Hyperspectral image (HSI) cross-scene classification is a challenging task in remotesensing, particularly when real-time processing of Target Domain (TD) HSI is required, and data cannot be reused for training. While...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
Hyperspectral image (HSI) cross-scene classification is a challenging task in remotesensing, particularly when real-time processing of Target Domain (TD) HSI is required, and data cannot be reused for training. While deep learning methods have shown promising results, the generalization ability of HSI representations remains limited, mainly due to class label imbalance. This paper introduces a dual-stage learning framework based on transfer learning to enhance classification accuracy in the TD. The framework includes a self-supervised learning stage and a supervised fine-tuning stage. The self-supervised stage focuses on learning robust representations by leveraging inherent structures within HSI data, while the fine-tuning stage uses training labels to extract semantic information. A masked diffusion model predicts masked tokens from unmasked ones, capturing both high-level structures and fine details in HSI data. An efficient spatiospectral Transformer, which removes self-attention from the decoder, is proposed to enhance the self-supervised process. This design allows mask tokens to obtain information from visible tokens without interacting with each other, reducing sequence length and computational costs. By decoding each mask token conditionally independently, only a subset of masked tokens is processed. Extensive experiments on two public HSI datasets demonstrate that the proposed method outperforms state-of-the-art techniques.
The Intensity Hue Saturation (IHS) transform is a widely used method to enhance the spatial resolution of multispectral images by substituting the Intensity component by the high resolution of the panchromatic image. ...
详细信息
ISBN:
(纸本)081943826X
The Intensity Hue Saturation (IHS) transform is a widely used method to enhance the spatial resolution of multispectral images by substituting the Intensity component by the high resolution of the panchromatic image. However, such a direct substitution introduces important modifications on spectral properties. A more rigorous approach should consist in enhancing the spatial resolution of the intensity component through an appropriate combination with the panchromatic image. Such a combination is performed in the redundant wavelet domain by using a fusion model. SPOT images are used to illustrate the superiority of our approach compared to the MHS method for preserving spectral properties.
In the last decade, the application of statistical and neural network classifiers to remote-sensingimages has been deeply investigated. Therefore, performances, characteristics, and pros and cons of such classifiers ...
详细信息
ISBN:
(纸本)081943826X
In the last decade, the application of statistical and neural network classifiers to remote-sensingimages has been deeply investigated. Therefore, performances, characteristics, and pros and cons of such classifiers are quite well known, even from remote-sensing practitioners. In this paper, we present the application to remote-sensingimage classification of a new pattern recognition technique recently introduced within the framework of the Statistical Learning Theory developed by V. Vapnik and his co-workers, namely, the Support Vector Machines (SVMs). In section 1, the main theoretical foundations of SVMs are presented. In section 2, experiments carried out on a data set of multisensor remote-sensingimages are described, with particular emphasis on the design and training phase of a SVM. In section 3, the experimental results are reported, together with a comparison between the performances of SVMs, neural network, and k-NN classifiers.
A concept of the restoration problem in the multidimensional imageprocessing is presented. Three algorithms for image restoration when the signal operator has disturbances are discussed. Simulation results prove the ...
详细信息
ISBN:
(纸本)0879426756
A concept of the restoration problem in the multidimensional imageprocessing is presented. Three algorithms for image restoration when the signal operator has disturbances are discussed. Simulation results prove the efficiency of the proposed algorithms.
The problem of identifying terrains in Landsat-TM images on the basis of non-uniformly distributed labeled data is discussed in this paper. Our approach is based on the use of neural network classifiers that learn to ...
详细信息
ISBN:
(纸本)081943826X
The problem of identifying terrains in Landsat-TM images on the basis of non-uniformly distributed labeled data is discussed in this paper. Our approach is based on the use of neural network classifiers that learn to predict posterior class probabilities. Principal Component Analysis (PCA) is used to extract features from spectral and contextual information. The proposed scheme obtains lower error rates that other model-based approaches.
This paper proposes a new regularization algorithm combining the wavelet-based and contourlet-based regularization items based on the Compressive sensing (CS) theorem. The new algorithm aims at gaining maximum benefit...
详细信息
ISBN:
(纸本)9780819483065
This paper proposes a new regularization algorithm combining the wavelet-based and contourlet-based regularization items based on the Compressive sensing (CS) theorem. The new algorithm aims at gaining maximum benefit by combining the multiscale and multiresolution properties common to both wavelet and contourlet schemes, while simultaneously incorporating their individual properties of point singularity and line singularity respectively. CS is applied to remotesensingimage deblurring. It has great practical significance due to saving the hardware cost and aiding fast transmission. Experimental results show the method achieves improvement in peak-signal-noise-ratio and correlation function as compared to traditional regularization algorithms.
We propose a new method of kernel density estimation with a varying adaptive window width. This method is different from traditional ones in two aspects. First, we use symmetric as well as nonsymmetric left and right ...
详细信息
ISBN:
(纸本)081943826X
We propose a new method of kernel density estimation with a varying adaptive window width. This method is different from traditional ones in two aspects. First, we use symmetric as well as nonsymmetric left and right kernels with discontinuities and show that the fusion of these estimates results in accuracy improvement. Second, we develop estimates with adaptive varying window widths based on the so-called intersection of confidence intervals (ICI) rule. Several examples of the proposed method are given for different types of densities and the quality of the adaptive density estimate is assessed by means of numerical simulations.
The problem of selecting an appropriate wavelet filter is always present in signal compression based on the wavelet transform. In this report, we give a method to select a wavelet filter for multispectral image compre...
详细信息
ISBN:
(纸本)081943826X
The problem of selecting an appropriate wavelet filter is always present in signal compression based on the wavelet transform. In this report, we give a method to select a wavelet filter for multispectral image compression. The wavelet filter selection is based on the Learning Vector Quantization (LVQ). In the training phase for the test images, the best wavelet filter has been found by a careful compression-decompression evaluation. Certain spectral features are used in characterizing the pixel spectra. The LVQ is used to form the best wavelet filter class for different types of spectral images. When a new image is to be compressed, a set of spectra from that image is selected, the spectra are classified by the trained LVQ and the filter associated to the largest class is selected for the compression of the whole multispectral image. The results show, that our method finds the most suitable wavelet filter for compression of multispectral images.
Synthetic aperture ladar (SAL) is a newly developed imaging device for remotesensing application. Owing to its short wavelength (3-5 orders of magnitude shorter than radar), SAL is very sensitive to platform vibratio...
详细信息
Synthetic aperture ladar (SAL) is a newly developed imaging device for remotesensing application. Owing to its short wavelength (3-5 orders of magnitude shorter than radar), SAL is very sensitive to platform vibration. For frequency-modulated continuous-wave SAL (FMCW-SAL), the platform vibration induces an additional range cell migration (RCM) to the SAL image. The vibration-induced RCM (vi-RCM) deteriorates the image quality. The vi-RCM is a unique problem for the FMCW-SAL imaging. To address this problem, a raw-data-driven method is proposed to correct the vi-RCM in this paper. First, the signal model was developed to show the vi-RCM in FMCW-SAL echo. Then, based on the model, the differential phase function (DPF) is constructed for the adjacent range profiles. The DPF is a single-frequency signal with its frequency being proportional to the relative range shift between the adjacent range profiles. Based on the DPF, the relative range shift is estimated. After the estimation of all the relative range shifts, the vi-RCM is calculated and corrected. Experiments are performed. The simulated experiment demonstrated the feasibility, accuracy, and efficiency of the proposed method, and the real data processing result verified the effectiveness of the proposed method for FMCW-SAL in practical applications. (C) 2020 Optical Society of America
暂无评论