TY - GEN
T1 - “AsyncELF”
T2 - AIAA SciTech Forum and Exposition, 2024
AU - Tan, Zu Puayen
AU - Hsu, Kyle
AU - Tan, Jia Ming
N1 - Publisher Copyright:
© 2024 by Zu Puayen Tan.
PY - 2024
Y1 - 2024
N2 - Wind-tunnel and flight-tests are crucial for hypersonic research in areas such as fluidstructure interactions (FSI). However, existing diagnostic tools based on high-speed camera and point-sensors contain limiting trade-offs: Point-sensors are ideal for embedding in test articles but do not easily provide spatially-resolved data. On the other hand, 10-100kHz highspeed cameras are too large (and costly) to mount within most test articles, while their large data throughputs are not amenable to in-flight downlink during flight-tests. To address these challenges, we aim to develop a technique called “AsyncELF” (Asynchronous Embedded Light-Field), which combines nascent event-based (EB) cameras with stereo-photogrammetry to enable 3D time-resolved measurements of structural flutter. The AsyncELF technique reduces camera footprint to ~72mm and ~80g while supporting 100kHz-equivalent response, 720p resolution, and sparse-data encoding. This paper reports the preliminary results from benchtop tests of AsyncELF, with particular focus on techniques to acquire and post-process asynchronously-generated EB data. Successful camera calibration of two synchronized EB cameras in a 3D stereo-pair was achieved with sub-pixel accuracy. A Mean-Shift Filteringinspired algorithm was implemented to track photogrammetric surface-markers on a benchtop moving panel (i.e. model of hypersonic thin-panel flutter) with 1-2px of trackingerror. Finally, the panel’s dynamics were successfully reconstructed in 3D, demonstrating preliminary feasibility of the AsyncELF concept.
AB - Wind-tunnel and flight-tests are crucial for hypersonic research in areas such as fluidstructure interactions (FSI). However, existing diagnostic tools based on high-speed camera and point-sensors contain limiting trade-offs: Point-sensors are ideal for embedding in test articles but do not easily provide spatially-resolved data. On the other hand, 10-100kHz highspeed cameras are too large (and costly) to mount within most test articles, while their large data throughputs are not amenable to in-flight downlink during flight-tests. To address these challenges, we aim to develop a technique called “AsyncELF” (Asynchronous Embedded Light-Field), which combines nascent event-based (EB) cameras with stereo-photogrammetry to enable 3D time-resolved measurements of structural flutter. The AsyncELF technique reduces camera footprint to ~72mm and ~80g while supporting 100kHz-equivalent response, 720p resolution, and sparse-data encoding. This paper reports the preliminary results from benchtop tests of AsyncELF, with particular focus on techniques to acquire and post-process asynchronously-generated EB data. Successful camera calibration of two synchronized EB cameras in a 3D stereo-pair was achieved with sub-pixel accuracy. A Mean-Shift Filteringinspired algorithm was implemented to track photogrammetric surface-markers on a benchtop moving panel (i.e. model of hypersonic thin-panel flutter) with 1-2px of trackingerror. Finally, the panel’s dynamics were successfully reconstructed in 3D, demonstrating preliminary feasibility of the AsyncELF concept.
UR - http://www.scopus.com/inward/record.url?scp=85197763606&partnerID=8YFLogxK
U2 - 10.2514/6.2024-2492
DO - 10.2514/6.2024-2492
M3 - Conference contribution
AN - SCOPUS:85197763606
SN - 9781624107115
T3 - AIAA SciTech Forum and Exposition, 2024
BT - AIAA SciTech Forum and Exposition, 2024
PB - American Institute of Aeronautics and Astronautics Inc, AIAA
Y2 - 8 January 2024 through 12 January 2024
ER -