Pankromatik Bandın Piksel Tabanlı Sınıflandırmaya Etkisi

Year 2021, , 32 - 40, 13.03.2021

Ekrem Saralıoğlu

https://doi.org/10.48123/rsgis.852055

Abstract

Bu çalışmada, pankromatik bandın çok bantlı uydu görüntülerinin piksel tabanlı sınıflandırma doğruluğuna etkisi deneysel olarak araştırılmıştır. Çok yüksek uzamsal çözünürlüklü multispektral uydu görüntüleri multispektral bantlar yanında daha yüksek uzamsal çözünürlükte bir pankromatik bant içermektedir. Bu nedenle sınıflandırma aşamasında çeşitli seçenekler oluşmaktadır.Örneğin, sınıflandırma yapıldığında bu bant kullanılmalı mı? Kullanılacaksa nasıl kullanılmalı? Sınıflandırma doğrulukları arasında ne kadar fark olur? Pankromatik bandın sınıflandırma sonuçlarına etkisini incelemek amacıyla 4 adet senaryo oluşturulmuştur. İlk senaryoda pankromatik bant görüntü kaynaştırma yapılarak sınıflandırmada kullanılmıştır. İkinci senaryoda sadece multispektral bantlar üzerinden sınıflandırma yapılmıştır. Üçüncü senaryoda, multispektral bantların boyutu en yakın komşuluk algoritması kullanılarak pankromatik bant boyutuna getirilmiştir. Daha sonra tüm bantların arkasına pankromatik bant eklenerek sınıflandırma yapılmıştır. Son senaryoda ise sadece pankromatik bant sınıflandırılmıştır. En yüksek sınıflandırma doğruluğu pan-keskinleştirme yapılan görüntülerde elde edilmiştir. WorldView-2 görüntüsünün kendi pankromatik bandı ile kaynaştırılması sonucu elde edilen görüntünün destek vektör makineleri ve rastgele orman ile sınıflandırma sonuçları sırasıyla %78 ve %75 olarak bulunmuştur. IKONOS görüntüsü için pan-keskinleştirme yapılmış görüntüde sınıflandırma doğrulukları ise aynı sırada %70 ve %66 olarak bulunmuştur.

Keywords

Multispektral uydu görüntüleri, Görüntü sınıflandırma, Pan-keskinleştirme

The Effect of Panchromatic Band on Pixel-Based Classification

Abstract

In this study, the effect of panchromatic band on pixel-based classification accuracy of multispectral satellite images was investigated experimentally. Very high spatial resolution multispectral satellite images contain not only multispectral bands but also a higher resolution panchromatic band, which offers a couple of options in the classification phase. For example, should this band be used within classification? If so, how should it be used? What is the difference in the classification accuracies? In order to examine the effect of the panchromatic band on classification results 4 scenarios were created. In the first scenario, panchromatic band was used in classification by pansharpening. In the second scenario, classification was conducted using only the multispectral bands. In the third scenario, the size of the multispectral bands was brought to that of the panchromatic band using the nearest neighborhood algorithm. Then, the classification was made by combining the panchromatic band with all the multispectral bands. In the last scenario, only the panchromatic band was classified. The highest classification accuracy was obtained with pansharpened images. The support vector machines and random forest classification accuracies of the image obtained by pansharpening the WorldView-2 image with its own panchromatic band were found to be 78% and 75%, respectively. The IKONOS pansharpened image resulted in classification accuracies of 70% and 66% in the same order.

Keywords

Multispectral satellite images, Image classification, Pan-sharpening

References

• Abou EL-Magd, I., & Tanton, T. W. (2003). Improvements in land use mapping for irrigated agriculture from satellite sensor data using a multi-stage maximum likelihood classification. International Journal of Remote Sensing, 24(21), 4197-4206.
• Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.
• 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.
• Civco, D. L. (1993). Artificial neural networks for land-cover classification and mapping. International journal of geographical information science, 7(2), 173-186.
• Du, Q., & Chang, C. I. (2001). A linear constrained distance-based discriminant analysis for hyperspectral image classification. Pattern Recognition, 34(2), 361-373.
• Gomariz-Castillo, F., Alonso-Sarría, F., & Cánovas-García, F. (2017). Improving classification accuracy of multi-temporal Landsat Images by Assessing the Use of different algorithms, textural and ancillary information for a mediterranean semiarid area from 2000 to 2015. Remote Sensing, 9(10), 1058.
• Huang, X., Wang, C., & Li, Z. (2018). A near real-time flood-mapping approach by integrating social media and post-event satellite imagery. Annals of GIS, 24(2), 113-123.
• Khatami, R., Mountrakis, G., & Stehman, S.V. (2016). A meta-analysis of remote sensing research on supervised pixel-based land-cover image classification processes: General guidelines for practitioners and future research. Remote Sensing of Environment, 177, 89-100.
• Laben, C.A., & Brower, B.V., (2000). Process for enhancing the spatial resolution of multispectral imagery using pan-sharpening, United States Eastman Kodak Company (Rochester, NY). US Patent 6011875.
• Li, C., Liu, L., Wang, J., Zhao, C., & Wang, R. (2004, September). Comparison of two methods of the fusion of remote sensing images with fidelity of spectral information. In IGARSS 2004. 2004 IEEE International Geoscience and Remote Sensing Symposium (Vol. 4, pp. 2561-2564). IEEE.
• Melgani, F., & Bruzzone, L. (2004). Classification of hyperspectral remote sensing images with support vector machines. IEEE Transactions on geoscience and remote sensing, 42(8), 1778-1790.
• Pedergnana, M., Marpu, P. R., Dalla Mura, M., Benediktsson, J. A., & Bruzzone, L. (2012). Classification of remote sensing optical and LiDAR data using extended attribute profiles. IEEE Journal of Selected Topics in Signal Processing, 6(7), 856-865.
• Petropoulos, G.P., Kalaitzidis, C., & Vadrevu, K. P. (2012). Support vector machines and object-based classification for obtaining land-use/cover cartography from Hyperion hyperspectral imagery. Computers & Geosciences, 41, 99-107.
• Samaniego, L., Bárdossy, A., & Schulz, K. (2008). Supervised classification of remotely sensed imagery using a modified k-NN technique. IEEE Transactions on Geoscience and Remote Sensing, 46(7), 2112-2125.
• Saralioglu, E., & Gungor, O. (2020). Semantic segmentation of land cover from high resolution multispectral satellite images by spectral-spatial convolutional neural network. Geocarto International, 1-21.
• Thanh Noi, P., & Kappas, M. (2018). Comparison of random forest, k-nearest neighbor, and support vector machine classifiers for land cover classification using Sentinel-2 imagery. Sensors, 18(1), 18.
• Wady, S.M.A., Bentoutou, Y., Bengermikh, A., Bounoua, A., & Taleb, N. (2020). A new IHS and wavelet based pansharpening algorithm for high spatial resolution satellite imagery. Advances in Space Research, 66(7), 1507-1521.
• Yıldırım, D., & Güngör, O. (2012). A novel image fusion method using IKONOS satellite images. Journal of Geodesy and Geoinformation, 1(1), 75-83.
• Yılmaz, V. (2020). Metasezgisel Guguk Kuşu Arama Algoritması ile Görüntü Kaynaştırma. Türk Uzaktan Algılama ve CBS Dergisi, 1(1), 1-12.
• Zhang, L., Zhang, L., & Du, B. (2016). Deep learning for remote sensing data: A technical tutorial on the state of the art. IEEE Geoscience and Remote Sensing Magazine, 4(2), 22-40.
• Zheng, H., Du, P., Chen, J., Xia, J., Li, E., Xu, Z., ... & Yokoya, N. (2017). Performance evaluation of downscaling Sentinel-2 imagery for land use and land cover classification by spectral-spatial features. Remote Sensing, 9(12), 1274.