TY - JOUR
T1 - A Method for Generating Large-Scale Spaceborne Long-Wave Infrared Point-Target Images Using Landsat 8 TIRS Data
AU - Shan, Ruiyan
AU - Dong, Liquan
AU - Li, Kang
AU - Dong, Lianqing
AU - Zheng, Guoxian
AU - Zhang, Muyao
AU - Wang, Yu
AU - Ren, Weihe
N1 - Publisher Copyright:
© 2026 IEEE.
PY - 2026
Y1 - 2026
N2 - Deep learning has demonstrated remarkable potential in spaceborne infrared small target detection; however, its application is severely bottlenecked by the scarcity of high-quality, annotated on-orbit data. To bridge this 'Sim-to-Real' domain gap, this article proposes a radiometrically consistent image generation framework capable of synthesizing large-scale, high-fidelity long-wave infrared (LWIR) datasets. Unlike traditional methods that rely on simplistic image injection or computationally expensive ray-tracing, our approach adopts a hybrid strategy: it leverages the global archive of calibrated Landsat-8 data to ensure authentic background diversity and couples it with physically modeled targets derived from computational fluid dynamics. A key innovation is the radiance-domain fusion mechanism, which integrates targets and backgrounds strictly in physical units, considering atmospheric transmission and sensor degradation effects (PSF, motion blur, noise equivalent delta temperature), thereby guaranteeing energy conservation for subpixel targets. Quantitative validation against real on-orbit imagery from the SDG satellite demonstrates that the proposed framework achieves superior fidelity, with intensity profiles and statistical metrics (HP-PSNR) closely matching real observations. This work provides a scalable solution to the data scarcity problem, offering a valuable foundation for training and benchmarking robust deep learning models in spaceborne surveillance tasks.
AB - Deep learning has demonstrated remarkable potential in spaceborne infrared small target detection; however, its application is severely bottlenecked by the scarcity of high-quality, annotated on-orbit data. To bridge this 'Sim-to-Real' domain gap, this article proposes a radiometrically consistent image generation framework capable of synthesizing large-scale, high-fidelity long-wave infrared (LWIR) datasets. Unlike traditional methods that rely on simplistic image injection or computationally expensive ray-tracing, our approach adopts a hybrid strategy: it leverages the global archive of calibrated Landsat-8 data to ensure authentic background diversity and couples it with physically modeled targets derived from computational fluid dynamics. A key innovation is the radiance-domain fusion mechanism, which integrates targets and backgrounds strictly in physical units, considering atmospheric transmission and sensor degradation effects (PSF, motion blur, noise equivalent delta temperature), thereby guaranteeing energy conservation for subpixel targets. Quantitative validation against real on-orbit imagery from the SDG satellite demonstrates that the proposed framework achieves superior fidelity, with intensity profiles and statistical metrics (HP-PSNR) closely matching real observations. This work provides a scalable solution to the data scarcity problem, offering a valuable foundation for training and benchmarking robust deep learning models in spaceborne surveillance tasks.
KW - Long-wave infrared (LWIR)
KW - point target simulation
KW - radiometric consistency
KW - sim-to-real validation
KW - spaceborne infrared imaging
UR - https://www.scopus.com/pages/publications/105031845407
U2 - 10.1109/JSTARS.2026.3670447
DO - 10.1109/JSTARS.2026.3670447
M3 - Article
AN - SCOPUS:105031845407
SN - 1939-1404
VL - 19
SP - 9112
EP - 9126
JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
ER -