Evaluation of quantized CNN architectures for land use classification for onboard cube satellite computing

Authors

• Wynnezel Wayne Naoto P. Akeboshi, De La Salle University, Manila
• Robert Kerwin C. Billones, De La Salle University, Manila
• Jerahmeel Kua Coching, De La Salle University, Manila
• Adrian Jenssen Lim Pe, De La Salle University, Manila
• Seth Gabriel Dy Yeung, De La Salle University, Manila
• Richard Josiah Tan Ai, De La Salle University, Manila
• Edwin Sybingco, De La Salle University, Manila
• Elmer Jose P. Dadios, De La Salle University, Manila
• Mark Angelo Purio, Adamson University

College

Gokongwei College of Engineering

Department/Unit

Manufacturing Engineering and Management

Document Type

Conference Proceeding

Source Title

2024 9th International Conference on Mechatronics Engineering (ICOM)

First Page

45

Last Page

50

Publication Date

2024

Abstract

Nanosatellites have increased in popularity, but its limited communication bandwidth and computing resource constraints remain a challenge. Orbital edge computing has been explored to address this issue by utilizing neural networks onboard the satellites for land use classification. This study explores the performance of pre-trained quantized neural networks in land use classification by fine-tuning ResN et50, MobileNetV3, and GoogLeNet using the EuroSAT RGB dataset with an 80–20 split. The study evaluated the three models based on its accuracy and file size and against its non-quantized counterparts. It was found that the best model is the MobileNetV3, with an 89.04% accuracy at a 3,090 KB file size. MobileNetV3 was also found to be the best architecture for quantization with the least impact to accuracy, with a reduction of only 5.05%, and having the most significant reduction in file size of 85.71 %.

html

Recommended Citation

Akeboshi, W. P., Billones, R. C., Coching, J. K., Pe, A. L., Yeung, S. D., Tan Ai, R., Sybingco, E., Dadios, E. P., & Purio, M. (2024). Evaluation of quantized CNN architectures for land use classification for onboard cube satellite computing. 2024 9th International Conference on Mechatronics Engineering (ICOM), 45-50. Retrieved from https://animorepository.dlsu.edu.ph/faculty_research/15465

Disciplines

Electrical and Computer Engineering

Keywords

CubeSats; Neural networks (Computer science)

Upload File

wf_no