Integrating EfficientNet, Cosine Annealing, and Advanced Data Augmentation for Enhanced Aircraft Detection in Satellite Imagery

DOI

10.2991/978-94-6463-447-1_25How to use a DOI?

Keywords

EfficientNet; Image Recognition Optimization; Aircraft Detection; Data Augmentation; Mixed Precision Training; Cosine Annealing

Abstract

This study introduces a groundbreaking approach to aircraft detection in satellite imagery, featuring an integrated suite of advanced methodologies that include the EfficientNet architecture, sophisticated data augmentation, mixed precision training, and cosine annealing learning rate optimization. We demonstrate how the synergy of these innovations significantly enhances model performance, providing a nuanced solution to the challenges of small target recognition and environmental variability in satellite imagery analysis. EfficientNet, renowned for its balance between computational efficiency and accuracy, is meticulously fine-tuned with a comprehensive set of data augmentation techniques such as RandomResizedCrop and RandomHorizontalFlip, enriching the training dataset and bolstering the model’s generalization capacity. The incorporation of mixed precision training facilitates faster computation and reduced memory usage, while the cosine annealing scheduler adeptly modulates the learning rate, fostering improved model convergence and robustness. The empirical outcomes underscore the superior detection capabilities of our model, marked by a high accuracy close to 95%, high precision, recall, and F1 scores, thereby establishing a new standard in satellite-based aircraft detection. This research not only propels the domain of remote sensing forward but also offers a scalable and efficient framework for real-time aerial surveillance and monitoring.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Lianzheng He
AU  - Junxin Wang
AU  - Rui Wang
PY  - 2024
DA  - 2024/07/14
TI  - Integrating EfficientNet, Cosine Annealing, and Advanced Data Augmentation for Enhanced Aircraft Detection in Satellite Imagery
BT  - Proceedings of the 2024 3rd International Conference on Engineering Management and Information Science (EMIS 2024)
PB  - Atlantis Press
SP  - 219
EP  - 227
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-447-1_25
DO  - 10.2991/978-94-6463-447-1_25
ID  - He2024
ER  -