Cloud Detection Algorithm for Remote Sensing Images Using Fully Convolutional Neural Networks

TL;DR

This paper introduces a deep learning framework using Fully Convolutional Neural Networks for precise pixel-level cloud detection in remote sensing images, enhancing accuracy by combining threshold-based and neural network methods.

Contribution

The paper proposes a hybrid cloud detection method that integrates FCN with gradient-based snow/ice exclusion, improving detection performance without manual ground truth correction.

Findings

Jaccard index improved by 4.36%

Recall measure increased by 3.62%

Hybrid approach enhances cloud detection accuracy

Abstract

This paper presents a deep-learning based framework for addressing the problem of accurate cloud detection in remote sensing images. This framework benefits from a Fully Convolutional Neural Network (FCN), which is capable of pixel-level labeling of cloud regions in a Landsat 8 image. Also, a gradient-based identification approach is proposed to identify and exclude regions of snow/ice in the ground truths of the training set. We show that using the hybrid of the two methods (threshold-based and deep-learning) improves the performance of the cloud identification process without the need to manually correct automatically generated ground truths. In average the Jaccard index and recall measure are improved by 4.36% and 3.62%, respectively.