AI Landing Platform Detection
Projects | | Links:
For the IMAV 2025 competition, I developed a real-time computer vision system dedicated to detecting landing platforms for autonomous drones.
Key Milestones:
- YOLO Architecture: I trained a YOLO-based neural network optimized for high-speed detection and low latency, which is critical for autonomous flight decision-making.
- Dataset Augmentation with Blender: To overcome the lack of diverse real-world images, I used Blender to generate a massive synthetic dataset. This allowed the model to train on various lighting conditions, camera angles, and textures.
- High Reliability: The integration of synthetic data resulted in exceptional detection robustness, ensuring precise landings even in visually complex or cluttered environments.
