This paper proposes a Cascaded Random Forest (CRF) method, which can improve the classification performance by means of combining two different enhancements into the Random Forest (RF) algorithm. In detail, on the one hand, a neighborhood rough sets based Hierarchical Random Subspace Method is designed for feature selection, which can improve the strength of base classifiers and increase the diversity between each two of the base classifiers; and on the other hand, Boosting is introduced into RF. As the minimization of the training error for updating the weights of samples in Boosting often leads to overfitting, we added out-of-bag error to update the sample weights in CRF. Different from the existing Boosting strategy that only one base classifier is generated at each iteration, the proposed CRF trains several base classifiers at each iteration. To evaluate the performance of our method, CRF is compared with other related RF methods and support vector machine on three benchmark hyperspectral datasets, and experimental results show that CRF can provide competitive solutions for hyperspectral image classification.

Random forest (RF) has obtained great success in hyperspectral image (HSI) classification. However, RF cannot leverage its full potential in the case of limited labeled samples. To address this issue, we propose a unified framework that embeds active learning (AL) and semi-supervised learning (SSL) into RF (ASSRF). Our aim is to utilize AL and SSL simultaneously to improve the performance of RF. The objective of the proposed method is to use a small number of manually labeled samples to train classifiers with relative high classification accuracy. To achieve this goal, a new query function is designed to query the most informative samples for manual labeling, and a new pseudolabeling strategy is introduced to select some samples for pseudolabeling. Compared with other AL- and SSL-based methods, the proposed method has several advantages. First, ASSRF utilizes the spatial information to construct a query function for AL, which can select more informative samples. Second, in addition to providing more labeled samples for SSL, the proposed pseudolabeling method avoids bias caused by AL-labeled samples. Finally, the proposed model retains the advantages of RF. To demonstrate the effectiveness of ASSRF, we conducted experiments on three real hyperspectral data sets. The experimental results have shown that our proposed method outperforms other state-of-the-art methods.

In the hyperspectral image (HSI) classification, there are numerous deep learning-based research routes that have emerged recently. Among them, two methodologies attract our attention. One is CNN-based classification and the other is transformer-based classification. The essence of these two methodologies is to interchange information locally or at a long distance for HSI pixels in the spatial or spectral-spatial domain. There are two principles underlying this essence—the information mixing mechanism and the information mixing domain. Although both CNN-based and transformer-based have made efforts in these two principles and obtained favorable classification results, there is still room for improvement in terms of accuracy and efficiency. To further enhance the accuracy and efficiency under the two principles, fast Fourier transform (FFT) is introduced to HSI classification and a fast Fourier filter is designed to mix information efficiently in the frequency domain by means of FFT. The parametric-free characteristic and fast computation of FFT can assist us in efficiently learning interactions among features in the frequency domain. Furthermore, a fast Fourier filter block is built upon the fast Fourier filter for repeatedly using as a basic block. In addition, we propose a spectral-spatial convolution tokenizer (SSCT) to extract shallow features and prepare spectral-spatial tokens for fast Fourier filter blocks. Finally, by employing SSCT and fast Fourier filter blocks, a novel deep neural network architecture—fast Fourier filter network (F ³ Net) is proposed for HSI classification. F ³ Net-P as a pyramidal variant of F ³ Net is also investigated. Experimental results on four datasets comprehensively evaluate our models and indicate that they are competitive with several current state-of-the-art methods, especially when the training samples are extremely limited. Specifically, F ³ Net-P achieves the highest accuracy of 97.25%, 98.08%, 97.49% and 97.95% on the four datasets, respectively, outperforming second best compared model by 1.49%, 2.03%, 2.14% and 1.94%.

Convolutional neural networks (CNNs) have dominated the hyperspectral image (HSI) classification due to their tremendous feature learning capability. However, the formidable local sensitivity is both a strength and a weakness. Recently, the vision transformers have exhibited impressive performances on various vision problems. Compared with CNNs, they can model long-range dependencies to learn more abundant interactions between spatial locations. Nevertheless, the existing transformer-based HSI classification methods also concentrate too much on the advantages of the transformer architecture and disregard the importance of local dependencies. In addition, token generation and token mixers in transformer-like architectures have not been adequately explored, leading to difficulties in obtaining the best classification performance. To deal with these problems, a novel multiscale hierarchical conv-aided Fourierformer (MHCFormer) is proposed for HSI classification. To the best of our knowledge, this is the first time that CNN, transformer, and Fourier transform are skillfully combined for HSI classification. The proposed MHCFormer involves three stages, i.e., multiscale spectral–spatial token generation, hierarchical token learning, and a classification head. The multiscale spectral–spatial token generation is constructed to transform HSI into tokens with multiscale-enhanced spectral–spatial information. The hierarchical token learning is designed to explore multiscale tokens globally and locally by integrating the design philosophy of transformers and CNNs along with Fourier transforms into a block and stacking the blocks hierarchically. Extensive experimental results on the new WHU-Hi-HanChuan dataset and the widely used Indian Pines and Houston 2013 datasets have demonstrated the superiority of MHCFormer over other state-of-the-art methods. The code of our work will be available publicly at https://github.com/Tikiten/MHCFormer .

Deep-learning methods, especially convolutional neural networks (CNN), have become the first choice for hyperspectral image (HSI) classification to date. It is a common procedure that small cubes are cropped from hyperspectral images and then fed into CNNs. However, standard CNNs find it difficult to extract discriminative spectral–spatial features. How to obtain finer spectral–spatial features to improve the classification performance is now a hot topic of research. In this regard, the attention mechanism, which has achieved excellent performance in other computer vision, holds the exciting prospect. In this paper, we propose a double-branch network consisting of a novel convolution named pyramidal convolution (PyConv) and an iterative attention mechanism. Each branch concentrates on exploiting spectral or spatial features with different PyConvs, supplemented by the attention module for refining the feature map. Experimental results demonstrate that our model can yield competitive performance compared to other state-of-the-art models.