Our team and mission

Candidates interested are encouraged to visit the ESA website: http://www.esa.int

Field(s) of activity for the internship

Internship topic: Lunar Crater Dataset Analysis and Database Generation

In 2031, ESA’s Argonaut lander will provide Europe independent access to the lunar surface. One of the many challenges connected to Moon landings is the vast amount of craters of all scales. Especially in the landing region, a reliable detection of them is required in order to navigate to a safe landing site and to avoid them as hazards during the landing process. AI-based networks seem to be a very promising candidate to solve this challenge, at least for the a-priori analysis. However, as for any AI application, a ground truth training data set is required. In the ESA IMPACT project, a lunar base building game was developed and published, which combined gaming, crowd science and space exploration by integrating crater marking in lunar satellite images into the gameplay. As of today, the result comprises over 10 million crater markings by the players, making it the biggest existing lunar crater database.

While first tests of using the dataset to train an AI have already started, the need for investigating the data with respect to several quality parameters becomes obvious.

You shall work on this dataset to maximise its potential for lunar landing application, AI applications and for scientists around the globe, since such a dataset can enable research way beyond mission analysis needs. For this the dataset needs to be analyised further, since unfortunately not all players have marked all craters in an images and not all markings are correct.

The task for this traineeship is to prepare best-practise guidelines and software script to post-process and filter the initial lunar crater dataset. Your initial tasks would include:

• familiarise yourself with the game and the dataset.
• familiarise yourself with the AI algorithms state of the art for crater detection.
• generate an overview of available meta data in coordination with the game developers.
• identify potential use cases in the science community and document their requirements.
• write filter and other post-processing scripts (Python) to eliminate misleading / useless data parts.
• document post-processing guidelines to be published along with the dataset.
• propose smart algorithms to be used by the game to improve marking quality: e.g. identify super users, compare user results, majority voting, use images under different lighting conditions and compare markings, etc..

Once the dataset can be reliably used you shall try to expand the lunar crater database to other regions of the moon using e.g. the trained AI algorithms.

Behavioural competencies

Result Orientation
Operational Efficiency
Fostering Cooperation
Relationship Management
Continuous Improvement
Forward Thinking

For more information, please refer to ESA Core Behavioural Competencies guidebook

Education

You must be a university student, preferably studying at master’s level. In addition, you must be able to prove that you will be enrolled at your University for the entire duration of the internship.

Additional requirements

The working languages of the Agency are English and French. A good knowledge of one of these is required. Knowledge of another ESA Member State language is an asset. 

During the interview, your motivation for applying to this role will be explored.

You should have knowledge in the following fields:  

• experience in Python software development.
• basic understanding of AI networks.
• the ability to work independently.

A background in space exploration or geology is considered an asset.