A scSE-LinkNet Deep Learning Model for Daytime Sea Fog Detection

Xiaofei Guo,Jianhua Wan,Mingming Xu,Hui Sheng,Shanwei Liu,Muhammad Yasir

doi:10.3390/rs13245163

Xiaofei Guo, Jianhua Wan + Show 4 more

Open Access

https://doi.org/10.3390/rs13245163

Copy DOI

Journal: Remote Sensing	Publication Date: Dec 20, 2021
Citations: 6	License type: CC BY 4.0

Affiliation: China University of Petroleum, East China

Abstract

Sea fog is a precarious weather disaster affecting transportation on the sea. The accuracy of the threshold method for sea fog detection is limited by time and region. In comparison, the deep learning method learns features of objects through different network layers and can therefore accurately extract fog data and is less affected by temporal and spatial factors. This study proposes a scSE-LinkNet model for daytime sea fog detection that leverages residual blocks to encoder feature maps and attention module to learn the features of sea fog data by considering spectral and spatial information of nodes. With the help of satellite radar data from Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP), a ground sample database was extracted from Moderate Resolution Imaging Spectroradiometer (MODIS) L1B data. The scSE-LinkNet was trained on the training set, and quantitative evaluation was performed on the test set. Results showed the probability of detection (POD), false alarm rate (FAR), critical success index (CSI), and Heidke skill scores (HSS) were 0.924, 0.143, 0.800, and 0.864, respectively. Compared with other neural networks (FCN, U-Net, and LinkNet), the CSI of scSE-LinkNet was improved, with a maximum increase of nearly 8%. Moreover, the sea fog detection results were consistent with the measured data and CALIOP products.

Highlights

IntroductionSea fog is a precarious weather phenomenon that appears when water vapor near the surface is condensed to form suspended water droplets [1]
The construction of the scSE-LinkNet deep learning model based on LinkNet and squeeze-andconstruction of the scSE-LinkNet deep learning model based on LinkNet and squeezeexcitation networks (SENet) is the most processprocess in this study
In order to show the experimental results of the scSE-LinkNet model, we used samples from the test set to obtain the sea fog detection results, which were compared with the

Summary

Introduction

Sea fog is a precarious weather phenomenon that appears when water vapor near the surface is condensed to form suspended water droplets [1]. It can result in the horizontal visibility of sea being less than 1 km and threatens the safety of navigation, aviation, and transportation, which in turn can affect the economy and threaten lives. Earth observation (EO) satellites offer cost-effective and timely images covering large areas with high temporal and spatial resolutions [3]. They have become indispensable technical means for real-time observation of the occurrence, development, and extinction of sea fog. Daytime sea fog detection is still a significant problem because of the similarity in spectral characteristics of fog and other types of cloud (middle/high level clouds, stratus clouds, and low clouds) [4]

Methods

Results

Discussion

Conclusion