Xiaofeng Su

Dr. Xiaofeng Su


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
29 April 2021 in Signal, Image and Video Processing
Verified
Journal Article
Closed access
PDF file
LLLiyuan Li
Liyuan Li
LXLi Xiaoyan
Li Xiaoyan
LJLinyi Jiang
Linyi Jiang
XSXiaofeng Su
Xiaofeng Su
Fansheng ChenFansheng Chen
Fansheng Chen

11 September 2022 in Remote Sensing
Verified
Journal Article
Open access
PDF file
XHxinZhuoyue Hu
xinZhuoyue Hu
MZMin Zhu
Min Zhu
QWQiyao Wang
Qiyao Wang
XSXiaofeng Su
Xiaofeng Su
Fansheng ChenFansheng Chen
Fansheng Chen
SDGSAT-1 was launched in November 2021, and TIS (Thermal infrared sensor) is a major instrument onboard this satellite. An analysis of the radiometric calibration and noise performance of the TIS has been carried out in the thermal vacuum chamber before launch in order to ensure that it meets the requirements. The prelaunch test results show NEdT (noise equivalent temperature difference) is 0.034 K, 0.047 K and 0.076 K@300 K for the three bands, respectively. The maximum fitting residuals are less than 0.5 K at measured blackbody temperatures. In addition, this paper also discusses the dependence between TIS performance and instrument temperature and focal plane array (FPA) temperature. The good radiometric and noise performance of TIS demonstrates it has potential to provide high resolution thermal remote sensing data in urban heat island, and other environmental issues research.

9 February 2022 in Optical and Quantum Electronics
Verified
Journal Article
Closed access
PDF file
XNXinyue Ni
Xinyue Ni
SYShutian Yu
Shutian Yu
XSXiaofeng Su
Xiaofeng Su
Fansheng ChenFansheng Chen
Fansheng Chen

2 September 2020 in Optical and Quantum Electronics
Verified
Journal Article
Open access
PDF file
JZJiaqiang Zhang
Jiaqiang Zhang
LXLi Xiaoyan
Li Xiaoyan
LLLiyuan Li
Liyuan Li
PSPengcheng Sun
Pengcheng Sun
XSXiaofeng Su
Xiaofeng Su
Fansheng ChenFansheng Chen
Fansheng Chen
Accurate and rapid cloud detection is exceedingly significant for improving the downlink efficiency of on-orbit data, especially for the microsatellites with limited power and computational ability. However, the inference speed and large model limit the potential of on-orbit implementation of deep-learning-based cloud detection method. In view of the above problems, this paper proposes a lightweight network based on depthwise separable convolutions to reduce the size of model and computational cost of pixel-wise cloud detection methods. The network achieves lightweight end-to-end cloud detection through extracting feature maps from the images to generate the mask with the obtained maps. For the visible and thermal infrared bands of the Landsat 8 cloud cover assessment validation dataset, the experimental results show that the pixel accuracy of the proposed method for cloud detection is higher than 90%, the inference speed is about 5 times faster than that of U-Net, and the model parameters and floating-point operations are reduced to 12.4% and 12.8% of U-Net, respectively.

13 June 2023 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
SWSiqi Wang
Siqi Wang
LLLiyuan Li
Liyuan Li
LXLi Xiaoyan
Li Xiaoyan
JZJingwen Zhang
Jingwen Zhang
LZLixing Zhao
Lixing Zhao
Fansheng ChenFansheng Chen
Fansheng Chen
The quality of hyperspectral images seriously impedes subsequent high-level vision tasks such as image segmentation, image encoding and target detection. However, the frequency, spectral, and spatial properties of the hyperspectral noise pictures are not utilized fully by existing image denoising algorithms. To address this issue, a novel convolutional network based on united octave and attention mechanism (UOANet) is proposed to extract the frequency-spectral-spatial-feature for denoising the actual noise of HSIs. In particular, the negative residual mapping embedded in Unet is proposed for multi-scale abstract representation and two modules are designed for modeling global noisy HSI features in the frequency-spectral-spatial domain. Firstly, with the use of residual octave convolution module (ResOct), our model can focus on the intrinsic properties of HSI noise distribution for desirable noise removal. Next, a parallel spatial-spectral attention module (SSAT) is used to fully utilize the rich spectrum data and the various spatial data of each band in HSI, which improves the richness of HSI details after denoising. Experimental results on both synthetic and real HSIs demonstrate the validity and superiority of UOANet compared with the state-of-the-arts under various noise settings.

1 November 2022 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
ZCZhouxia Chen
Zhouxia Chen
XHxinZhuoyue Hu
xinZhuoyue Hu
XSXiaofeng Su
Xiaofeng Su
THtingliang hu
tingliang hu
Fansheng ChenFansheng Chen
Fansheng Chen
Radiometric calibration of satellites is one of the core technologies for analyzing satellite data quantitatively. For large-aperture remote sensing cameras, the blackbody is placed in the rear optical path due to its weight and size. As a result, the front optics’ self-emission cannot be evaluated when the inner blackbody is observed. To achieve full optical path calibrations, stars are used as radiation calibration sources for large-aperture cameras. In high energy concentration detection systems, the point spread function (PSF), intrapixel sensitivity (IPS), capacitive coupling and sampling phase may cause some energy of the point source to be lost, resulting in an energy difference between the extended and point sources. It is important to note that conventional aperture photometry is not always the ideal method for obtaining high-precision photometry. This paper proposes a method for compensation of point-source signals based on capacitive coupling correction, PSF reconstruction and IPS model, and establishes a response conversion model for stellar and inner blackbodies. According to calibration coefficients based on stars and inner blackbody, the error between the star calibration coefficient and the inner blackbody calibration coefficient is 0.18%, which is better than 59.4% before the proposed method was applied. The error between the two methods can be stabilized within 0.3% within 140 days of the launch of the satellite.