UCloudNet: A Residual U-Net with Deep Supervision for Cloud Image Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Yijie, Wang, Hewei, Wang, Shaofan, Lee, Yee Hui, Pathan, Muhammad Salman, Dev, Soumyabrata
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909454420672512
author Li, Yijie
Wang, Hewei
Wang, Shaofan
Lee, Yee Hui
Pathan, Muhammad Salman
Dev, Soumyabrata
author_facet Li, Yijie
Wang, Hewei
Wang, Shaofan
Lee, Yee Hui
Pathan, Muhammad Salman
Dev, Soumyabrata
contents Recent advancements in meteorology involve the use of ground-based sky cameras for cloud observation. Analyzing images from these cameras helps in calculating cloud coverage and understanding atmospheric phenomena. Traditionally, cloud image segmentation relied on conventional computer vision techniques. However, with the advent of deep learning, convolutional neural networks (CNNs) are increasingly applied for this purpose. Despite their effectiveness, CNNs often require many epochs to converge, posing challenges for real-time processing in sky camera systems. In this paper, we introduce a residual U-Net with deep supervision for cloud segmentation which provides better accuracy than previous approaches, and with less training consumption. By utilizing residual connection in encoders of UCloudNet, the feature extraction ability is further improved.
format Preprint
id arxiv_https___arxiv_org_abs_2501_06440
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle UCloudNet: A Residual U-Net with Deep Supervision for Cloud Image Segmentation
Li, Yijie
Wang, Hewei
Wang, Shaofan
Lee, Yee Hui
Pathan, Muhammad Salman
Dev, Soumyabrata
Computer Vision and Pattern Recognition
Image and Video Processing
Recent advancements in meteorology involve the use of ground-based sky cameras for cloud observation. Analyzing images from these cameras helps in calculating cloud coverage and understanding atmospheric phenomena. Traditionally, cloud image segmentation relied on conventional computer vision techniques. However, with the advent of deep learning, convolutional neural networks (CNNs) are increasingly applied for this purpose. Despite their effectiveness, CNNs often require many epochs to converge, posing challenges for real-time processing in sky camera systems. In this paper, we introduce a residual U-Net with deep supervision for cloud segmentation which provides better accuracy than previous approaches, and with less training consumption. By utilizing residual connection in encoders of UCloudNet, the feature extraction ability is further improved.
title UCloudNet: A Residual U-Net with Deep Supervision for Cloud Image Segmentation
topic Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2501.06440