Master's Thesis: Completion of gaps in paired point cloud and image data capturing urban environments
One of the main challenges when working with scanned point cloud and image data to create a digital twin of a certain environment is how to compensate for missing or problematic parts that result from occlusions, distortions or noise. Neural Radiance Field (NeRF) models have recently proven to be a viable alternative to photogrammetry that can at times additionally fill in the kinds of mentioned gaps in the available data with contextually-relevant content. This project seeks to investigate the application of NeRFs or similar models in enhancing the reconstruction of urban scenes by leveraging image content to complement the existing 3D information from the associated point clouds.
- Department
- ML Development
- Locations
- Göteborg
- Remote status
- Hybrid
Colleagues
Göteborg
What we value.
We’re all about collaboration. That’s because we’re solving a puzzle that cannot be completed alone. It takes trust, courage, creativity and transparency to deliver on our promise. Our team buys into this, so our customers buy into it. It’s just our way of staying on the same page.
About Repli5
We enable the development of autonomous vehicles with synthetic data. Our software reduces the cost and time required to train autonomous vehicles.
Master's Thesis: Completion of gaps in paired point cloud and image data capturing urban environments
Loading application form
Already working at Repli5?
Let’s recruit together and find your next colleague.