Abstract:Thick cloud and cloud shadow often exist in optical remote sensing (RS) images, which limit the application range of RS data. To solve this challenging problem, we propose a multitemporal-based method of cloud removal to reconstruct the large cloud-cover region in RS images. The method can combine the advantages of high accuracy of deep-learning method and low computation cost of traditional method, and reduce the difficulty of establishing multi-temporal dataset to a certain extent. In the proposed method, the complex problem of thick cloud removal is decomposed into two simple processes: coarse stage and refined stage. In the stage of coarse cloud removal, it is found that the cloud-cover images are distributed more regularly in frequency domain than in the space domain. Therefore, the frequency domain residual module is introduced, and the frequency domain attention module is designed to achieve more focused fusion of three-temporal features. In the stage of refined cloud removal, the refined network is designed to retain the effective features of the coarse stage and further refine the reconstruction result based on the image self-correlation. We carried out sufficient experiments to demonstrate the superiority of the proposed method compared with four classical methods, which include WLR, STSCNN, RFRnet and PSTCR. In qualitative experiments, the proposed method has obvious advantages in information accuracy, visual naturalness and semantic rationality. In quantitative experiments, compared with the best of four comparison methods, average values of Peak Signal to Noise Ratio (PSNR) on three-temporal images are improved by 5.1934 dB, 2.7729 dB and 7.2364 dB respectively, and average values of Correlation Coefficient (CC) on three-temporal images are increased by 0.1048, 0.0110 and 0.0966 respectively. Moreover, the proposed method has an obvious advantage in computation cost because it can reconstruct multiple cloud-cover images synchronously.

Deep Feature Gaussian Processes for Single-Scene Aerosol Optical Depth Reconstruction

Deep learning for Aerosol Forecasting

Aerosol optical depth retrieval using scaled digital number (DN) values of multi-spectral satellite and a generating adversarial model based on deep learning application

Deep Learning with Pretrained Framework Unleashes the Power of Satellite-Based Global Fine-Mode Aerosol Retrieval

Deep Parameterized Neural Networks for Hyperspectral Image Denoising

A Deep-Learning and Transfer-Learning Hybrid Aerosol Retrieval Algorithm for FY4-AGRI: Development and Verification over Asia

Deep Gaussian Processes for Air Quality Inference

Deep Random Features for Scalable Interpolation of Spatiotemporal Data

Improved 1-km-Resolution Hourly Estimates of Aerosol Optical Depth Using Conditional Generative Adversarial Networks

Deep Ground Filtering of Large-Scale ALS Point Clouds Via Iterative Sequential Ground Prediction

Gaussian Process and Deep Learning Atmospheric Correction

Retrieval of Atmospheric Aerosol Optical Depth From AVHRR Over Land With Global Coverage Using Machine Learning Method

Reconstructing aerosol optical depth using spatiotemporal Long Short-Term Memory convolutional autoencoder

Dense Point Clouds Matter: Dust-GS for Scene Reconstruction from Sparse Viewpoints

Wide and Deep Learning Model for Satellite-Based Real-Time Aerosol Retrievals in China

A deep-learning reconstruction method for remote sensing images with large thick cloud cover

Synthesis of Multi-Sensor Top of Atmosphere and Ground Level Reflectances to Support High-Resolution AOD Estimation with Machine Learning

DEM Extraction from Airborne Lidar Point Cloud in Thick-Forested Areas Via Convolutional Neural Network

Deep Feature Aggregation Framework Driven by Graph Convolutional Network for Scene Classification in Remote Sensing.

Deep-AIR: A Hybrid CNN-LSTM Framework forFine-Grained Air Pollution Forecast

Single Remote Sensing Image Dehazing Using Gaussian and Physics-Guided Process