Paper
9 December 2021 A remote sensing image scene classification framework based Efficientnet-b7
Yuanhui Chen, Kaile Ye, Chunmei Xu, Lili Zhan
Author Affiliations +
Proceedings Volume 12129, International Conference on Environmental Remote Sensing and Big Data (ERSBD 2021); 1212903 (2021) https://doi.org/10.1117/12.2625586
Event: 2021 International Conference on Environmental Remote Sensing and Big Data, 2021, Wuhan, China
Abstract
Remote sensing image scene classification (RSSC) is a hot topic in remote sensing. Aiming at the problem of remote sensing image of limited labeling samples and class-imbalanced, we proposed a RSSC framework based on efficientnetB7. The framework uses mirroring, rotation, cropping, HSV disturbance, and gamma transform to improve the problem of class-imbalance, and restricts the rotation angle of the high-rise-sparse-buildings to make it in line with the actual situation., and then the pre-trained model is used for training. The results show that the kappa and OA of the model increased 0.139 and 0.055, respectively, and the classification deviation caused by class-imbalance is alleviated.
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yuanhui Chen, Kaile Ye, Chunmei Xu, and Lili Zhan "A remote sensing image scene classification framework based Efficientnet-b7", Proc. SPIE 12129, International Conference on Environmental Remote Sensing and Big Data (ERSBD 2021), 1212903 (9 December 2021); https://doi.org/10.1117/12.2625586
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Data modeling

Remote sensing

Scene classification

Classification systems

Image classification

Statistical modeling

Neural networks

Back to Top