Skip to content

Imaging Home

Jump to bottom
Eric Pedley edited this page Sep 22, 2023 · 4 revisions

Imaging Home

Welcome to the Imaging sub-system GitHub Wiki! This page serves as a central hub for important information and repositories. In addtion we will keep development rules and conventions listed here for reference. Let me (vince) or eric know if you have any questions.

Table of Contents

  1. Introduction
  2. Team Overview
  3. Key Repositories
  4. Use Cases
  5. Getting Started
  6. Contribution Guidelines
  7. Resources

Team Overview

Welcome to Imaging, the computer vision subsystem responsible for the UAV's perception capabilities. In SUAS, our competition, we compete against other teams in a payload delievery task. Specifcally we will be given a set of waypoints to fly which will then lead into a specified area containing the targets. In addition to the waypoints, teams are also given specific target identifiers which are characteristics unique to each target. Here are the SUAS 2024 rules for more details regarding this. Almost all of the competition scoring is based on the ability to locate your targets and successfully deliever its associated payload to that target without damage or safety concerns.

Imaging plays a critical role in this process as we are tasks with distingishing the target and also finding the location of the target in 3D based on the image in 2D in order to find a reference location to navigate to. The procedure for this was a complete mystery to me when I first joined the team so if you are interested this process primary utilitizes the camera projection matrix which is unique to each camera. Camera calibration will influence that matrix and there are many techniques and procedures for this too.

Alongside this Imaging will also aid in development of a ground station application to meet the nessecary functionality described in the rule set as well. Both software subsystems (GNC & Imaging) will most likely split this responsbility.

Lastly, Imaging also will be apart of experimental techniques to improve the overall drone performance. This can mean VSLAM to better estimate the drones position in flight or stero vision for obstacle avoidance.

Key Repositories

Competition Software Package

  • uavf_2024: [Repository]
  • Description: [The competition python package that will be used during SUAS. This contains both GNC and Imaging code]
  • Use Cases: [The home for all software development in UAV Forge and a familar package oriented API for specific scripting nessecary to specific needs.]
  • Contributing: [Link to the repository's contribution guidelines, if applicable.]

Development and Deployment Docker

  • uavf_software_docker: [Repository]
  • Description: [Dockerfiles and images to set up both deployment and development environment. One image is for x86 machines while the other is for our onboard computer using ARM.]
  • Use Cases: [Environment management and dependency tracking over time to manage changes.]

Simulation and training Docker

  • uavforge_aws_docker: [Repository]
  • Description: [This repo contains instructions on how to connect to the AWS cloud instance for things like training and simulating.]
  • Use Cases: [Only for use in the aws cloud instance. This repo can be cloned for docker building on the instance. Its likely that there is already an initalized repo there so git pull for updates.]
  • Contributing: [Link to the repository's contribution guidelines, if applicable.]
Clone this wiki locally