Lack of Rescue Drone Footage:

Difficulty arose due to the absence of sufficient drone footage capturing disaster scenes, hindering the training and validation of our AI models.
Insufficient Images for Normal and Wounded Classification:

Obtaining an adequate number of images for training the classification models posed a challenge, particularly in distinguishing between normal individuals and those wounded in disaster scenarios.
Limited Availability of Drones or Suitable Images:

The project faced setbacks due to the scarcity of drones or suitable images depicting disaster-stricken areas, impacting our ability to gather diverse and representative data for model development.
Time Constraints:

Time constraints posed challenges in terms of model development, validation, and refinement, as well as in the overall project timeline for completion.
Resource Constraints for Model Training:

Resource limitations, including computing power and data annotation resources, constrained the training of the three AI models essential for object detection, classification, and tracking.