Kenar Koruyan Filtreler İçeren Geliştirilmiş Aktif Derin Öğrenme Çerçevesini Kullanan Hiperspektral Görüntü Sınıflandırılması

Year 2024, Volume: 5 Issue: 1, 54 - 68, 28.03.2024

Zainab Dheyaa Al-sammarraie , Ali Can Karaca

https://doi.org/10.48123/rsgis.1402066

Abstract

Tarım, jeolojik araştırma ve çevresel izleme gibi uygulamalar için uydu verilerinden değerli bilgiler elde etmek amacıyla hiperspektral görüntünün sınıflandırılması önemli bir görevdir. Bu süreçte her pikselin etiketlenmesi zaman alıcıdır ve mali kaynak gerektirmektedir. Bu amaçla az sayıda örnekle çalışmak çok önemlidir. Sınırlı sayıda örnek altında yüksek sınıflandırma performansı sağlamak için bu makale, performansı aktif bir öğrenme çerçevesiyle geliştirmeyi amaçlamaktadır. Çerçeve, boyut azaltma, kenar koruma filtresi ve aktif öğrenme adımlarını içermektedir. Bu açıdan bakıldığında performans üzerindeki etkilerini analiz etmek için farklı kenar koruyucu filtre yöntemlerini araştırılmıştır. Kenar koruyucu filtreleri boyut azaltmayla birleştiren çalışma, görüntü kalitesini korurken ve gürültüyü azaltırken sınıflandırma performansını artıran benzersiz bir yöntem sunmaktadır. Ağırlıklı En Küçük Kareler (WLS), Ortak Histogram Ağırlıklı Ortanca Filtre (Joint WMF), Hızlı Global Görüntü Yumuşatma (FGS), Bilateral Filtre (BF), and Static/dynamic (SD) olmak üzere toplam beş kenar koruyan filtre değerlendirilmiştir. Deneylerimiz referans araştırmayla (CNN+AL) karşılaştırıldığında önerilen çerçevenin Indian Pines, Pavia Üniversitesi ve Salinas veri kümeleri için genel ve ortalama doğruluğu yaklaşık %2-5 artırdığını göstermektedir.

Keywords

Aktif öğrenme, Kenar koruyan filtreler, Hiperspektral görüntüleme

Hyperspectral Image Classification Using Improved Active Deep Learning Framework Including Edge Preserving Filters

Abstract

To extract valuable information from satellite data for applications such as agriculture, geological research, and environmental monitoring, the classification of hyperspectral images is an essential task. Labeling each pixel in this process is time-consuming and requires financial resources. To this end, working with a small number of samples is very important. In order to provide high classification performances with a limited number of samples, this paper aims to enhance the performance with an active learning framework. The framework incorporates dimensionality reduction, an edge-preserving filter, and active learning steps. From this perspective, we investigated different edge-preserving filter methods to analyze the effects on performance. By combining edge-preserving filters with dimensionality reduction, the study presents a unique method that improves classification performance while maintaining image quality and reducing noise. The following five edge-preserving smoothing filters are evaluated: weighted least squares (WLS), Joint-Histogram weighted median filter (Joint WMF), fast global image smoother (FGS), bilateral filter (BF), and static/dynamic (SD). Our experiments demonstrate that compared to the reference research (CNN+AL+MRF), the proposed framework increased overall and average accuracies about 2-5% for Indian Pines, Pavia University, and Salinas datasets.

Keywords

Active learning, Edge preserving filters, Hyperspectral imaging

References

• Alcolea, A., Paoletti, M. E., Haut, J. M., Resano, J., & Plaza, A. (2020). Inference in supervised spectral classifiers for on-board hyperspectral imaging: An overview. Remote Sensing, 12(3), 534. https://doi.org/10.3390/rs12030534
• Cao, X., Yao, J., Xu, Z., & Meng, D. (2020). Hyperspectral image classification with convolutional neural network and active learning. IEEE Transactions on Geoscience and Remote Sensing, 58(7), 4604–4616.
• Chen, M., Wang, Q., & Li, X. (2018). Discriminant analysis with graph learning for hyperspectral image classification. Remote Sensing, 10(6), 836. https://doi.org/10.3390/rs10060836
• Chen, Y., Jiang, H., Li, C., Jia, X., & Ghamisi, P. (2016). Deep feature extraction and classification of hyperspectral images based on convolutional neural networks. IEEE Transactions on Geoscience and Remote Sensing, 54(10), 6232–6251.
• Gupta, V., Sastry, S., & Mitra, S. K. (2020). Hyperspectral image classification using trilateral filter and deep learning. In 2020 IEEE International Symposium on Sustainable Energy, Signal Processing and Cyber Security (iSSSC) (pp. 1-6). IEEE.
• Ham, B., Cho, M., & Ponce, J. (2015). Robust image filtering using joint static and dynamic guidance. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 4823-4831). IEEE.
• Haut, J. M., Paoletti, M. E., Plaza, J., Li, J., & Plaza, A. (2018). Active learning with convolutional neural networks for hyperspectral image classification using a new Bayesian approach. IEEE Transactions on Geoscience and Remote Sensing, 56(11), 6440–6461.
• He, L., Li, J., Liu, C., & Li, S. (2017). Recent advances on spectral–spatial hyperspectral image classification: An overview and new guidelines. IEEE Transactions on Geoscience and Remote Sensing, 56(3), 1579–1597.
• Hong, D., Yokoya, N., Chanussot, J., & Zhu, X. X. (2019). CoSpace: Common subspace learning from hyperspectral-multispectral correspondences. IEEE Transactions on Geoscience and Remote Sensing, 57(7), 4349–4359.
• Hong, D., Yokoya, N., Ge, N., Chanussot, J., & Zhu, X. X. (2019). Learnable manifold alignment (LeMA): A semi-supervised cross-modality learning framework for land cover and land use classification. ISPRS Journal of Photogrammetry and Remote Sensing, 147, 193–205.
• Hu, L., Luo, X., & Wei, Y. (2020). Hyperspectral image classification of convolutional neural network combined with valuable samples. Journal of Physics: Conference Series, 1549(5), 052011. https://doi.org/10.1088/1742-6596/1549/5/052011
• Hu, Q., Xu, W., Liu, X., Cai, Z., & Cai, J. (2021). Hyperspectral image classification based on bilateral filter with multispatial domain. IEEE Geoscience and Remote Sensing Letters, 19, 1–5. https://doi.org/10.1109/LGRS.2021.3058182
• Jia, S., Zhang, X., & Li, Q. (2015). Spectral–Spatial Hyperspectral Image Classification Using l1/2 Regularized Low-Rank Representation and Sparse Representation-Based Graph Cuts. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 8(6), 2473–2484.
• Joshi, A. J., Porikli, F., & Papanikolopoulos, N. (2009). Multi-class active learning for image classification. In 2009 IEEE conference on computer vision and pattern recognition (pp. 2372–2379). IEEE.
• Kang, X., Li, S., & Benediktsson, J. A. (2013). Spectral–spatial hyperspectral image classification with edge-preserving filtering. IEEE Transactions on Geoscience and Remote Sensing, 52(5), 2666–2677.
• Kilik, R. (2021). Histogram-based weighted median filtering used for noise reduction of digital elevation model data. Acta Geodaetica et Geophysica, 56(4), 743–764. https://doi.org/10.1007/s40328-021-00356-2
• Kotwal, K., & Chaudhuri, S. (2010). Visualization of hyperspectral images using bilateral filtering. IEEE Transactions on Geoscience and Remote Sensing, 48(5), 2308–2316.
• Li, J. (2015). Active learning for hyperspectral image classification with a stacked autoencoders based neural network. In 2015 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS) (pp. 1–4). IEEE.
• Liu, P., Zhang, H., & Eom, K. B. (2016). Active deep learning for classification of hyperspectral images. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(2), 712–724.
• Min, D., Choi, S., Lu, J., Ham, B., Sohn, K., & Do, M. N. (2014). Fast global image smoothing based on weighted least squares. IEEE Transactions on Image Processing, 23(12), 5638–5653.
• Santara, A., Mani, K., Hatwar, P., Singh, A., Garg, A., Padia, K., & Mitra, P. (2017). BASS net: Band-adaptive spectral-spatial feature learning neural network for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing, 55(9), 5293–5301.
• Thilagavathi, K., Nagendran, R., & Mary, I. T. B. (2021). Hyperspectral image classification using ensemble average method. In 2021 International Conference on Advancements in Electrical, Electronics, Communication, Computing and Automation (ICAECA) (pp. 1–6). IEEE.
• Wang, Q., Chen, M., Zhang, J., Kang, S., & Wang, Y. (2021). Improved active deep learning for semi-supervised classification of hyperspectral image. Remote Sensing, 14(1), 171. https://doi.org/10.3390/rs14010171
• Wan, X., & Chen, S. (2023). Hyperspectral image classification using improved multi-scale block local binary pattern and bi-exponential edge-preserving smoother. European Journal of Remote Sensing, 56(1), 2237654. https://doi.org/10.1080/22797254.2023.2237654
• Yang, Y., Wu, D., Zeng, L., & Li, Z. (2024). Weighted least square filter via deep unsupervised learning. Multimedia Tools and Applications, 83, 31361–31377. https://doi.org/10.1007/s11042-023-16844-2.
• Zhang, D., Kang, J., Xun, L., & Huang, Y. (2019). Hyperspectral image classification using spatial and edge features based on deep learning. International Journal of Pattern Recognition and Artificial Intelligence, 33(09), 1954027. https://doi.org/10.1142/S0218001419540272
• Zhu, F., Liang, Z., Jia, X., Zhang, L., & Yu, Y. (2019). A benchmark for edge-preserving image smoothing. IEEE Transactions on Image Processing, 28(7), 3556–3570.