Cloud-EGAN: Rethinking CycleGAN From a Feature Enhancement Perspective for Cloud Removal by Combining CNN and Transformer
نویسندگان
چکیده
Cloud cover presents a major challenge for geoscience research of remote sensing images with thick clouds causing complete obstruction information loss while thin blurring the ground objects. Deep learning (DL) methods based on convolutional neural networks (CNNs) have recently been introduced to cloud removal task. However, their performance is hindered by weak capabilities in contextual extraction and aggregation. Unfortunately, such play vital role characterizing complex In this work, conventional cycle-consistent generative adversarial network (CycleGAN) revitalized from feature enhancement perspective. More specifically, saliency (SE) module first designed replace original CNN CycleGAN re-calibrate channel attention weights capture detailed multi-level maps. Furthermore, high-level (HFE) developed generate contextualized cloud-free features suppressing components. particular, HFE composed both CNN- transformer-based modules. The former enhances local employing residual multi-scale strategies, latter captures long-range dependencies Swin transformer exploit global Capitalizing SE modules, an effective Cloud-Enhancement GAN, namely Cloud-EGAN, proposed accomplish tasks. Extensive experiments RICE WHUS2-CR datasets confirm impressive Cloud-EGAN.
منابع مشابه
A multi-scale convolutional neural network for automatic cloud and cloud shadow detection from Gaofen-1 images
The reconstruction of the information contaminated by cloud and cloud shadow is an important step in pre-processing of high-resolution satellite images. The cloud and cloud shadow automatic segmentation could be the first step in the process of reconstructing the information contaminated by cloud and cloud shadow. This stage is a remarkable challenge due to the relatively inefficient performanc...
متن کاملRethinking the Cloud Architecture for a Mobile Edge Cloud
The era of cloud computing is upon us. Cloud computing allows Internet applications to be hosted on remote server and storage resources that are leased on-demand from the cloud providers. Today's cloud platforms host a plethora of popular Internet applications in domains ranging from news, e-commerce, entertainment, banking, games and social media. A concurrent trend is the increasing ubiquity ...
متن کاملA risk model for cloud processes
Traditionally, risk assessment consists of evaluating the probability of "feared events", corresponding to known threats and attacks, as well as these events' severity, corresponding to their impact on one or more stakeholders. Assessing risks of cloud-based processes is particularly difficult due to lack of historical data on attacks, which has prevented frequency-based identification...
متن کاملCiteSeerx: A Cloud Perspective
Information retrieval applications are good candidates for hosting in a cloud infrastructure. CiteSeerx a digital library and search engine was built with the goal of efficiently disseminating scientific information and literature over the web. The framework for CiteSeerx as an application of the SeerSuite software is a design built with extensibility and scalability as fundamental features. Th...
متن کاملThin cloud removal from single satellite images.
A novel method for removing thin clouds from single satellite image is presented based on a cloud physical model. Given the unevenness of clouds, the cloud background is first estimated in the frequency domain and an adjustment function is used to suppress the areas with greater gray values and enhance the dark objects. An image, mainly influenced by transmission, is obtained by subtracting the...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
سال: 2023
ISSN: ['2151-1535', '1939-1404']
DOI: https://doi.org/10.1109/jstars.2023.3280947