Supervisor: Southwest Ordnance Industry Bureau
Organizer: Chongqing Ordnance Industry Society
Chongqing University of Technology

Remote sensing image scene classification algorithm based on the ResMLP model

DOI: 10.11809/bqzbgcxb2023.01.032
Keywords: remote sensing image; deep learning; scene classification; embedded realization; ResMLP model
Abstract: The research on remote sensing image scene classification based on deep learning has made rapid progress, but the algorithms under most schemes require a lot of computer resources, which is difficult to be directly deployed on the embedded system and cannot provide support for potential satellite on orbit missions. A remote sensing image scene classification algorithm based on the ResMLP model is researched to solve this problem. The algorithm is optimized according to the characteristics of the embedded system, and the embedded implementation of the model is completed. The experimental results show that the algorithm only needs 18.4 M params in the embedded system to reach 85.9% TOP 1 accuracy of remote sensing image classification, which is both accurate and concise.
Published: 2023-01-28
PDF HTML

Address: Room 614, 6th Floor, Mingde Building, Chongqing University of Technology, No. 69 Hongguang Avenue, Banan District, Chongqing City

Postal Code: 400054   Phone: 023-68852703, 023-62583262, 023-62569336   Fax: 023-68852703   Email: bzxb@vip.cqut.edu.cn