Research Article
BibTex RIS Cite

Derin Öğrenme Amaçlı Etiketli Veri Üretimi İçin Bir Web Platformu

Year 2020, Volume: 1 Issue: 1, 24 - 33, 31.03.2020

Abstract

Klasik RGB görüntülerine göre uydu görüntülerinin derin öğrenme amaçlı kullanımlarında önemli problemler bulunmaktadır. Uzaktan algılanmış görüntüler için etiketli eğitim verisi eksikliği bu problemlerden biridir. Uydu görüntüleri için yetersiz etiketli eğitim verisi sorununun çözümünü doğru ve kısa sürede sağlayabilecek en önemli yöntemlerden biri kitle kaynak kullanımıdır. Bu çalışmada yüksek çözünürlüklü uydu görüntüleri için etiketli verilerin kitleler tarafından toplanmasını sağlamak amacıyla oluşturulan web platform tanıtılmaktadır. Bu platform dinamik bir yapıda olup aynı anda farklı kullanıcıların da kullanabileceği şekilde tasarlanmıştır. Platformda kullanıcılar Dünya yüzeyini kapsayan Google uydu altlığı kullanılabildiği gibi veri tabanına eklenecek görüntüler üzerinden de etiketli veri oluşturabilmektedir. Bu sayede yüksek çözünürlüklü uydu görüntüleri için Dünya çapında görüntülerden çıkarılabilecek bütün sınıflara ait (bina, yol, orman, akarsu, fındık, çay, gemi, uçak vb.) etiketli verilerin hazırlanması mümkün olmaktadır. İstenen eğitim sınıflarının tanımı ve kullanıcıların platformu efektif kullanmaları amacıyla yardım dokümanları web platforma eklenmiştir. Kullanıcılar görüntüler üzerinden poligon aracı ile yardım dokümanında belirtilen sınıfları çevirebilmektedir. Daha sonra kullanıcılar tarafından bu sınıflara etiket değeri verilmektedir. Oluşturulan etiketli verilerden doğru ve yanlışların tespit edilebilmesi amacıyla veri doğrulama modülü de web platforma eklenmiştir. Bu modülde kullanıcılar önceki kişilerin çevirdikleri ve etiketledikleri sınıfları yardım dokümanından kontrol ederek puanlandırmaktadır. Sonuçta en yüksek puan alan doğru etiketli veriler seçilmektedir.

References

  • Akar, Ö., & Güngör, O. (2015). Integrating multiple texture methods and NDVI to the Random Forest classification algorithm to detect tea and hazelnut plantation areas in northeast Turkey. International Journal of Remote Sensing, 36(2), 442-464.
  • Arel, I., Rose, D. C., & Karnowski, T. P. (2010). Deep machine learning-a new frontier in artificial intelligence research. IEEE computational intelligence magazine, 5(4), 13-18.
  • Carranza-García, M., García-Gutiérrez, J., & Riquelme, J. C. (2019). A framework for evaluating land use and land cover classification using convolutional neural networks. Remote Sensing, 11(3), 274.
  • Chen, Y., Lin, Z., Zhao, X., Wang, G., & 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.
  • Chen, Y., Zhao, X., & Jia, X. (2015). Spectral–spatial 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.
  • Ghaffar, M. A. A., McKinstry, A., Maul, T., & Vu, T. T. (2019). Data augmentation approaches for satellite image super-resolution. ISPRS Annals of Photogrammetry, Remote Sensing & Spatial Information Sciences, Vol. IV-2/W7, 47–54, https://doi.org/10.5194/isprs-annals-IV-2-W7-47-2019.
  • He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
  • Howe, J. (2008). Crowdsourcing: Kalabalıkların gücü bir işin geleceğine nasıl şekil verebilir?. Birinci Baskı, Koç Sistem Yayınları, İstanbul.
  • Howe, J. (2006). The Rise of Crowdsourcing, Wired, https://www.wired.com/wired/archive/14.06/crowds.html, Erişim Tarihi: 10 Ekim 2019.
  • Kemker, R., Salvaggio, C., & Kanan, C. (2018). Algorithms for semantic segmentation of multispectral remote sensing imagery using deep learning. ISPRS Journal of Photogrammetry and Remote Sensing, 145, 60-77.
  • Li, Y., Zhang, H., Xue, X., Jiang, Y., & Shen, Q. (2018). Deep learning for remote sensing image classification: A survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 8(6), e1264.
  • Maggiori, E., Tarabalka, Y., Charpiat, G., & Alliez, P. (2016). Convolutional neural networks for large-scale remote-sensing image classification. IEEE Transactions on Geoscience and Remote Sensing, 55(2), 645-657.
  • Saralioglu, E., & Gungor, O. (2019). Use of crowdsourcing in evaluating post-classification accuracy. European Journal of Remote Sensing, 52(sup1), 137-147.
  • Song, J., Gao, S., Zhu, Y., & Ma, C. (2019). A survey of remote sensing image classification based on CNNs. Big Earth Data, 3(3), 232-254.
  • Thenkabail, P. S. (2016). Remote Sensing Handbook; Volume 1: Remotely Sensed Data Characterization, Classification, and Accuracies. Taylor & Francis.
  • Yu, X., Wu, X., Luo, C., & Ren, P. (2017). Deep learning in remote sensing scene classification: a data augmentation enhanced convolutional neural network framework. GIScience & Remote Sensing, 54(5), 741-758.
  • Zhang, C. (2018). Deep learning for land cover and land use classification (Doctoral dissertation, Lancaster University).

A Web Platform for the Generation of Labeled Data for Deep Learning

Year 2020, Volume: 1 Issue: 1, 24 - 33, 31.03.2020

Abstract

Compared to conventional RGB images, new problems arise in the use of satellite images for deep learning. The absence of labeled training data for remotely sensed images is one of these problems. One of the methods that can solve the problem of insufficient labelled data available for satellite images in a short time and accurately is crowdsourcing. This study introduces a web platform created to ensure that the labelled data for high-resolution satellite images can be collected by the masses. This platform has a dynamic structure and is designed to be used by different users simultaneously. In order to create tagged data, users can use Google Earth satellite images covering the entire Earth’s surface as well as new images added to the database. In this way, it will be possible to generate labelled data for all types of classes (buildings, roads, forests, streams, hazelnuts, tea, ships, planes, etc.) that can be extracted from images around the world. Help documents have been added to the web platform to identify training classes for users, enabling them to use the platform effectively. Users can use the polygon tool to create descriptive fields for the classes specified in the help document to create labels. Data verification module has also been added to the web platform in order to determine the correct and incorrect labels. In this module, users verify and score the labels created by other people using the help document. As a result, the correctly labelled data with the highest score are selected.

References

  • Akar, Ö., & Güngör, O. (2015). Integrating multiple texture methods and NDVI to the Random Forest classification algorithm to detect tea and hazelnut plantation areas in northeast Turkey. International Journal of Remote Sensing, 36(2), 442-464.
  • Arel, I., Rose, D. C., & Karnowski, T. P. (2010). Deep machine learning-a new frontier in artificial intelligence research. IEEE computational intelligence magazine, 5(4), 13-18.
  • Carranza-García, M., García-Gutiérrez, J., & Riquelme, J. C. (2019). A framework for evaluating land use and land cover classification using convolutional neural networks. Remote Sensing, 11(3), 274.
  • Chen, Y., Lin, Z., Zhao, X., Wang, G., & 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.
  • Chen, Y., Zhao, X., & Jia, X. (2015). Spectral–spatial 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.
  • Ghaffar, M. A. A., McKinstry, A., Maul, T., & Vu, T. T. (2019). Data augmentation approaches for satellite image super-resolution. ISPRS Annals of Photogrammetry, Remote Sensing & Spatial Information Sciences, Vol. IV-2/W7, 47–54, https://doi.org/10.5194/isprs-annals-IV-2-W7-47-2019.
  • He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
  • Howe, J. (2008). Crowdsourcing: Kalabalıkların gücü bir işin geleceğine nasıl şekil verebilir?. Birinci Baskı, Koç Sistem Yayınları, İstanbul.
  • Howe, J. (2006). The Rise of Crowdsourcing, Wired, https://www.wired.com/wired/archive/14.06/crowds.html, Erişim Tarihi: 10 Ekim 2019.
  • Kemker, R., Salvaggio, C., & Kanan, C. (2018). Algorithms for semantic segmentation of multispectral remote sensing imagery using deep learning. ISPRS Journal of Photogrammetry and Remote Sensing, 145, 60-77.
  • Li, Y., Zhang, H., Xue, X., Jiang, Y., & Shen, Q. (2018). Deep learning for remote sensing image classification: A survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 8(6), e1264.
  • Maggiori, E., Tarabalka, Y., Charpiat, G., & Alliez, P. (2016). Convolutional neural networks for large-scale remote-sensing image classification. IEEE Transactions on Geoscience and Remote Sensing, 55(2), 645-657.
  • Saralioglu, E., & Gungor, O. (2019). Use of crowdsourcing in evaluating post-classification accuracy. European Journal of Remote Sensing, 52(sup1), 137-147.
  • Song, J., Gao, S., Zhu, Y., & Ma, C. (2019). A survey of remote sensing image classification based on CNNs. Big Earth Data, 3(3), 232-254.
  • Thenkabail, P. S. (2016). Remote Sensing Handbook; Volume 1: Remotely Sensed Data Characterization, Classification, and Accuracies. Taylor & Francis.
  • Yu, X., Wu, X., Luo, C., & Ren, P. (2017). Deep learning in remote sensing scene classification: a data augmentation enhanced convolutional neural network framework. GIScience & Remote Sensing, 54(5), 741-758.
  • Zhang, C. (2018). Deep learning for land cover and land use classification (Doctoral dissertation, Lancaster University).
There are 17 citations in total.

Details

Primary Language Turkish
Subjects Photogrammetry and Remote Sensing
Journal Section Research Articles
Authors

Ekrem Saralıoğlu 0000-0002-0609-3338

Oğuz Güngör 0000-0002-3280-5466

Publication Date March 31, 2020
Submission Date January 30, 2020
Acceptance Date March 16, 2020
Published in Issue Year 2020 Volume: 1 Issue: 1

Cite

APA Saralıoğlu, E., & Güngör, O. (2020). Derin Öğrenme Amaçlı Etiketli Veri Üretimi İçin Bir Web Platformu. Türk Uzaktan Algılama Ve CBS Dergisi, 1(1), 24-33.