Spatial and Spectral-Channel Attention Network for Denoising on Hyperspectral Remote Sensing Image

Dou, Hong-Xia; Pan, Xiaomiao; Wang, Chao; Shen, Hao-Zhen; Deng, Liang-Jian

doi:10.3390/rs14143338

Cited by 10 publications

(5 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, due to the continuity and universality of the distribution of hyperspectral objects, self-similarity is an inherent property of hyperspectral images, which has been widely used in many popular denoising methods. At present, the development trend of spectral denoising methods is to combine the spatial and spectral similarity of images to improve the quality of images [28,29]. In [30], the authors proposed a low-rank restoration method that combines spatial and spectral information for image denoising, which simultaneously embeds the TV regularization, nuclear norm, and L 1 norm (LRTV).…”

Section: Optimization-based Methodsmentioning

confidence: 99%

Memory Augmentation and Non-Local Spectral Attention for Hyperspectral Denoising

Dong,

Mo,

Sun

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

In this paper, a novel hyperspectral denoising method is proposed, aiming at restoring clean images from images disturbed by complex noise. Previous denoising methods have mostly focused on exploring the spatial and spectral correlations of hyperspectral data. The performances of these methods are often limited by the effective information of the neighboring bands of the image patches in the spectral dimension, as the neighboring bands often suffer from similar noise interference. On the contrary, this study designed a cross-band non-local attention module with the aim of finding the optimal similar band for the input band. To avoid being limited to neighboring bands, this study also set up a memory library that can remember the detailed information of each input band during denoising training, fully learning the spectral information of the data. In addition, we use dense connected module to extract multi-scale spatial information from images separately. The proposed network is validated on both synthetic and real data. Compared with other recent hyperspectral denoising methods, the proposed method not only demonstrates good performance but also achieves better generalization.

show abstract

Section: Optimization-based Methodsmentioning

confidence: 99%

Memory Augmentation and Non-Local Spectral Attention for Hyperspectral Denoising

Dong,

Mo,

Sun

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Some previous studies on remote sensing images using CNN have demonstrated differences in the weights of different bands of an image across various application scenarios. The utilization of attentional mechanisms allows for improved focus on these weights [44,45]. Given that various bands or polarization modes in remote sensing data may respond differently to changes in water level, the incorporation of these attention mechanisms into the models could be a valuable addition.…”

Section: Model Architecturementioning

confidence: 99%

Combining Satellite Imagery and a Deep Learning Algorithm to Retrieve the Water Levels of Small Reservoirs

Wu,

Huang,

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

There are an estimated 800,000 small reservoirs globally with a range of uses. Given the collective importance of these reservoirs to water resource management and wider society, it is essential that we can monitor and understand the hydrological dynamics of ungauged reservoirs, particularly in a changing climate. However, unlike large reservoirs, continuous and systematic hydrological observations of small reservoirs are often unavailable. In response, this study has developed a retrieval framework for water levels of small reservoirs using a deep learning algorithm and remotely sensed satellite data. Demonstrated at four reservoirs in California, satellite imagery from both Sentinel-1 and Sentinel-2 along with corresponding water level field measurements was collected. Post-processed images were fed into a water level inversion convolutional neural network model for water level inversion, while different combinations of these satellite images, sampling approaches for training/testing data, and attention modules were used to train the model and evaluated for accuracy. The results show that random sampling of training data coupled with Sentinel-2 satellite imagery was generally the most accurate initially. Performance is improved by incorporating a channel attention mechanism, with the average R2 increasing by 8.6% and the average RMSE and MAE decreasing by 15.5% and 36.4%, respectively. The proposed framework was further validated on three additional reservoirs in different regions. In conclusion, the retrieval framework proposed in this study provides a stable and accurate methodology for water level estimation of small reservoirs and can be a powerful tool for small reservoir monitoring over large spatial scales.

show abstract

“…Yulong Guo et al [16] presented a sparse representation-based denoising method for water color hyperspectral data, adeptly separating signal from noise. Hongxia Dou et al [17] utilized a 3D U-Net architecture with SSCA modules, taking into account both spatial and spectral features of hyperspectral images to enhance denoising performance.Lintao Han [18] introduced the RSIDNet deep learning network, which bolsters denoising capabilities through multi-scale feature extraction and attention mechanisms. Masoud Moradi [19] determined the optimal wavelet functions for ocean color time series data.…”

Section: Introductionmentioning

confidence: 99%

Parallel Acceleration Algorithm for Wavelet Denoising of UAVAGS Data Based on CUDA

Xiong,

Wang,

Qiao

et al. 2024

Preprint

View full text Add to dashboard Cite

The computational efficiency is low when the vast volume of unmanned aerial vehicle airborne gamma-ray spectrum (UAVAGS) data is handled by wavelet denoising in CPU. So, a CUDA-based GPU parallel solution is recommended to resolve this issue in this paper. This proposed solution aims to significantly enhance the efficiency of parallel acceleration for wavelet denoising of UAVAGS data. In the preliminary stage, experiments were conducted with varying block sizes to investigate the influence of different block sizes on processing time. The objective was to identify the most suitable block size for efficiently processing UAVAGS data. Subsequently, a performance evaluation was conducted by comparing the acceleration ratios of GPU and CPU for different data volumes, as well as varying wavelet basis functions under the same data volume conditions. Finally, by intentionally introducing noise, calculations were performed to determine the optimal wavelet basis function concerning signal-to-noise ratio after denoising. The research findings indicate that the optimal two-dimensional block size falls within the range of 64×64 to 128×128. The majority of wavelet basis functions achieved acceleration ratios exceeding 100-fold in total processing time, with the coif5 wavelet basis function reaching an acceleration ratio of 185-fold. Comparative analysis of various denoising functions revealed that, under low signal-to-noise ratios, these functions exhibited insufficient denoising effects, while at high signal-to-noise ratios, there was a risk of excessive denoising. However, significant denoising effects were observed when employing hard thresholding with coif5, soft thresholding, and an improved thresholding method with db3.

show abstract

Spatial and Spectral-Channel Attention Network for Denoising on Hyperspectral Remote Sensing Image

Cited by 10 publications

References 66 publications

Memory Augmentation and Non-Local Spectral Attention for Hyperspectral Denoising

Memory Augmentation and Non-Local Spectral Attention for Hyperspectral Denoising

Combining Satellite Imagery and a Deep Learning Algorithm to Retrieve the Water Levels of Small Reservoirs

Parallel Acceleration Algorithm for Wavelet Denoising of UAVAGS Data Based on CUDA

Contact Info

Product

Resources

About