On-Orbit Real-Time Variational Image Destriping: FPGA Architecture and Implementation

Chen, Liqun; Chang, Yi; Yan, Luxin

doi:10.1109/tgrs.2022.3140428

Cited by 6 publications

(3 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Striping noise is an important topic for line array detector 2 . According to Equation (3), the striping noise comes from two ways [3][4][5] . One is the inconsistency of response spectra of different detectors, and the other is the difference of the radiation response coefficients of the detectors.…”

Section: Removal Of Striping Noisementioning

confidence: 99%

On-orbit performance of Fengyun-4A advanced geosynchronous radiation imager

Chen

Han²,

Guo³

et al. 2023

Earth and Space: From Infrared to Terahertz (ESIT 2022)

View full text Add to dashboard Cite

The performance of the Advanced Geosynchronous Radiation Imager (AGRI) onboard the Fengyun-4 (FY-4) satellite was evaluated. The results showed that its performance has approached the international advanced level of similar instruments. A wavelet filter was designed to filter the striping noises of the AGRI, which come from the inconsistency of the spectrum and thus varies with the observed target. The non-uniformity of the whole image was evaluated by weighting the probability variance of its count values. After filtering, the infrared image of the AGRI had a uniform visual effect on a single channel, and the pattern noise was reduced to below 0.2 K, which ensures the uniformity of products such as cloud detection and sea temperature that use the infrared splitting window channels. Furthermore, the AGRI carries the surface source blackbody as the on-orbit radiation calibration benchmark and rotates the north-south scanning mirror to point to the blackbody to obtain the blackbody observation data in each infrared band. In the current mode, the infrared calibration is performed every 15 minutes. The long-term evaluation results showed that the radiation calibration accuracy of the infrared channels of the AGRI was better than 0.5 K, and some of the channels reached 0.2 K. In addition, the long-term monitoring of the sensitivity and radiative cooling of the AGRI has shown the on-orbit variation of the instrument, which provides a reference for the development of the subsequent instrument.

show abstract

Section: Removal Of Striping Noisementioning

confidence: 99%

On-orbit performance of Fengyun-4A advanced geosynchronous radiation imager

Chen

Han²,

Guo³

et al. 2023

Earth and Space: From Infrared to Terahertz (ESIT 2022)

View full text Add to dashboard Cite

show abstract

“…In addition to high performance requirements, the system should consume low power and have minimal size since onboard systems have limited resources. Field programmable gate arrays (FPGAs) have emerged as a suitable option for onboard processing of hyperspectral data [3]- [8] due to their reconfigurability and hardware efficiency. In many cases, instead of a direct FPGA implementation of the algorithm, certain reformulations to the algorithm can be performed to further increase the performance, especially of the FPGA architecture.…”

Section: Introductionmentioning

confidence: 99%

Improved Real-Time Fast Simplex Growing Algorithm for Finding Endmembers in Hyperspectral Data and its FPGA architecture

Ganjimala¹

2022

Preprint

View full text Add to dashboard Cite

<p>This paper proposes a fast-processing algorithm for finding endmembers in hyperspectral image data and it’s FPGA architecture. The proposed algorithm named as Improved Real-Time Fast Simplex Growing Algorithm (IRT-FSGA) removes redundant computation for the pixels which do not qualify as endmember replacements and thereby achieves a significant reduction in computation time over the RT-FSGA algorithm. The improvement obtained is without compromising on the accuracy of the endmembers found. FPGA implementation results of the designed architecture show that the proposed IRT-FSGA achieves 33× improvement in performance at the cost of only a small area overhead. Thus, it is potentially useful in real-time applications which have stringent throughput and power requirements.</p>

show abstract

“…The field-programmable gate array (FPGA) has the advantages of a small size, a low power consumption, and reconfigurability, and it has received more and more attention in the direction of accelerating CNNs [13]. Compared with GPUs with a high power consumption and ASICs with long development cycles, FPGAs are more suitable for the hardware acceleration of neural network models on satellite platforms with limited resources, space, and power consumption [14]. The performance of FPGA acceleration is closely related to the on-chip resources of hardware and different CNN structures.…”

Section: Introductionmentioning

confidence: 99%

Hardware Acceleration and Implementation of YOLOX-s for On-Orbit FPGA

et al. 2022

View full text Add to dashboard Cite

The rapid development of remote sensing technology has brought about a sharp increase in the amount of remote sensing image data. However, due to the satellite’s limited hardware resources, space, and power consumption constraints, it is difficult to process massive remote sensing images efficiently and robustly using the traditional remote sensing image processing methods. Additionally, the task of satellite-to-ground target detection has higher requirements for speed and accuracy under the conditions of more and more remote sensing data. To solve these problems, this paper proposes an extremely efficient and reliable acceleration architecture for forward inference of the YOLOX-s detection network an on-orbit FPGA. Considering the limited onboard resources, the design strategy of the parallel loop unrolling of the input channels and output channels is adopted to build the largest DSP computing array to ensure a reliable and full utilization of the limited computing resources, thus reducing the inference delay of the entire network. Meanwhile, a three-path cache queue and a small-scale cascaded pooling array are designed, which maximize the reuse of on-chip cache data, effectively reduce the bandwidth bottleneck of the external memory, and ensure an efficient computing of the entire computing array. The experimental results show that at the 200 MHz operating frequency of the VC709, the overall inference performance of the FPGA acceleration can reach 399.62 GOPS, the peak performance can reach 408.4 GOPS, and the overall computing efficiency of the DSP array can reach 97.56%. Compared with the previous work, our architecture design further improves the computing efficiency under limited hardware resources.

show abstract

On-Orbit Real-Time Variational Image Destriping: FPGA Architecture and Implementation

Cited by 6 publications

References 48 publications

On-orbit performance of Fengyun-4A advanced geosynchronous radiation imager

On-orbit performance of Fengyun-4A advanced geosynchronous radiation imager

Improved Real-Time Fast Simplex Growing Algorithm for Finding Endmembers in Hyperspectral Data and its FPGA architecture

Hardware Acceleration and Implementation of YOLOX-s for On-Orbit FPGA

Contact Info

Product

Resources

About