Trinity-Net: 54,000 Remote Sensing Image Pairs for Dehazing

Trinity-Net is a remote sensing image dehazing dataset and model repository published in 2023 by K. Chi, Y. Yuan, and Q. Wang in the IEEE Transactions on Geoscience and Remote Sensing. It provides paired imagery designed to support gradient-guided Swin Transformer architectures for atmospheric correction in satellite data.

Use Cases

• Training dehazing models using paired hazy and clear remote sensing images
• Benchmarking gradient-guided Swin Transformer performance on satellite data
• Developing atmospheric correction algorithms for multispectral imagery

Strengths

• Peer-reviewed publication in IEEE T-GRS (2023)
• Specifically optimized for gradient-guided Swin Transformer architectures
• Focuses on remote sensing-specific atmospheric noise rather than general photography

Limitations

• Unknown total record count and file size
• Lack of explicit license information in the repository
• No sample data preview available for immediate inspection

Provenance

Source

IEEE Transactions on Geoscience and Remote Sensing (T-GRS)

Freshness

Last updated January 2023, coinciding with its publication in IEEE T-GRS.