CaBuAr Wildfire Spread Prediction

Tools

• CaBuAr_Pipeline_predict_on_cabuar_test_sample Run the CaBuAr V4 burn-scar projection on a held-out test sample of the CaBuAr ChaBuD split and score it against the ground truth: SSIM, pixel MSE and PSNR against the true post-fire image, and burn-scar IoU, F1, precision and recall against the CaBuAr ground-truth burned-area mask (ΔNBR > 0.05). Returns: Tuple of (pre_fire_false_colour_image, predicted_post_fire_false_colour_image, comparison_image with the true post-fire scene and the predicted burn scar against the ground-truth mask, metrics_markdown).
• CaBuAr_Pipeline_predict_post_fire_from_pre_fire_npy Project the post-fire Sentinel-2 scene and the burn scar from a pre-fire Sentinel-2 image of an area in which a fire later occurs, using the CaBuAr V4 latent diffusion model. The burn scar is derived with the Normalized Burn Ratio difference, ΔNBR = NBR(pre-fire) - NBR(predicted post-fire), and a pixel counts as burned when ΔNBR > 0.05. The model does not predict whether or where a fire will ignite; it assumes a fire occurs in the scene. If a user passes a file as an input, use the upload_file_to_gradio tool, if present, to upload the file to the gradio app and create a Gradio File Input. Then use the returned path as the input to the tool. Returns: Tuple of (pre_fire_false_colour_image, predicted_post_fire_false_colour_image, delta_nbr_map_image, summary_markdown with the predicted burned fraction).

No tool matches.

As the server's own listing gives them. We never call a tool.