fairfolio
Saved Projects

Hawkeye: Unmanned Search and Rescue Missions through Intelligent Drones Guided by Computer Vision & Dynamic Pathfinding

CSEF · 2016 Mathematics & Computer Science Fourth Award

Adiyan Kaul, Sohan M. Vichare · Cupertino High School, Cupertino, Santa Clara County

Overview

Objectives/Goals Build and program an autonomous drone that can carry out search and rescue missions in realistically dangerous terrain, without human control. More specifically, the drone should be able to: 1) Fly a flight pattern around an area of land, 2) Search for and identify people, 3) Bring these people back to pre-designated safe locations, and 4) Detect and navigate around obstacles throughout the process. Methods/Materials Software Components: OpenCV Computer Vision Library (python), modified facial recognition classifier algorithm with 500 additional positives and 400 additional negatives for drone capability. Paper written by Sven Koenig and Maxim Likhachev (http://robotics.cs.tamu.edu/dshell/cs625/aaai02b.pdf) detailing D*Lite Pathfinding Algorithm. Computer with Python 2.6 Suite and stock math libraries. Hardware Components: 3DR Y6 Drone Body. SF and SFP Propellers. Raspberry Pi 2 B+ Running Debian OS. 5500 mAH LiPo battery. Various wires. Soldering materials. Wi-Fi Module for Ad-Hoc Network. Results Drone modified by us to hold a Raspberry Pi and Camera successfully searched for and identified 2 out of 2 people in windy environment using computer vision algorithms/classifier we trained. Drone proceeded to lead said people back to safe locations while detecting/avoiding obstacles using the D*Lite pathfinding algorithm. Hardware: Successful interfacing between Raspberry Pi Camera, Raspberry Pi, Pixhawk Drone CPU, and Drone. Software: 1) Successful person recognition (modified OpenCV) 2) Successful spoofing of MavLink commands to control drone 3) Successful obstacle detection using OpenCV 4) Successful obstacle avoidance using D* Lite Pathfinding Algorithm (same algorithm was used on Mars Rovers Spirit and Opportunity) Conclusions/Discussion In the status quo, search and rescue missions remain immensely dependent on a large human volunteer base - which can be problematic in rural or dangerous locations. Our project allows for a $40 modification to an existing industrial drone that can automate it to carry out search and rescue missions independently - something which we believe will be immensely useful to government organizations. Secondly, we have developed a modular way for a Raspberry Pi and Camera to interface with a common drone CPU (the Pixhawk) and create an Ad-Hoc network for computer connection - something that any developer can use to automate drones.

Summary statement

We created a $40 method of automating industrial drones to carry out unmanned search and rescue missions in realistically rough and windy terrain.

Help received

None, save for open source libraries (OpenCV, math libraries) and this paper: Pathfinding with D* Lite (http://robotics.cs.tamu.edu/dshell/cs625/aaai02b.pdf)

Awards (1)

Competition history

  • CSEF 2016 Mathematics & Computer Science · Entry S1412

Resources

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

Source: California Science & Engineering Fair public projects

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google