Functie

Background

The field of neural rendering has seen significant growth following the release of the seminal paper on neural radiance fields (NeRF) (Mildenhall, et al., 2020), which presented a novel technique for deriving and rendering 3D representations of real-world scenes from 2D images by means of neural networks. Since the introduction of NeRF, many follow-up works have emerged proposing improvements in multiple directions, e.g. by seeking new ways to accelerate training/rendering (Instant-NGP, KiloNeRF) or to enable the representation of larger scenes (MegaNeRF). However, these advancements present new challenges, especially when it comes to improving speed of training / inference without compromising image quality / resolution.

Assignment

In this project the goal is to optimize the method used to create a 3D representation of a part of an aircraft that is due for inspection; this 3D representation will be visualized in a dashboard for a remote inspector that is offering support to the onsite ground engineer. There are many challenging aspects that have to be considered. The part that has to be inspected may be geometrically complex, captured with suboptimal lighting, and may be highly reflective. Moreover, the number of pictures that can be taken should be as low as possible, since an on-site engineer may not be in the position to take a large set of photos due to time constraints. Lastly, the time needed to obtain the representation needs to be reduced to a minimum while the resolution should be as high as possible to facilitate detailed inspection of the part. Fortunately, recent improvements focused on using neural primitives such as Signed Distance Functions (SDF), point clouds (Gaussian Splatting) and plenoctrees/plenoxels, show promising results when it comes down to training and inference speed as well as maintaining quality of output.

In this project, the student will research neural primitives to generate (neural) radiance fields that are fast, produce high resolution output and is (ideally) easy to render (with, for example, traditional rendering pipelines).  The project is focused on aviation and offers you the opportunity to work on cutting edge neural rendering techniques and aviation maintenance. The outcome of the work will contribute to a more environmentally friendly and sustainable aircraft maintenance operation.  Furthermore, your work will directly contribute to the prestigious BrightSky project which focuses on optimizing aircraft maintainance and making it more environmentally friendly.

Result

• A master thesis report, including recommendations on how to integrate NeRF in aerospace operations.
• A presentation where you present your work to your colleagues at the NLR

Duration

The duration of this master thesis internship is in the range of six to nine months.

Profiel

Profile

• A motivated master student in Computer Science, Artificial Intelligence, Data Science, Aerospace Engineering or a related field.
• Solid Python skills, with preferably experience or affinity with OpenCV
• Experience or affinity with computer vision
• Assertive and self-motivated, able to be part of the project team and also proceed individually

Arbeidsvoorwaarden

What we offer

• A challenging graduation project/internship in a high-tech result orientated work environment
• Weekly supervision and availability of the technical staff for support
• An internship allowance
• Working in an actual R&D project as part of the team
• Internship results to be used in the current and future projects

Informatie

About NLR

Royal NLR has been the ambitious research organisation with the will to keep innovating for over 100 years. With that drive, we make the world of transportation safer, more sustainable, more efficient and more effective. We are on the threshold of breakthrough innovations. Plans and ideas start to move when these are fed with the right energy. Over 750 driven professionals work on research and innovation. From aircraft engineers to psychologists and from mathematicians to application experts.

Our colleagues are happy to tell you what it’s like to work at NLR.

This assignment will be managed by the Modelling & Simulation group within the Aerospace Operations Training & Simulation (AOTS) department.

Solliciteren

Interested?

Send your application, together with your motivation letter and CV to Chihab.Amghane@NLR.nl and we will contact you as soon as possible.