Abstract- For this project, we recommend using the YOLOv4 and YOLOv5 models to create a system that can accurately detect and classify drones and birds. The use of drones is increasingly threatening to bird populations, and it is crucial to develop a solution that can identify them. To do this, we will train the YOLOv4 and YOLOv5 models on the training set, using transfer learning. After that, we will assess their performance on the validation set and test their accuracy on a separate test set. Finally, we will compare the performance of the two models and select the best one for bird and drone detection.

| DOI: 10.17148/IARJSET.2023.10839