A company needed to segment satellite images to detect environmental changes. Learn how Aya’s team enabled accurate land-use change detection.

Overview

Orbiting 1,200 miles above Earth, hundreds of satellites zip around at 17,000 miles per hour, capturing incredibly detailed images of our planet. This technology helps us monitor and understand our world more deeply than ever before.

One key application of these satellite images is tracking environmental changes, like deforestation and natural disasters such as wildfires. These changes can lead to serious issues, including loss of wildlife, increased flooding, and other extreme weather events.

Our Client, a U.S. company, needed to create a model to detect and analyse changes in the environment. Their data scientists required detailed images of land to classify different land uses accurately.

Challenge

Aya Data was tasked with classifying different types of land use—such as roads, rivers, lakes, residential areas, industrial zones, and natural terrain—within 2,000 high-resolution satellite images of South America. These images spanned several years.

A major challenge was accurately identifying deforestation caused by human activity. This required specialised knowledge in agriculture and forestry, as well as an understanding of the environmental context of the images.

Due to the complexity and scale of the images, each one had to be analysed both in parts and as a whole to provide context, especially for large features like man-made rivers originating from hydroelectric plants.

Solutions

Aya Data’s experts in Ghana meticulously labelled the 2,000 high-resolution satellite images. Our experienced team ensured precise labelling within the project’s timeline and at a cost-effective rate compared to other providers.

Results

The labelled images were used to train a custom computer vision model that now accurately detects changes in land use across South America.

This model will be featured in an upcoming research paper, which aims to influence policy and investment in sustainable farming practices in the region.

Interested in precise ecological monitoring? Contact us today for a free quote on how our advanced solutions can help you track and manage environmental changes effectively.

Disclaimer: Aya Data is committed to client confidentiality and will not disclose any client or project-specific information. As such, names and specific details are kept anonymous in our case studies.

  • Category:
    Data Annotation
  • Industry
    Agriculture
  • Headquarters
    Accra, Ghana

Preparing Self-Driving Cars for the Rise of E-Scooters

Collecting and Processing Data on a 3000Ha Cocoa Field for Automated Crop Monitoring