Abstract
This study proposes a method for integrating artificial intelligence, specifically Convolutional Neural Networks (CNN), into the camera system of the DJI Tello drone to recognize basic geometric objects such as circles, triangles, rectangles, and regular pentagons. A black-and-white image dataset was automatically generated, featuring varied sizes and positions of objects to optimize the training process. A simple CNN architecture was trained on this dataset and achieved near-perfect accuracy on both training and validation sets. The trained model was then integrated into the drone's camera, and experimental testing demonstrated its real-time recognition capability with no signs of overfitting. The results confirm the effectiveness and application potential of the proposed solution for intelligent drone systems.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright (c) 2025 Array