Kelp Wanted: Segmenting Kelp Forests

Help researchers estimate the extent of Giant Kelp Forests by segmenting Landsat imagery. #climate

$15,000 in prizes
Completed feb 2024
671 joined

Overview

Kelp, we need somebody! In this challenge, participants are invited to help map and monitor kelp forests—underwater habitats that cover large swaths of ocean coastlines around the world and are essential to many ecosystems and species.

Giant kelp serves as a foundation of many coastal marine ecosystems, providing habitat and food for thousands of species. Humans benefit too; it's estimated that kelp forests generate over US$500 billion annually through services like fisheries production. Yet, kelp forests face mounting threats from climate change, overfishing, and unsustainable harvesting practices. To preserve and protect these critical ecosystems, there is a pressing need to improve methods for mapping and monitoring them.

However, comprehensive and robust monitoring is a tricky task. Kelp forests are dynamic environments that respond quickly to factors such as rising temperatures, large wave disturbances, and nutrient availability. To address this, researchers at Kelpwatch.org have introduced an innovative approach by leveraging machine learning to analyze coastal satellite imagery. This breakthrough enables scientists to estimate the presence of kelp forests across expansive areas over time. This cost-effective method is a first step to comprehensively monitoring kelp forest dynamics on a large scale.

Task

In this challenge, we invite you to help advance this critical research. While Kelpwatch currently focuses on the west coast of North America, the next version aims to have a global application. Sucessful models will ideally be adaptable to new locations. Furthermore, models should exhibit robustness in the face of potential data issues, both in the satellite data and in the labels.

Your goal is to detect the presence of kelp canopy using Landsat satellite imagery and labels generated by citizen scientists participating in the Floating Forests project. Successful algorithms will not only advance scientific knowledge, but also provide a critical tool for kelp forest managers and policy makers to preserve these at-risk ecosystems.

Prizes

Competition End Date:

Feb. 21, 2024, 11:59 p.m. UTC

Place Prize Amount
1st $7,500
2nd $3,000
3rd $1,500
Bonus $3,000

Bonus Award: Top MATLAB User

We're offering a bonus prize of $3,000 to the top contributor using MATLAB for a significant portion of their submission. It can be used in combination with other tools or languages. Just make sure to fill out the Software Environment form to be considered for the prize. This is required to be eligible for the bonus prize, though we'd love to hear from you regardless of what environment you're using!

See the MathWorks competition page for more information about complimentary software licenses and learning resources for this challenge.

Note: Bonus Award recipient can also be a winner of the general prize pool. MathWorks employees are not eligible for prizes.


How to compete

  1. Click the "Join the competition" button in the sidebar to enroll in the competition.
  2. Get familiar with the problem through the about page and problem description. You might also want to reference some of the additional resources from the about page.
  3. Download the data from the data tab.
  4. Create and train your own model. The benchmark blog post, which is written in MATLAB, is a good place to start.
  5. Use your model to generate predictions that match the submission format.
  6. Click “Submit” in the sidebar, and “Make new submission”. You’re in!

Prize generously supplied by our friends at MathWorks


"Kelp Canopy" banner image courtesy of Tom Bell, All Rights Reserved.