Rice Blast Resistance Imagery from UAV Visible-Infrared Sensors

UAV-based visible-infrared imagery supports identification of rice blast resistance. The 4.6 MB dataset, published by Huasheng Huang, was last updated in April 2026. It includes associated codes for a parallel ResNet-Transformer and cross-modal attention network.

Use Cases

• Train a parallel ResNet-Transformer model to classify rice blast resistance labels from paired visible and infrared image channels.
• Apply cross-modal attention mechanisms to fuse features from UAV-based visible and infrared imagery for disease detection.
• Benchmark multimodal fusion architectures using the provided visible-infrared image pairs and associated model codes.
• Develop automated pipelines for identifying rice blast resistance to support breeding programs from aerial image data.

Strengths

• Dataset includes paired visible and infrared image modalities from UAVs.
• Released under a permissive CC-BY-4.0 license for reuse and modification.
• Associated source codes for the proposed neural network architecture are provided.

Limitations

• Small dataset scale at 4.6 MB, limiting model training and generalization.
• Unknown sample count and potential geographic bias from a single study location.
• Image resolution, labeling methodology, and class balance details are not provided.

Provenance

Source

Huasheng Huang via figshare.

Collection Method

Data gathered using UAV-based visible and infrared sensors.

Time Range

null

Freshness

Last updated in April 2026.

Geography

null