Learning Representations of Satellite Images with Evaluations on Synoptic Weather Events

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yo, Ting-Shuo, Su, Shih-Hao, Wu, Chien-Ming, Chen, Wei-Ting, Chu, Jung-Lien, Chang, Chiao-Wei, Kuo, Hung-Chi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916887556784128
author Yo, Ting-Shuo
Su, Shih-Hao
Wu, Chien-Ming
Chen, Wei-Ting
Chu, Jung-Lien
Chang, Chiao-Wei
Kuo, Hung-Chi
author_facet Yo, Ting-Shuo
Su, Shih-Hao
Wu, Chien-Ming
Chen, Wei-Ting
Chu, Jung-Lien
Chang, Chiao-Wei
Kuo, Hung-Chi
contents This study applied representation learning algorithms to satellite images and evaluated the learned latent spaces with classifications of various weather events. The algorithms investigated include the classical linear transformation, i.e., principal component analysis (PCA), state-of-the-art deep learning method, i.e., convolutional autoencoder (CAE), and a residual network pre-trained with large image datasets (PT). The experiment results indicated that the latent space learned by CAE consistently showed higher threat scores for all classification tasks. The classifications with PCA yielded high hit rates but also high false-alarm rates. In addition, the PT performed exceptionally well at recognizing tropical cyclones but was inferior in other tasks. Further experiments suggested that representations learned from higher-resolution datasets are superior in all classification tasks for deep-learning algorithms, i.e., CAE and PT. We also found that smaller latent space sizes had minor impact on the classification task's hit rate. Still, a latent space dimension smaller than 128 caused a significantly higher false alarm rate. Though the CAE can learn latent spaces effectively and efficiently, the interpretation of the learned representation lacks direct connections to physical attributions. Therefore, developing a physics-informed version of CAE can be a promising outlook for the current work.
format Preprint
id arxiv_https___arxiv_org_abs_2508_06122
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning Representations of Satellite Images with Evaluations on Synoptic Weather Events
Yo, Ting-Shuo
Su, Shih-Hao
Wu, Chien-Ming
Chen, Wei-Ting
Chu, Jung-Lien
Chang, Chiao-Wei
Kuo, Hung-Chi
Computer Vision and Pattern Recognition
Atmospheric and Oceanic Physics
This study applied representation learning algorithms to satellite images and evaluated the learned latent spaces with classifications of various weather events. The algorithms investigated include the classical linear transformation, i.e., principal component analysis (PCA), state-of-the-art deep learning method, i.e., convolutional autoencoder (CAE), and a residual network pre-trained with large image datasets (PT). The experiment results indicated that the latent space learned by CAE consistently showed higher threat scores for all classification tasks. The classifications with PCA yielded high hit rates but also high false-alarm rates. In addition, the PT performed exceptionally well at recognizing tropical cyclones but was inferior in other tasks. Further experiments suggested that representations learned from higher-resolution datasets are superior in all classification tasks for deep-learning algorithms, i.e., CAE and PT. We also found that smaller latent space sizes had minor impact on the classification task's hit rate. Still, a latent space dimension smaller than 128 caused a significantly higher false alarm rate. Though the CAE can learn latent spaces effectively and efficiently, the interpretation of the learned representation lacks direct connections to physical attributions. Therefore, developing a physics-informed version of CAE can be a promising outlook for the current work.
title Learning Representations of Satellite Images with Evaluations on Synoptic Weather Events
topic Computer Vision and Pattern Recognition
Atmospheric and Oceanic Physics
url https://arxiv.org/abs/2508.06122