Real-Time Ellipse Detection for Robotics Applications

We propose a new algorithm for real-time detection and tracking of elliptic patterns suitable real-world robotics applications. The method fits ellipses to each contour in the image frame rejects that do not yield good fit. resulting is lightweight enough be used on robots' resource-limited onboard computers, can deal with lighting variations detect pattern even when view partial. tested an example application autonomous UAV landing fast-moving vehicle show its performance indoors, outdoors, simulation task. comparison other well-known ellipse methods shows our proposed outperforms F1 score 0.981 dataset over 1500 frames. videos experiments, source codes, collected are provided paper at https://theairlab.org/landing-on-vehicle .