A REMOTE SENSING APPROACH TO IDENTIFY SHIPPING EMISSIONS

Öz

Shipping emissions are recognized as a significant contributor to climate change, increasing the frequency of extreme weather events. It is important to monitor transportation mobility for estimating emissions, and remote sensing data and methods are valuable due to their continuity, wide coverage, and ease of integration with other relevant data. Sentinel-2 data enables monitoring of transportation mobility and can be used as proxy data to infer shipping emissions. For this purpose, this study aims to monitor ship movements on a regional scale with YOLOX-assisted ship detection to estimate emissions. A new dataset was curated from Sentinel-2 RGB images provided by the European Space Agency’s Copernicus program, in which environmental and climate monitoring data are available free of charge. The dataset covers various locations of busy container ports to support data diversity. The YOLOX object detector, introduced to the computer vision community in 2021, is known for its versatility in detecting multi-scale objects. Thus, YOLOX was chosen to perform ship detection, and each standard YOLOX model (YOLOX-s-m-l-x) was trained on the dataset in separate sessions with default hyperparameters. As a result, the model with the highest evaluation score, YOLOX-l, was utilized to extract ship instances and localize ships from the dataset. Average precision (AP), average recall (AR), and F1-score were found to be 56.39%, 65.45%, and 60.48%, respectively. Ship extractions were utilized to estimate carbon monoxide (CO) emissions for 2023 in the port of Rotterdam, and a total of 148.504 tons of shipping-related CO emissions were inferred as a result.

Anahtar Kelimeler

• Shipping Emissions
• Remote Sensing
• Ship Detection
• Deep Learning.

DENİZ TAŞIMACILIĞI KAYNAKLI EMİSYONLARIN BELİRLENMESİNDE UZAKTAN ALGILAMA YAKLAŞIMI

Öz

Aşırı hava olaylarının görülme sıklığının artması ve iklim değişikliğinin önemli nedenlerinden birinin ulaştırma kaynaklı emisyonlar olduğu bilinmektedir. Emisyonların tahmini için ulaştırma hareketliliğinin izlenmesi gerekli olup, uzaktan algılama veri ve yöntemleri; sürekliliği geniş kapsama alanı ve ilgili diğer veriler ile bütünleştirilme kolaylığı nedeniyle öne çıkmaktadır. Bu çalışmada deniz taşımacılığı kaynaklı emisyonların tahmin edilmesi için bölgesel ölçekte gemi hareketlerinin derin öğrenme algoritmalarından YOLOX ile tespit edilmesi amaçlanmıştır. Sentinel-2 verisi ulaştırma hareketliliğinin izlenmesinde ve emisyon tahmininde yardımcı bir veri kaynağıdır. Bu sebeple, kara-deniz, çevre ve iklim izleme verileri araştırmacılara ücretsiz sunulmakta olduğu Avrupa Uzay Ajansı Copernicus programının bir parçası olan Sentinel-2 RGB görüntülerinden yeni bir veri seti oluşturulmuş olup, bu veri seti gemilerin otomatik çıkarımında kullanılmıştır. Veri çeşitliliği, model eğitiminde ve modelin genelleme yeteneğini artırmada önemli bir parametredir. Veri çeşitliliğini sağlamak ve modelin öğrenme kapasitesini artırmak için dünyanın farklı konteyner limanlarından alınan Sentinel-2 görüntüleri bir araya getirilmiş ve çeşitli ön işlem adımlarından geçirilmiştir. İlk kez 2021 yılında kullanıcılara tanıtılan YOLOX algoritması farklı büyüklükteki nesneleri tespit edebilme kapasitesine sahip bir algoritma olup, bu yönü ile önceki diğer YOLO mimarilerinden ayrılmaktadır. Bu sebeple, bu çalışma özelinde YOLOX algoritması tercih edilmiştir. Tespit doğruluğunu artırmak amacıyla, her bir standart YOLOX modeli (YOLOX-s-m-l-x) ayrı oturumlarda, hiperparametreler üzerinde değişiklik yapılmadan eğitilmiştir ve aralarından en yüksek metrik değerlere ulaşan YOLOX-l modeli ile gemiler tespit edilmiştir. Model ortalama kesinlik, ortalama duyarlılık ve F1 puanı sırasıyla %56.39, %65.45 ve %60.48 bulunmuştur. Gemi çıkarımları, Rotterdam limanında 2023 yılı deniz taşımacılığı kaynaklı karbonmonoksit (CO) emisyonlarının tahmininde kullanılmıştır ve sonuç olarak emisyon miktarı 148,504-ton olarak öngörülmüştür.

Anahtar Kelimeler

• Deniz Taşımacılığı Kaynaklı Emisyonlar
• Uzaktan Algılama
• Gemi Tespiti
• Derin Öğrenme.

Kaynakça

1. Alamoush, A. S., Ölçer, A. I. & Ballini, F. (2022). Ports’ role in shipping decarbonisation: A common port incentive scheme for shipping greenhouse gas emissions reduction. In Cleaner Logistics and Supply Chain (Vol. 3). Elsevier Ltd. https://doi.org/10.1016/j.clscn.2021.100021
2. Balasundram, A., Shaik, A., Banga, J. K. & Singh, A. K. (2024). U-Net inspired deep neural network-based smoke plume detection in satellite images. In Computers, Materials & Continua, 79(1), 779-799. https://doi.org/10.32604/cmc.2024.048362
3. Christodoulou, A. & Cullinane, K. (2024). The prospects for, and implications of, emissions trading in shipping. Maritime Economics and Logistics, 26(1), 168–184. https://doi.org/10.1057/s41278-023-00261-1
4. Couture, H. D., Alvara, M., Freeman, J., Davitt, A., Koenig, H., Rouzbeh Kargar, A., O’Connor, J., Söldner-Rembold, I., Ferreira, A., Jeyaratnam, J., Lewis, J., McCormick, C., Nakano, T., Dalisay, C., Lewis, C., Volpato, G., Gray, M. & McCormick, G. (2024). Estimating Carbon Dioxide Emissions from Power Plant Water Vapor Plumes Using Satellite Imagery and Machine Learning. Remote Sensing, 16(7). https://doi.org/10.3390/rs16071290
5. Deniz, C. & Durmuşoǧlu, Y. (2008). Estimating shipping emissions in the region of the Sea of Marmara, Turkey. Science of the Total Environment, 390(1), 255–261. https://doi.org/10.1016/j.scitotenv.2007.09.033
6. Drusch, M., del Bello, U., Carlier, S., Colin, O., Fernandez, V., Gascon, F., Hoersch, B., Isola, C., Laberinti, P., Martimort, P., Meygret, A., Spoto, F., Sy, O., Marchese, F. & Bargellini, P. (2012). Sentinel-2: ESA’s Optical High-Resolution Mission for GMES Operational Services. Remote Sensing of Environment, 120, 25–36. https://doi.org/10.1016/j.rse.2011.11.026
7. Er, M. J., Zhang, Y., Chen, J. & Gao, W. (2023). Ship detection with deep learning: a survey. Artificial Intelligence Review, 56(10), 11825–11865. https://doi.org/10.1007/s10462-023-10455-x
8. Everingham, M., van Gool, L., Williams, C. K. I., Winn, J. & Zisserman, A. (2010). The pascal visual object classes (VOC) challenge. International Journal of Computer Vision, 88(2), 303–338. https://doi.org/10.1007/s11263-009-0275-4

9. Fontelle, J.-P., Fridell, E., Grigoriadis, A., Hill, N., Kilde, N., Lavender, K., Mamarikas, S., Reynolds, G., Rypdal, K., Thomas, R., Webster, A. & Winther, M. (n.d.). EMEP/EEA air pollutant emission inventory guidebook 2023 1 Category Title NFR 1.A International maritime and inland navigation, national navigation, national fishing, recreational boats International maritime navigation, international inland navigation, national navigation (shipping), national fishing EMEP/EEA air pollutant emission inventory guidebook 2023 2.
10. Gascon, F., Bouzinac, C., Thépaut, O., Jung, M., Francesconi, B., Louis, J., Lonjou, V., Lafrance, B., Massera, S., Gaudel-Vacaresse, A., Languille, F., Alhammoud, B., Viallefont, F., Pflug, B., Bieniarz, J., Clerc, S., Pessiot, L., Trémas, T., Cadau, E., … Fernandez, V. (2017). Copernicus Sentinel-2A calibration and products validation status. Remote Sensing, 9(6). https://doi.org/10.3390/rs9060584
11. Ge, Z., Liu, S., Wang, F., Li, Z. & Sun, J. (n.d.). YOLOX: Exceeding YOLO Series in 2021 V100 batch 1 Latency (ms) YOLOX-L YOLOv5-L YOLOX-DarkNet53 YOLOv5-Darknet53 EfficientDet5 COCO AP (%) Number of parameters (M) Figure 1: Speed-accuracy trade-off of accurate models (top) and Size-accuracy curve of lite models on mobile devices (bottom) for YOLOX and other state-of-the-art object detectors (Vol. 5). https://github.com/ultralytics/yolov3
12. Geilenkirchen, G., Bolech, M., Hulskotte, J., Dellaert, S., Ligterink, N., Sijstermans, M., Geertjes, K. & Felter, K. (n.d.). Methods for Calculating the Emissions of Transport in the Netherlands Colophon Methods for calculating the emissions of transport in the Netherlands Author(s).
13. Girshick, R. (n.d.). Fast R-CNN. https://github.com/rbgirshick/
14. Girshick, R., Donahue, J., Darrell, T. & Malik, J. (n.d.). Rich feature hierarchies for accurate object detection and semantic segmentation. https://arxiv.org/abs/1311.2524
15. Gorroño, J., Varon, D. J., Irakulis-Loitxate, I. & Guanter, L. (2023). Understanding the potential of Sentinel-2 for monitoring methane point emissions. Atmospheric Measurement Techniques, 16(1), 89–107. https://doi.org/10.5194/amt-16-89-2023
16. Hanna, J., Borth, D. & Mommert, M. (2023). Physics-Guided Multitask Learning for Estimating Power Generation and CO2 Emissions From Satellite Imagery. IEEE Transactions on Geoscience and Remote Sensing, 61. https://doi.org/10.1109/TGRS.2023.3286444
17. Hanna, J., Mommert, M. & Borth, D. (n.d.). Estimation of Power Generation and CO2 Emissions Using Satellite Imagery. https://doi.org/10.5281/zenodo
18. Helber, P., Bischke, B., Dengel, A. & Borth, D. (2019). Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(7), 2217–2226. https://doi.org/10.1109/JSTARS.2019.2918242
19. Hobbs, M., Kargar, A. R., Couture, H., Freeman, J., Söldner-Rembold, I., Ferreira, A., Jeyaratnam, J., O’Connor, J., Lewis, J., Koenig, H., McCormick, C., Nakano, T., Dalisay, C., Davitt, A., Gans, L., Lewis, C., Volpato, G., Gray, M. & McCormick, G. (2023). Inferring Carbon Dioxide Emissions From Power Plants Using Satellite Imagery and Machine Learning. IGARSS 2023 - IEEE International Geoscience and Remote Sensing Symposium, 4911–4914. https://doi.org/10.1109/igarss52108.2023.10283046
20. Hosang, J., Benenson, R., Dollar, P. & Schiele, B. (2016). What Makes for Effective Detection Proposals? IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(4), 814–830. https://doi.org/10.1109/TPAMI.2015.2465908
21. Jain, A. (2022). Employing Deep Learning to Quantify Power Plant Greenhouse Gas Emissions via Remote Sensing Data. AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges. https://www.climatechange.ai/papers/aaaifss2022/11
22. Li, B., Xie, X., Wei, X. & Tang, W. (2021). Ship detection and classification from optical remote sensing images: A survey. In Chinese Journal of Aeronautics (Vol. 34, Issue 3, pp. 145–163). https://doi.org/10.1016/j.cja.2020.09.022
23. Lin, Y., Li, J., Wei, S. & Liu, S. (2024). OptiShipNet: Efficient Ship Detection in Complex Marine Environments Using Optical Remote Sensing Images. Journal of Marine Science and Engineering, 12(10). https://doi.org/10.3390/jmse12101786
24. Luo, Z., He, T., Yi, W., Zhao, J., Zhang, Z., Wang, Y., Liu, H. & He, K. (2024). Advancing shipping NOxpollution estimation through a satellite-based approach. PNAS Nexus, 3(1). https://doi.org/10.1093/pnasnexus/pgad430
25. Mommert, M., Sigel, M., Neuhausler, M., Scheibenreif, L. & Borth, D. (2020). Characterization of Industrial Smoke Plumes from Remote Sensing Data. http://arxiv.org/abs/2011.11344
26. Nunes, R. A. O., Alvim-Ferraz, M. C. M., Martins, F. G. & Sousa, S. I. V. (2017). Assessment of shipping emissions on four ports of Portugal. Environmental Pollution, 231, 1370–1379. https://doi.org/10.1016/j.envpol.2017.08.112
27. Passig, J., Schade, J., Irsig, R., Li, L., Li, X., Zhou, Z., Adam, T. & Zimmermann, R. (2021). Detection of Ship Plumes from Residual Fuel Operation in Emission Control Areas using Single-Particle Mass Spectrometry. https://doi.org/10.5194/amt-2020-482
28. Poucin, F., Ouerghi, E., Lajouanie, S., de Almeida Rodrigues, H., Facciolo, G., de Franchis, C. & Hessel, C. (2023). Machine Learning and Feature Extraction for Industrial Smoke Plumes Detection from Sentinel-2 Images. International Geoscience and Remote Sensing Symposium (IGARSS), 2023-July, 6113–6116. https://doi.org/10.1109/IGARSS52108.2023.10282662
29. Ransheng, Y., Ying, H. & Hongfei, Z. (2023). Ship Detection in Maritime Surveillance System Based on YOLOX. 2023 8th International Conference on Image, Vision and Computing, ICIVC 2023, 146–153. https://doi.org/10.1109/ICIVC58118.2023.10269862
30. Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. (n.d.). You Only Look Once: Unified, Real-Time Object Detection. http://pjreddie.com/yolo/
31. Ren, S., He, K., Girshick, R. & Sun, J. (n.d.). Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. https://github.com/
32. Rouet-Leduc, B. & Hulbert, C. (2024). Automatic detection of methane emissions in multispectral satellite imagery using a vision transformer. Nature Communications, 15(1). https://doi.org/10.1038/s41467-024-47754-y
33. Rowley, A. & Karakuş, O. (2023). Predicting air quality via multimodal AI and satellite imagery. Remote Sensing of Environment, 293. https://doi.org/10.1016/j.rse.2023.113609
34. Scheibenreif, L., Mommert, M. & Borth, D. (2021). Estimation of Air Pollution with Remote Sensing Data: Revealing Greenhouse Gas Emissions from Space. http://arxiv.org/abs/2108.13902
35. Strandgren, J., Krutz, D., Wilzewski, J., Paproth, C., Sebastian, I., R. Gurney, K., Liang, J., Roiger, A. & Butz, A. (2020). Towards spaceborne monitoring of localized CO2 emissions: An instrument concept and first performance assessment. Atmospheric Measurement Techniques, 13(6), 2887–2904. https://doi.org/10.5194/amt-13-2887-2020
36. van Geffen, J., Folkert Boersma, K., Eskes, H., Sneep, M., ter Linden, M., Zara, M. & Pepijn Veefkind, J. (2020). S5P TROPOMI NO2 slant column retrieval: Method, stability, uncertainties and comparisons with OMI. Atmospheric Measurement Techniques, 13(3), 1315–1335. https://doi.org/10.5194/amt-13-1315-2020
37. Wan, Z., Ji, S., Liu, Y., Zhang, Q., Chen, J. & Wang, Q. (2020). Shipping emission inventories in China’s Bohai Bay, Yangtze River Delta, and Pearl River Delta in 2018. Marine Pollution Bulletin, 151. https://doi.org/10.1016/j.marpolbul.2019.110882
38. Wang, Q., Shi, W., Li, Z. & Atkinson, P. M. (2016). Fusion of Sentinel-2 images. Remote Sensing of Environment, 187, 241–252. https://doi.org/10.1016/j.rse.2016.10.030
39. Wu, J., O’Sullivan, C., Orlandi, F., O’Sullivan, D. & Dev, S. (2023). Measurement of Industrial Smoke Plumes from Satellite Images. International Geoscience and Remote Sensing Symposium (IGARSS), 2023-July, 5680–5683. https://doi.org/10.1109/IGARSS52108.2023.10282713
40. Wu, Q. (2020). geemap: A Python package for interactive mapping with Google Earth Engine. Journal of Open Source Software, 5(51), 2305. https://doi.org/10.21105/joss.02305
41. Zhang, F., Chen, Y., Tian, C., Wang, X., Huang, G., Fang, Y. & Zong, Z. (2014). Identification and quantification of shipping emissions in Bohai Rim, China. Science of the Total Environment, 497–498, 570–577. https://doi.org/10.1016/j.scitotenv.2014.08.016
42. Zhang, Q., Geng, G. N., Wang, S. W., Richter, A. & He, K. bin. (2012). Satellite remote sensing of changes in NO x emissions over China during 1996-2010. Chinese Science Bulletin, 57(22), 2857–2864. https://doi.org/10.1007/s11434-012-5015-4
43. Zhang, Q., Huang, Y. & Song, R. (2022). A Ship Detection Model Based on YOLOX with Lightweight Adaptive Channel Feature Fusion and Sparse Data Augmentation. AVSS 2022 - 18th IEEE International Conference on Advanced Video and Signal-Based Surveillance. https://doi.org/10.1109/AVSS56176.2022.9959441
44. WHO global air quality guidelines. (2021). Particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide.
45. Web 1, https://www.esa.int/Applications/Observing_the_Earth/Copernicus/Sentinel-2, (20.08.2024 tarihinde alınmıştır).
46. Web 2, https://dataspace.copernicus.eu/, (20.08.2024 tarihinde alınmıştır).
47. Web 3, https://unctadstat.unctad.org/datacentre/dataviewer/US.PortCalls, (20.08.2024 tarihinde alınmıştır).
48. Web 4, https://github.com/Megvii-BaseDetection/YOLOX?tab=readme-ov-file#standard-models, (10.10.2024 tarihinde alınmıştır).
49. Web 5, https://github.com/ultralytics/yolov5, (10.10.24 tarihinde alınmıştır).