Distributed Deep Learning for Remote Sensing Data Interpretation

Haut, Juan M.; Paoletti, Mercedes E.; Moreno‐Álvarez, Sergio; Plaza, Javier; Rico‐Gallego, Juan‐Antonio

doi:10.1109/jproc.2021.3063258

Cited by 27 publications

(12 citation statements)

References 132 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, the unchanged pixels are firstly searched and selected to prepare for weakly supervised learning. Here, CVA is applied to obtain a twoclass saliency map in which spectral samples with a higher probability of belonging to unchanged pixels are selected as pseudo-labels that can be expressed as U= {(u 1 , p 1 ) , ..., (u i , p i ) , ..., (u n , p n )} (7) where u i is the unchanged spectral sample with p i =0 and n is the number of training spectral samples. The pseudo-labels are not always referenced.…”

Section: Change Discriminationmentioning

confidence: 99%

“…To tackle this problem, parallel computing and cloud computing with distributed structures have received great attention. In practice, distributed cloud computing has become a natural solution for processing remote sensing HSIs by virtue of their powerful processing performance and comprehensive resources [6], [7]. However, the existing distributed structures are inconvenient to be applied for HCD that depends on multi-temporal HSIs, due to the following limitations: (1) high cost of transferring large volumes of raw HSIs between multiple times, (2) no guarantee of raw data privacy and security, (3) information loss in the process of uploading from the data source to the cloud, and (4) unsuitability for handling non-independent identical distributed data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Decentralized Federated GAN for Hyperspectral Change Detection in Edge Computing

Xie,

Xu,

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Change detection on hyperspectral images is an essential task for Earth observation. Due to the vast amounts of remote sensing big data resulting from the ongoing advancements in remote sensing hardware, employing centralized learning through cloud computing emerges as a logical and convenient solution. Nevertheless, this approach overlooks the influence of the isolation and heterogeneity of remote sensing data on the reliability of change detection outcomes. In contrast, federated learning enables collaborative change detection on non-independent identical distributed remote sensing data without the need to transfer the original data. Simultaneously, it is important to acknowledge that the dependency of federated learning on the central node may pose potential data security risks. To address this issue, this paper proposes a decentralized federated learning GAN network. This ensures that raw data remains stationary during participation in unsupervised learning. Additionally, the network employs edge devices to implement federated learning, allowing adaptation to diverse devices in practical scenarios. Lastly, blockchain is integrated into a decentralized architecture for the dynamic selection of leader nodes, effectively enhancing the robustness and security of the framework. DFGAN is a novel approach for cooperative privacy preservation introduced into hyperspectral change detection systems. The method outperforms most of existing methods on three datasets, achieving more than 90% detection accuracy on all of them. Additionally, by simulating in a real on-orbit environment using a satellite constellation simulator on a Jeston TX2, FedGAN demonstrates superior accuracy in hyperspectral image change compared to existing algorithms while significantly reducing training time by 17.3%.

show abstract

Section: Change Discriminationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Decentralized Federated GAN for Hyperspectral Change Detection in Edge Computing

Xie,

Xu,

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In addition, the client also supports various application functions to provide solutions for users' problems when browsing the website. In addition, the client also has modules that can help administrators to achieve information release, management and other work tasks [13][14]. Second, it can connect to the server remotely, with high security.…”

Section: Characteristicsmentioning

confidence: 99%

Development of Distributed Remote Processing System Based on Net Remoting Technology

2022

DPS

View full text Add to dashboard Cite

With the development of network technology, multimedia teaching has become an indispensable part of today's education, and remting is also a new type of remote processing server. The system is designed and implemented with java development framework and JS architecture. This topic mainly introduces the function modules such as management of simulation interactive single machine training center, database management and transaction logic engine based on net platform, and analyzes and explains them respectively. At the same time, the front-end page is designed and built in detail, including the scheme of web page layout and data access control and the application of related technologies. After that, this paper tests the operation and processing of the distributed remote processing system. The test results show that the IP address information is sent to the client through the serial port for interactive operation. At the same time, it can also receive the return command from the terminal device to support the normal 5-10S time control of the remote processing, and complete the analysis and test of the performance and function.

show abstract

“…The size of RS images is typically large as they are captured by toptier camera devices with multiple bands or layers. As the RS data are produced daily [8] in the era of big data [9,10] , the order of magnitude of the data grows to terabyte [11] or even petabyte [12] . As a result, the Hadoop Distributed File System (HDFS) is often utilised as storage for the RS data [13][14][15][16] .…”

Section: Introductionmentioning

confidence: 99%

MIX-RS: A Multi-Indexing System Based on HDFS for Remote Sensing Data Storage

Xiong

Dai

et al. 2022

Tsinghua Sci. Technol.

View full text Add to dashboard Cite

A large volume of Remote Sensing (RS) data has been generated with the deployment of satellite technologies. The data facilitate research in ecological monitoring, land management and desertification, etc.The characteristics of RS data (e.g., enormous volume, large single-file size, and demanding requirement of fault tolerance) make the Hadoop Distributed File System (HDFS) an ideal choice for RS data storage as it is efficient, scalable, and equipped with a data replication mechanism for failure resilience. To use RS data, one of the most important techniques is geospatial indexing. However, the large data volume makes it time-consuming to efficiently construct and leverage. Considering that most modern geospatial data centres are equipped with HDFS-based big data processing infrastructures, deploying multiple geospatial indices becomes natural to optimise the efficacy.Moreover, because of the reliability introduced by high-quality hardware and the infrequently modified property of the RS data, the use of multi-indexing will not cause large overhead. Therefore, we design a framework called Multi-IndeXing-RS (MIX-RS) that unifies the multi-indexing mechanism on top of the HDFS with data replication enabled for both fault tolerance and geospatial indexing efficiency. Given the fault tolerance provided by the HDFS, RS data are structurally stored inside for faster geospatial indexing. Additionally, multi-indexing enhances efficiency.The proposed technique naturally sits on top of the HDFS to form a holistic framework without incurring severe overhead or sophisticated system implementation efforts. The MIX-RS framework is implemented and evaluated using real remote sensing data provided by the Chinese Academy of Sciences, demonstrating excellent geospatial indexing performance.

show abstract

Distributed Deep Learning for Remote Sensing Data Interpretation

Cited by 27 publications

References 132 publications

Decentralized Federated GAN for Hyperspectral Change Detection in Edge Computing

Decentralized Federated GAN for Hyperspectral Change Detection in Edge Computing

Development of Distributed Remote Processing System Based on Net Remoting Technology

MIX-RS: A Multi-Indexing System Based on HDFS for Remote Sensing Data Storage

Contact Info

Product

Resources

About