SAFDAS uses convolutional neural networks coupled with transfer learning. This architecture enables users to overcome training data limitations and train the model with region-specific information. Validated alerts are vital in establishing trust with end users, facilitating reliable information dispatch to the field, and enabling unbiased area estimations. This reliability aids in effective resource allocation and fosters accurate responses to forest disturbances. SARFDAS was created by SIG and was tested in Palawan, in the Philippines, and is currently being adapted for use with SERVIR SEA.
