Earth Observation Intern: On-Board Hyperspectral Processing

EU Satellite Centre (SatCen); Noordwijk; On-site; EUR 13,000 - 20,000; Full time

Job summary

ESA's Copernicus Hyperspectral Imaging Mission for the Environment (CHIME) internship at ESTEC in Noordwijk invites university students to help develop event-aware on-board processing, from detection to operations. You will join with system and satellite engineering teams to integrate machine-learning prototypes into an end-to-end mission simulator and evaluate performance.

You will define processing modes for land/ocean/night, connect detection with compression and downlink priorities, and

Qualifications

• Background in machine learning, computer vision, signal processing or remote sensing.
• Practical experience with Python and a deep-learning framework.
• Familiarity with hyperspectral data, transfer learning, model compression or uncertainty estimation is an asset.

Responsibilities

• define configurable processing modes for land, ocean and night-time observations;
• integrate existing detection methods into the mission end-to-end simulator;
• connect detection masks with selective compression and data-prioritisation policies;
• quantify detection performance, false alarms, scientific-data preservation and potential downlink savings;
• assess robustness across different geographic regions, seasons, illumination conditions and representative instrument effects;
• define operating limits, verification criteria, traceability information and fallback behaviour;
• deliver reproducible software, trained models, documented datasets and a technical report suitable for public presentation or publication.

Skills

Machine learning

Python

Deep learning

Remote sensing

Computer vision

Hyperspectral data

Transfer learning

Model compression

Uncertainty estimation

Education

Master's student

Tools

TensorFlow

PyTorch

NumPy/SciPy

Jupyter

Job description

ESA's Copernicus Hyperspectral Imaging Mission for the Environment (CHIME) internship at ESTEC in Noordwijk invites university students to help develop event-aware on-board processing, from detection to operations. You will join with system and satellite engineering teams to integrate machine-learning prototypes into an end-to-end mission simulator and evaluate performance.

You will define processing modes for land/ocean/night, connect detection with compression and downlink priorities, and