Cloud removal from satellite imagery using multispectral edge-filtered conditional generative adversarial networks

Hasan, Cengis; Horne, Ross; Mauw, Sjouke; Mizera, Andrzej

doi:10.1080/01431161.2022.2048915

Cited by 10 publications

(9 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To use satellite images acquired in every season, more images with higher cloud cover should be included. Consequently, cloud removal in the preprocessing step must be done to lower the risk of loss of information leading to a spatiotemporal discontinuity that degrades the quality and usefulness of satellite images (Hasan et al, 2022).…”

Section: Discussionmentioning

confidence: 99%

Identifying Climate Change Vulnerability Based on Land Cover Indicators: A Case Study in Surabaya, Indonesia

BINARTI¹,

SANTOSO²

2022

View full text Add to dashboard Cite

Surabaya is facing the threat of climate change indicated by the increase in air and surface temperature. The city has a risk of sinking by 2050 if the global warming cannot be resolved. Several related studies established that the change in land cover and land use is accompanied by the increase in surface temperature, which will be addressed in this present study. Therefore, this study aimed to examine the impact of land use/cover on the increase in air/surface temperature and investigate the contribution of land cover indicators to climate change in Surabaya as the basis of the identification of spatial climate change vulnerability. Data were collected from satellite images obtained over a long period and processed with GIS-based software to obtain an overview of changes. Mined long-term historical climate data and satellite imagery were processed into a land surface temperature map (LST), describing the tendency of climate change. The satellite imagery data from 2013 to 2021 was used to have an overview of land use and land cover changes based on indicators of built-up area (NDBI), surface imperviousness (NDISI), vegetation (NDVI), and water (NDWI). The contribution of each indicator to the surface temperature was analyzed using the multivariate regression method. The significant contribution of the land cover indicators to the surface temperature as the results means that NDBI, NDISI, NDVI, and NDWI can be used as indicators in climate change vulnerability assessment. The sequential contribution weights to the surface temperature are NDISI, NDWI, NDVI, and NDBI. Furthermore, the climate change vulnerability map of Surabaya City was developed based on the contribution weights, which the pattern of vulnerability levels corresponds to the pattern of water index values.

show abstract

Section: Discussionmentioning

confidence: 99%

Identifying Climate Change Vulnerability Based on Land Cover Indicators: A Case Study in Surabaya, Indonesia

BINARTI¹,

SANTOSO²

2022

View full text Add to dashboard Cite

show abstract

“…To verify the effectiveness of the proposed method, we evaluate the performance of GAN‐UD on the RICE1 dataset [33] and Paris dataset [10]. In addition, we extend GAN‐UD to the NIR band, and since the RICE1 dataset does not contain the NIR band, we choose the Sentinel‐2A dataset provided by the CR‐GAN‐PM [39] method for our experiments.…”

Section: Methodsmentioning

confidence: 99%

“…Paris dataset was obtained from WorldView‐2 by Hasan et al. [10] and contains 4043 images with a resolution of 256 × 256. We use the literature nine method to synthesize the cloudy image

{I}_x

.…”

Section: Methodsmentioning

confidence: 99%

“…Recently, deep learning has been introduced into thin cloud removal methods. Deep learning‐based cloud removal methods learn the mapping function between the remotely sensed cloud images and the cloud‐free images from a large number of training samples with the powerful feature representation capability of convolutional neural networks, and this class of methods greatly improves the performance of remote sensing image cloud removal [6–11, 37, 38]. However, most deep learning‐based cloud removal methods are based on supervised learning.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised remote sensing image thin cloud removal method based on contrastive learning

Tan,

Du,

Man

et al. 2024

IET Image Processing

View full text Add to dashboard Cite

Cloud removal algorithm is a crucial step of remote sensing image preprocessing. The current mainstream remote sensing image cloud removal algorithms are implemented based on deep learning, and most of them are supervised. A large number of data pairs are required for training to achieve cloud removal. However, real with/without cloud image pairs datasets are difficult to obtain in the real world, and the models obtained by training on synthetic datasets often need to generalize better to natural scenes. And the existing unsupervised thin cloud removal methods based on Cycle‐GAN framework with considerable model complexity and unstable training are not an excellent solution to the problem of lack of paired datasets. Based on this, in this paper, the authors propose an unsupervised remote sensing image thin cloud removal method based on contrastive learning—GAN‐UD. It is a network consisting of a frequency‐spatial attention generator and a discriminator. In addition, the authors introduce local contrastive loss and global content loss to constrain the content of the generated images to ensure that the generated cloud‐free images are consistent with the input cloud images in terms of image content. Experimental results show that the proposed method in this paper can still effectively remove thin clouds from remote sensing images without paired training datasets, outperforms current unsupervised cloud removal methods, and achieves comparable performance to supervised methods.

show abstract

“…As a complement, it is necessary to test the algorithm's performance in more heterogeneous areas such as urban areas. We conduct comparative experiments on the Paris dataset proposed by [64]. The cloud-free images are obtained from WorldView-2 European Cities.…”

Section: Additional Datasetmentioning

confidence: 99%

Cloudformer: A Cloud-Removal Network Combining Self-Attention Mechanism and Convolution

Pan

Tang

et al. 2022

Remote Sensing

View full text Add to dashboard Cite

Optical remote-sensing images have a wide range of applications, but they are often obscured by clouds, which affects subsequent analysis. Therefore, cloud removal becomes a necessary preprocessing step. In this paper, a novel and superior transformer-based network is proposed, named Cloudformer. The proposed method novelly combines the advantages of convolution and a self-attention mechanism: it uses convolution layers to extract simple features over a small range in the shallow layer, and exerts the advantage of a self-attention mechanism in extracting correlation in a large range in the deep layer. This method also introduces Locally-enhanced Positional Encoding (LePE) to flexibly generate suitable positional encodings for different inputs and to utilize local information to enhance encoding capabilities. Exhaustive experiments on public datasets demonstrate the superior ability of the method to remove both thin and thick clouds, and the effectiveness of the proposed modules is validated by ablation studies.

show abstract

Cloud removal from satellite imagery using multispectral edge-filtered conditional generative adversarial networks

Cited by 10 publications

References 13 publications

Identifying Climate Change Vulnerability Based on Land Cover Indicators: A Case Study in Surabaya, Indonesia

Identifying Climate Change Vulnerability Based on Land Cover Indicators: A Case Study in Surabaya, Indonesia

Unsupervised remote sensing image thin cloud removal method based on contrastive learning

Cloudformer: A Cloud-Removal Network Combining Self-Attention Mechanism and Convolution

Contact Info

Product

Resources

About