Project summary

This project, conducted by H. Jekel at TU Delft - Aerospace Engineering, aimed to develop an autonomous obstacle avoidance system for a drone, targeting the largest distance coverage in a 10-minute flying contest. The team explored two primary methods: optical flow and color filtering. The optical flow method was adept at identifying obstacles and calculating distances between the drone’s camera and these obstacles, utilizing circular flight patterns for efficient detection and evasion. On the other hand, the color filtering approach allowed for the recognition of specific colors, facilitating obstacle detection. Despite the potential of the optical flow method, the team ultimately chose color filtering for the competition due to its simplicity and effectiveness in recognizing colored objects. The project culminated in successfully creating a drone equipped with an obstacle avoidance system capable of covering 67 meters in the contest, highlighting significant advancements in drone navigation and safety.

Project documents

Report:

• Download PDF

Presentation:

• Download PDF

Code implementation

Code for this blog:

• WEBSITE

Code to reproduce the results:

• Thresholds

• Optical flow circle

• Orange avoider