Journals / Harita Dergisi / 2021 / Cilt: 87 - Sayı: 166

Basic Principles of Deep Learning and Its Applications in Remote Sensing

Derin Öğrenmenin Temel Prensipleri ve Uzaktan Algılama Alanındaki Uygulamaları

Pages
25–43
DOI
—
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

As a result, the place and importance of deep learning in the current remote sensing literature is reviewed and the current literature summary is also presented

📌 3 Key Takeaways

🎯
Objective & Hypothesis:Today, the use of deep learning architectures in remote sensing applications is becoming more and more common every day.
📊
Key Findings & Data:Compared to traditional methods, deep learning models appear to produce more successful results in the processing of remotely detected data.
💡
Conclusion & Impact:As a result, the place and importance of deep learning in the current remote sensing literature is reviewed and the current literature summary is also presented
🔬Design:Lab & Field Experiment
📍Scope:Comprehensive Sample
⚙️Variables:Domain Parameters

Abstract

Today, the use of deep learning architectures in remote sensing applications is becoming more and more common every day. Compared to traditional methods, deep learning models appear to produce more successful results in the processing of remotely detected data. These models are often preferred, especially in the field of image processing, in parallel with the development of computer hardware. The deep learning approach has important advantages such as the minimum level of human intervention during its implementation, the ability to solve complex problems such as object detection and classification, and its effectiveness in processing high-resolution satellite images. However, the need for a large data set labeled for training deep learning models is a major problem. Creating these data sets with high accuracy requires time and workload. In this study, convolution, pooling flattening, and dropout layers, which are main components of convolutional neural networks, a deep learning model, are described in detail. In addition, the principles of activation and cost functions affecting deep learning architecture are presented. In addition, the use of deep learning architectures in remote sensing has been studied under the headings of preprocessing, image classification, object detection, image interpretation and other application areas. As a result, the place and importance of deep learning in the current remote sensing literature is reviewed and the current literature summary is also presented

Özet

Günümüzde, derin öğrenme mimarilerinin uzaktanalgılama uygulamalarında kullanımı her geçen gündaha yaygın hale gelmektedir. Geleneksel yöntemlerlekarşılaştırıldığında, uzaktan algılanmış verilerinişlenmesinde derin öğrenme modellerinin daha başarılısonuçlar ürettiği görülmektedir. Bu modeller, bilgisayardonanımlarının gelişimine paralel olarak özelliklegörüntü işleme alanında sıklıkla tercih edilmektedir.Derin öğrenme yaklaşımının uygulanması esnasındainsan müdahalesinin minimum seviyede olması, nesnetespiti ve sınıflandırma gibi karmaşık problemleriçözebilmesi ve yüksek çözünürlüklü uydu görüntülerininişlenmesindeki etkinliği gibi önemli avantajlarıbulunmaktadır. Ancak, derin öğrenme modellerinineğitimi için etiketlenmiş büyük veri seti ihtiyacı iseönemli bir problem olarak karşımıza çıkmaktadır. Buveri setlerinin yüksek doğrulukla oluşturulabilmesizaman ve iş yükü gerektirmektedir. Bu çalışmada, birderin öğrenme modeli olan evrişimsel sinir ağımimarisine ait evrişim, ortaklama, düzleştirme veseyreltme katmanları ayrıntılı bir şekilde açıklanmıştır.Ayrıca, derin öğrenme mimarisini etkileyen hareketegeçirme ve maliyet fonksiyonlarının çalışma prensipleridetaylı şekilde sunulmuştur. Ayrıca, uzaktan algılamadaderin öğrenme mimarilerinin kullanımı; önişleme,görüntü sınıflandırma, nesne tespiti, görüntüyorumlama ve diğer uygulama alanları başlıkları altındairdelenmiştir. Sonuç olarak, derin öğrenmenin günceluzaktan algılama literatüründeki yeri ve önemi ortayakoyularak güncel literatür özeti de sunulmuştur.

🔖

Cite This Article

YILMAZ, E. Ö., KAVZOĞLU, T. (2021). Basic Principles of Deep Learning and Its Applications in Remote Sensing. Harita Dergisi, Cilt: 87 - Sayı: 166, 25–43. https://ajindex.com/articles/347939

References (1)

  1. Abdel-Hamid, O., Mohamed, A.R., Jiang, H., Deng, L., Penn, G. ve Yu, D. (2014). Convolutional neural networks for speech recognition. IEEE Transactions on Audio, Speech and Language Processing, 22(10), 1533-1545. doi:10.1109/TASLP2014.2339736 Al-amri, S.S., Kalyankar, N.V. ve Khamitkar, S.D. (2010). A comparative Study of Removal Noise from Remote Sensing Image. International Journal of Computer Science Issues, 7(1), 32– 36. arXiv:1002.1148 Alganci, U., Soydas, M. ve Sertel, E. (2020). Comparative research on deep learning approaches for airplane detection from very high-resolution satellite images. Remote Sensing,12(3), 458. doi:10.3390/rs12030458 Alom, M.Z., Taha, T.M., Yakopcic, C., Westberg, S., Sidike, P., Nasrin, M.S., … Asari, V.K. (2019). A state-of-the-art survey on deep learning theory and architectures. Electronics, 8(3), 292. doi:10.3390/electronics8030292 Audebert, N., Le Saux, B. ve Lefevre, S. (2019). Deep learning for classification of hyperspectral data: A comparative review. IEEE Geoscience and Remote Sensing Magazine, 7(2), 159-173. doi:10.1109/ MGRS.2019.2912563 Ball, J.E., Anderson, D.T. ve Chan, C.S. (2017). Comprehensive survey of deep learning in remote sensing: theories, tools, and challenges for the community. Journal of Applied Remote Sensing, 11(04), 042609. doi:10.1117/1.jrs.11.042609 Chen, K., Fu, K., Gao, X., Yan, M., Sun, X. ve Zhang, H. (2017, Aralık). Building extraction from remote sensing images with deep learning in a supervised manner. 2017 IEEE International Geoscience and Remote Sensing Symposium, Teksas, ABD. Chen, J., Huang, H., Peng, J., Zhu, J., Chen, L., Tao, C. ve Li, H. (2021). Contextual Information-Preserved Architecture Learning for Remote-Sensing Scene Classification. IEEE Transactions on Geoscience and Remote Sensing (Early Access). doi: 10.1109/TGRS.2021.3063287 Chen, Y., Lin, Z., Zhao, X., Wang, G. ve Gu, Y. (2014). Deep learning-based classification of hyperspectral data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 7(6), 2094-2107. doi: 10.1109/JSTARS.2014.2329330 Chen, Y., Zhao, X. ve Jia, X. (2015). Spectralspatial classification of hyperspectral data based on deep belief network. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 8(6), 2381-2392. doi:10.1109/JSTARS.2015.2388577 Cheng, G. ve Han, J. (2016). A survey on object detection in optical remote sensing images. ISPRS Journal of Photogrammetry and Remote Sensing, 117, 11–28.