期刊
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
卷 14, 期 11, 页码 1928-1932出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/LGRS.2017.2737823
关键词
Deep learning; hyperspectral image (HSI) classification; recursive autoencoders (RAE); unsupervised feature learning
类别
资金
- National Natural Science Foundation of China [61772400, 61377011, 61373111]
- Program for New Scientific and Technological Star of Shaanxi Province [2014KJXX-45]
- UK EPSRC [EP/N508664/1, EP/R007187/1, EP/N011074/1]
- Royal Society-Newton Advanced Fellowship [NA160342]
- EPSRC [EP/N011074/1, EP/N508664/1, EP/R007187/1] Funding Source: UKRI
- Engineering and Physical Sciences Research Council [EP/N011074/1, EP/R007187/1, EP/N508664/1] Funding Source: researchfish
For hyperspectral image (HSI) classification, it is very important to learn effective features for the discrimination purpose. Meanwhile, the ability to combine spectral and spatial information together in a deep level is also important for feature learning. In this letter, we propose an unsupervised feature learning method for HSI classification, which is based on recursive autoencoders (RAE) network. RAE utilizes the spatial and spectral information and produces high-level features from the original data. It learns features from the neighborhood of the investigated pixel to represent the whole local homogeneous area of the image. In addition, to obtain more accurate representation of the investigated pixel, a weighting scheme is adopted based on the neighboring pixels, where the weights are determined by the spectral similarity between the neighboring pixels and the investigated pixel. The effectiveness of our method is evaluated by the experiments on two hyperspectral data sets, and the results show that our proposed method has a better performance.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据