8000 GitHub - mayakundakci/MonoNav: MonoNav: MAV Navigation via Monocular Depth Estimation and Reconstruction
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

mayakundakci/MonoNav

 
 

Repository files navigation

Maya Kundakci 16362 Project Documentation

Project Description

Implemented Nathaniel Simon and Anirudha Majumdar's MonoNav monocular navigation stack.

Hardware

  • Crazyflie 2.1
  • Wolfwhoomp monocular camera
  • Flow deck v2
  • 350mAh LiPo battery
  • Video receiver
  • Crazyradio PA

Project Timeline

Setup and Configuration for Crazyflie

  1. Setting up the required dependencies:
  2. Driver Installation and Connectivity
    • Install the Crazyradio USB Driver (Zadig for Windows)
  3. Hardware-Software Integration:
    • Linking the Crazyflie hardware and Crazyflie Python library to interact via cfclient.

Setup and Configuration for MonoNav

  1. Setup a Linux environment
  2. Install dependencies for MonoNav
  3. Install dependencies for ZoeDepth
  4. Update MonoNav code base to be compatible with environment
    • ZoeDepth state dict parameter compatability
    • OpenCV path
  5. Configure video stream setup (video receiver and computer)

MonoNav Implementation

  1. Ran demo data to ensure MonoNav was setup correctly
  2. Calibrated camera using MonoNav calibrate.py (OpenCV)
  3. Collected depth data streaming from Crazyflie and mapped environments (available at data/Room and data/Hallway)

Limitations

We were not able to take off our Crazyflie due to a problem with the center of mass on our drone. Therefore, we could not activate the mononav mapping & path planning flight mode. The next step for our project is to 3D print a mound for the camera and execute mononav.

About

MonoNav: MAV Navigation via Monocular Depth Estimation and Reconstruction

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%
0