High-resolution satellite scene classification using a sparse coding based multiple feature combination

By Guofeng Sheng, Wen Yang, Tao Xu, Hong Sun. 2012

305 citations recorded by the source corpus.

Subfields: Atmospheric Science, Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atmospheric Science

1. Liangpei Zhang
2. Qian Du
3. Wei Li
4. Yanfei Zhong
5. Cheng Chen
6. Hongjun Su
7. Xiaoqiang Lu
8. Gui-Song Xia

In Computer Vision and Pattern Recognition

1. Liangpei Zhang
2. Junwei Han
3. Gong Cheng
4. Yanfei Zhong
5. Xiaoqiang Lu
6. Gui-Song Xia
7. Xuelong Li
8. Qian Du

In Media Technology

1. Liangpei Zhang
2. Yanfei Zhong
3. Xiaoqiang Lu
4. Qian Du
5. Wei Li
6. Gui-Song Xia
7. Gong Cheng
8. Junwei Han

Citations in the corpus

• EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification (2017, background): “dataset and outperformed previous results on the other three presented datasets by about 2%–4% (AID, SAT-6, BCS) [6], [19], [22].”
• A Review on Deep Learning in UAV Remote Sensing (2021, background): “Still, there are some known aerial repositories with labeled examples, which were presented in recent years, such as the DOTA [197], UAVDT [50], VisDrone [9], WHU-RS19 [171], RSSCN7 [220], RSC11 [209], Brazilian Co ff ee Scene [151] datasets.”
• Land-use scene classification using multi-scale completed local binary patterns (2016, methodology): “6 SIFT + LTP-HF + color histogram [24] 93.”
• Remote Sensing Image Scene Classification Using Multi-Scale Completed Local Binary Patterns and Fisher Vectors (2016, methodology): “The same experimental setup in [30] was used.”
• Training Deep Convolutional Neural Networks for Land–Cover Classification of High-Resolution Imagery (2017, background): “4% classification accuracy and the current state of the art is [21] with 93.”
• Self-Supervised Remote Sensing Feature Learning: Learning Paradigms, Challenges, and Future Works (2022, unknown): “[17], [18], [19], [20], autoencoder [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], and deep belief networks [31].”
• Feature significance-based multibag-of-visual-words model for remote sensing image scene classification (2016, methodology): “For the entire dataset 1, the highest accuracy is obtained by the method based on sparse codes of multiple features (SCMF).(2) The SCMF method utilizes an improved sparse coding-based SPM method for feature representation, which is much better than the traditional BOVW model, i.”
• Introducing Eurosat: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification (2018, methodology): “The deep CNNs achieve state-of-the-art results on the UCM dataset [2] and outperform previous results on the three datasets AID, SAT-6 and BCS by about 2-4% [4, 9, 10].”
• Transferring Deep Convolutional Neural Networks for the Scene Classification of High-Resolution Remote Sensing Imagery (2015, methodology): “We randomly select samples of each class for training the SVM classifier and the rest for testing, following the same sampling setting as [5,54] for the two datasets, respectively: 80 training samples per class for the UCM dataset and 30 training samples per class for the WHU-RS dataset.”
• PatternNet: A Benchmark Dataset for Performance Evaluation of Remote Sensing Image Retrieval (2017, methodology): “The images in RSD have a wide range of spatial resolutions which are up to 0.5m. RSSCN7 dataset (https://www.dropbox.com/s/j80iv1a0mvhonsa/RSSCN7.zip?dl=0).”
• MARTA GANs: Unsupervised Representation Learning for Remote Sensing Image Classification (2016, methodology): “As shown in Table II, our method outperforms the SCMF [14] (a sparse coding based multiple-feature fusion method) by 3.”
• A Comparative Study of Sampling Analysis in the Scene Classification of Optical High-Spatial Resolution Remote Sensing Imagery (2015, background): “Observe that, in order to do scene classification for a large-scale HSR-RS image, one often requires to first construct the basic scene elements for classification, by using either non-overlapping tiles [2,5,13], overlapping tiles [10,14,16] or super-pixels [4,11].”
• Fusion of Deep Convolutional Neural Networks for Land Cover Classification of High-Resolution Imagery (2017, methodology): “Chen et al. [20] achieved a 93.4% accuracy using their local binary patterns and extreme learning machine technique, while Sheng et al. [21] obtained 94.6% average classification accuracy.”