To address this challenge, we propose radiometric resolution and scale-aware masked autoencoder (RSMAE), a novel framework ... leverages SAR-specific radiometric characteristics to enhance ...

Methods: This study integrates rainfall, surface displacement, and vertical displacement monitoring data, and proposes an automatic failure mode identification method based on deep convolutional ...

The results indicate that different types of autoencoders are better suited for different tasks: denoising autoencoders excel at accurate reconstruction, while the sparse variant is more effective at ...