 4km2’lik bir alana yayılan yoğun müsilaj oluşumları tespit edilmiştir. Elde edilen nicel değerlendirmeler sonucunda daha geniş veri setlerinin derin öğrenme modellerinin sınıflandırma doğruluğuna katkı sağladığı görülmüştür.

Kaynakça

  • Acar, U., Yılmaz, O. S., Çelen, M., Ateş, A. M., Gülgen, F., & Şanlı, F. B. (2021). Determination of Mucilage in the Sea Of Marmara Using Remote Sensing Techniques with Google Earth Engine. International Journal of Environment and Geoinformatics, 8(4), 423-434.
  • Colkesen, I., Kavzoglu, T., Sefercik, U. G., & Ozturk, M. Y. (2023). Automated Mucilage Extraction İndex (AMEI): A Novel Spectral Water Index for Identifying Marine Mucilage Formations From Sentinel-2 Imagery. International Journal of Remote Sensing, 44(1), 105-141.
  • Colkesen, I., Ozturk, M. Y., & Altuntas, O. Y. (2024). Comparative Evaluation of Performances of Algae Indices, Pixel-And Object- Based Machine Learning Algorithms in Mapping Floating Algal Blooms Using Sentinel-2 Imagery. Stochastic Environmental Research and Risk Assessment, 38(4), 1613-1634.
  • Cozzi, S., Ivancic, I., Catalano, G., Djakovac, T., & Degobbis, D. (2004). Dynamics of the Oceanography Properties during Mucilage Appearance in the Northern Adriatic Sea: Analysis of the 1977 event in comparison to earlier events. Journal of Marine Systems, 50, 223–241. https://doi.org/10.1016/j.jmarsys.2004.01.007
  • Esi, Ç., Ertürk, A., & Erten, E. (2024). Nonnegative Matrix Factorization-Based Environmental Monitoring of Marine Mucilage. International Journal of Remote Sensing, 45(11), 3764-3788.
  • Hinton, G. E., and R. R. Salakhutdinov. (2006). Reducing the Dimensionality of Data with Neural Networks. Science 313 (5786): 504–507. doi:10.1126/science.1127647.
  • Hou, Y., Liu, Z., Zhang, T., & Li, Y. (2021). C-UNet: Complement UNet for Remote Sensing Road Extraction. Sensors, 21(6), 2153.
  • Hu, K., Zhang, D., & Xia, M. (2021). CDUNet: Cloud Detection UNet for Remote Sensing Imagery. Remote Sensing, 13(22), 4533.
  • Kattenborn, T., Leitloff, J., Schiefer, F., & Hinz, S. (2021). Review on Convolutional Neural Networks (CNN) in Vegetation Remote Sensing. ISPRS Journal of Photogrammetry and Remote Sensing, 173, 24-49.
  • Kaushal, A., Gupta, A. K., & Sehgal, V. K. (2024). A Semantic Segmentation Framework with UNet-Pyramid for Landslide Prediction using Remote Sensing Data. Scientific Reports, 14(1), 1-23.
  • Kavzoglu, T., & Goral, M. (2022). Google Earth Engine for Monitoring Marine Mucilage: Izmit Bay in Spring 2021. Hydrology, 9(8), 135.
  • Kavzoğlu, T., Tonbul, H., Çölkesen, İ., & Sefercik, U. G. (2021). The Use of Object-Based Image Analysis for Monitoring 2021 Marine Mucilage Bloom in the Sea of Marmara. International Journal of Environment and Geoinformatics, 8(4), 529-536.
  • Mishra, V. K., & Mishra, A. K. (2024). Mapping Phytoplankton and Algal Blooms with a Novel Multi Sensor Water Index (MSWI). IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.
  • Özalp, H. B. (2021). First massive mucilage event observed in deep waters of Çanakkale Strait (Dardanelles), Turkey. Journal of the Black Sea/mediterranean Environment, 27(1), 49–66.
  • Schiaparelli, S., Castellano, M., Povero, P., Sartoni, G., Cattaneo- Vietti, R. (2007). A Benthic Mucilage event in North- Western Mediterranean Sea and its Possible Relationships with the Summer 2003 European Heatwave: Short Term Effects on Littoral Rocky Assemblages. Marine Ecology, 28(3), 341– 353.
  • Sefercik, U. G., Colkesen, I., Kavzoglu, T., Ozdogan, N., & Ozturk, M. Y. (2024). Assessing the Physical and Chemical Characteristics of Marine Mucilage Utilizing In-Situ and Remote Sensing Data (Sentinel-1,-2,-3). PFG–Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 92(4), 415-430.
  • Sun, D., Gao, G., Huang, L., Liu, Y., & Liu, D. (2024). Extraction of Water Bodies from High-Resolution Remote Sensing Imagery Based on a Deep Semantic Segmentation Network. Scientific Reports, 14(1), 14604.
  • Taş, S., Kuş, D., & Yılmaz, I. N. (2020). Temporal variations in Phytoplankton Composition in the North-Eastern Sea of Marmara: Potentially Toxic Species and Mucilage Event. Mediterranean Marine Science, 21(3), 668–683. https://doi.org/10.12681/mms.22562
  • Wang, Q., Zhang, X., Chen, G., Dai, F., Gong, Y., & Zhu, K. (2018). Change Detection Based on Faster R-CNN for High-Resolution Remote Sensing Images. Remote sensing letters, 9(10), 923-932.
  • Wasehun, E. T., Hashemi Beni, L., & Di Vittorio, C. A. (2024). UAV and Satellite Remote Sensing for Inland Water Quality Assessments: A Literature Review. Environmental Monitoring and Assessment, 196(3), 277.
  • Xue, H., Liu, K., Wang, Y., Chen, Y., Huang, C., Wang, P., & Li, L. (2024). MAD-UNet: A Multi-Region UAV Remote Sensing Network for Rural Building Extraction. Sensors, 24(8), 2393.
  • Yagci, A. L., Colkesen, I., Kavzoglu, T., & Sefercik, U. G. (2022). Daily Monitoring of Marine Mucilage Using the MODIS Products: A Case Study of 2021 Mucilage Bloom in the Sea of Marmara, Turkey. Environmental Monitoring and Assessment, 194(3), 170.
  • Yilmaz, E. O., Tonbul, H., & Kavzoglu, T. (2024). Marine Mucilage Mapping with Explained Deep Learning Model Using Water- Related Spectral Indices: A Case Study of Dardanelles Strait, Turkey. Stochastic Environmental Research and Risk Assessment, 38(1), 51-68.
  • Zhang, W., Tang, P., & Zhao, L. (2019). Remote Sensing Image Scene Classification Using CNN-CapsNet. Remote Sensing, 11(5), 494.
  • Xu, H. (2006). Modification of Normalised Difference Water Index (NDWI) to Enhance Open Water Features In Remotely Sensed Imagery. International Journal of Remote Sensing, 27(14), 3025-3033.
  • TÜİK (Türkiye İstatistik Kurumu). (2023). 2023 Adrese Dayalı Nüfus Kayıt Sistemi Sonuçları. https://data.tuik.gov.tr/
Toplam 26 adet kaynakça vardır.

Ayrıntılar

Birincil Dil Türkçe
Konular Çevre Yönetimi (Diğer)
Bölüm Araştırma Makalesi
Yazarlar

Meltem Çelen 0000-0001-9487-497X

Mehmet Salim Öncel

Mustafacan Saygı 0009-0003-6015-1273

İsmail Çölkesen 0000-0001-9670-3023

Yayımlanma Tarihi 5 Şubat 2025
Gönderilme Tarihi 16 Aralık 2024
Kabul Tarihi 4 Şubat 2025
Yayımlandığı Sayı Yıl 2025 Sayı: 8 ÇEVRE

Kaynak Göster

APA Çelen, M., Öncel, M. S., Saygı, M., Çölkesen, İ. (2025). Evrişimsel Sinir Ağlarıyla Marmara Denizi’ndeki Müsilaj Oluşumlarının Tespitinde Eğitim Veri Seti Boyutunun Etkisinin İncelenmesi. Şura Akademi(8 ÇEVRE), 121-130.