FPGA-Based Implementation of a CNN Architecture for the On-Board Processing of Very High-Resolution Remote Sensing Images

Neris, Romen; Rodriguez, Adrian Lopez; Guerra, Raúl; López, Sebastián; Sarmiento, Roberto

doi:10.1109/jstars.2022.3169330

Cited by 11 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Reconfigurable computing, often known as FPGA-based computing, has recently gained popularity for implementing algorithms suited for RS applications [8][9] for accelerated performance. FPGAs exemplify an evolution over the Application-Specific Integrated Circuits (ASICs) [10] in the sense that they are explicitly designed to solve distinct problems.…”

Section: Specialized Hardware Architectures For Hpc-pmentioning

confidence: 99%

Exploring the Advancements in High-Performance Computing Paradigm for Remote Sensing Big Data Analytics

SK,

Aji S

2023

Cloud Computing and Data Science

View full text Add to dashboard Cite

The incredible growth in Remote Sensing (RS) data volume, with high spectral-spatial-temporal resolutions, has been utilized in various application domains. With the rapid advancements in modern sensor technologies, including the 3D acquisition sensors, RS data with a large variety, velocity, veracity, varied value and incredible volume are generated, leading to the Remote Sensing Big Data (RSBD). With the high availability of RSBD, we require High-Performance Computing (HPC) environments for storing and processing these High-Dimensional (HD), complex, heterogeneous and distributed data. Also, introducing Deep Learning (DL) techniques in the RS domain demands more computing power, higher memory and networking bandwidth throughput capabilities, and optimized software and libraries to deliver the required performance. Motivated by this, we explore HPC computing environments for handling RSBD across multiple application domains in this paper. With particular emphasis on architectures such as cloud-based HPC, clusters, heterogeneous networks of computers, and specialized hardware architectures like Field Programmable Gate Arrays (FPGAs) and Graphics Processing Units (GPUs), we investigate how HPC technologies are being used to process RSBD efficiently while including integrated intelligence. This critical analysis results in a multi-layered cloud-based framework for efficient RSBD processing tasks. Also, we identified several data challenges to be handled while designing HPC frameworks. The findings from the study can help researchers better understand the HPC design concepts for developing RSBD frameworks.

show abstract

Section: Specialized Hardware Architectures For Hpc-pmentioning

confidence: 99%

Exploring the Advancements in High-Performance Computing Paradigm for Remote Sensing Big Data Analytics

SK,

Aji S

2023

Cloud Computing and Data Science

View full text Add to dashboard Cite

show abstract

“…Onboard processing of complex problems with low energy consumption improves efficiency, accuracy, and performance [38]. Implementation of convolution neural network (CNN) on FPGAs for onboard processing of high-resolution satellite images requires less computational power and low memory resources [39]. The artificial neural network (ANN) model implemented on FPGA can also be used to extract the features from an input image and to encrypt the same image using asymmetric encryption algorithms [40].…”

Section: Introductionmentioning

confidence: 99%

Comparative study on satellite based image encryption methods: a survey

Vasudevaiah,

Shivaswamy

2024

IJECE

View full text Add to dashboard Cite

The availability of high-resolution satellite images increases with advancements in remote sensing technology. These satellite images are used in various earth observation applications such as disaster management, military applications, weather forecasting, land use and cover, and many more. Satellite images have large volumes stored in memory devices. These satellite images are transmitted to the ground station for processing and analysis. In these cases, images are vulnerable to privacy issues. As technology advances, onboard processing of satellite images using intelligent systems processes the images faster. A model such as field programmable gate arrays (FPGA) is used in onboard processing to process satellite images. However, images are susceptible to faults induced by harsh radiation environments in space. Encryption is one of the most assured methods to provide privacy to satellite images. Hence, encryption of satellite images during processing, storage, and transmission is the present rising demand. There are various encryption methods implemented using algorithms such as advanced encryption standard (AES), homomorphic, advanced encryption standard-counter (AES-CTR), and chaotic maps. Concurrent processing and encryption of images using MapReduce with Hadoop Framework perform the task faster. The focus of this paper is a comparative study of the various encryption methods used in recent years.

show abstract

“…Researchers have primarily focused on deploying CNN models on FPGA platforms. Neris et al [19] conducted a comparative analysis of commonly used remote sensing image processing models and determined the MobileNet1Lite model to be the most well-suited for FPGA deployment. They developed a MobileNet1Lite accelerator on the FPGA platform using high-level synthesis (HLS) technology, implementing both 32-bit floating-point and 16-bit fixed-point precision.…”

Section: Introductionmentioning

confidence: 99%

Algorithm–Hardware Co-Optimization and Deployment Method for Field-Programmable Gate-Array-Based Convolutional Neural Network Remote Sensing Image Processing

Ni,

Wei,

Zhang

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

In recent years, convolutional neural networks (CNNs) have gained widespread adoption in remote sensing image processing. Deploying CNN-based algorithms on satellite edge devices can alleviate the strain on data downlinks. However, CNN algorithms present challenges due to their large parameter count and high computational requirements, which conflict with the satellite platforms’ low power consumption and high real-time requirements. Moreover, remote sensing image processing tasks are diverse, requiring the platform to accommodate various network structures. To address these issues, this paper proposes an algorithm–hardware co-optimization and deployment method for FPGA-based CNN remote sensing image processing. Firstly, a series of hardware-centric model optimization techniques are proposed, including operator fusion and depth-first mapping technology, to minimize the resource overhead of CNN models. Furthermore, a versatile hardware accelerator is proposed to accelerate a wide range of commonly used CNN models after optimization. The accelerator architecture mainly consists of a parallel configurable network processing unit and a multi-level storage structure, enabling the processing of optimized networks with high throughput and low power consumption. To verify the superiority of our method, the introduced accelerator was deployed on an AMD-Xilinx VC709 evaluation board, on which the improved YOLOv2, VGG-16, and ResNet-34 networks were deployed. Experiments show that the power consumption of the accelerator is 14.97 W, and the throughput of the three networks reaches 386.74 giga operations per second (GOPS), 344.44 GOPS, and 182.34 GOPS, respectively. Comparison with related work demonstrates that the co-optimization and deployment method can accelerate remote sensing image processing CNN models and is suitable for applications in satellite edge devices.

show abstract

FPGA-Based Implementation of a CNN Architecture for the On-Board Processing of Very High-Resolution Remote Sensing Images

Cited by 11 publications

References 21 publications

Exploring the Advancements in High-Performance Computing Paradigm for Remote Sensing Big Data Analytics

Exploring the Advancements in High-Performance Computing Paradigm for Remote Sensing Big Data Analytics

Comparative study on satellite based image encryption methods: a survey

Algorithm–Hardware Co-Optimization and Deployment Method for Field-Programmable Gate-Array-Based Convolutional Neural Network Remote Sensing Image Processing

Contact Info

Product

Resources

About