Semantic Segmentation

Perform Semantic Segmentation From Downward-Facing Drone Images

Welcome to the Semantic Segmentation Challenge!

This is a competition run by AIcrowd & Amazon Prime Air in order to develop safe and reliable autonomous control systems for aircraft. This challenge focuses on using images from a single downward-facing camera to estimate scene depth and perform semantic segmentation, including a new dataset of drone images that will benchmark semantic segmentation and mono-depth perception.

The two fundamental components of this project pertaining to computer vision are semantic segmentation and depth perception. Through this challenge, the Computer Vision community is encouraged to innovate and enhance the cutting-edge perception tasks related to drone images. This particular challenge focuses on semantic segmentation.


  • Challenge Launch: 22nd December 2022
  • Challenge End: 28th April 2023
  • Winner Announcement: 30th June 2023


  • The Top scoring submission will receive $15,000 USD
  • The Second best submission will receive $7,500 USD
  • The Third place submission will receive $1,250 USD
  • The Most “Creative” solution submitted to the whole competition, as determined by the Sponsor’s sole discretion, will receive $2,500 USD.

Sponsored by Saturn Cloud

The dataset is hosted by Saturn Cloud, which will be available for download. Each participant can use the Saturn Cloud computing environment, which provides 100 free hours of compute per participant, and a Python environment. Message Saturn Cloud support and say, "I'm competing in the Semantic Segmentation challenge." You'll be upgraded from the standard free tier to 100 hours of compute! Click the link below to get started with the competition.