Image Processing / Machine Learning Internship

Job description

False positives and negatives reduction in a satellite based pipeline monitoring system

OrbitalEye is a high-tech company based in Delft, the Netherlands. Its core product, PIMSyS, is a complete solution for reliable and cost-effective pipeline monitoring. It combines accurate and up-to-date monitoring data, acquired by synthetic aperture radar (SAR) satellites, with smart tools for planned third-party activities and inspection planning.

PIMSyS is based on Synthetic Aperture Radar Coherent Change Detection (SAR-CCD) technology. It enables us to detect activities in the vicinity of the assets being monitored, independent of cloud cover and daylight at virtually any location on Earth. For this, SAR-CCD uses satellite images taken at different times to map locations where things have changed.

A crucial step in maximizing the value for our customers, is to further process the SAR-CCD change maps in order to extract all relevant changes, while minimizing the number of false positives (agricultural activities, parking/loading areas, etc.). This is a challenging goal, since there exists a delicate balance between reducing false positives and not introducing false negatives (relevant changes not being detected). Therefore, the further enhancement of our systems performance with respect to these false positives and negatives, is and will remain an ongoing R&D topic for OrbitalEye.

At this moment, we are looking for students who would be interested to work on this challenging problem together with our R&D Engineers as part of their master’s thesis project and/or internship (minimum 6 months). The project and thesis should in some way contribute to the reduction of false positives and/or negatives, but can still cover a wide-range of sub-topics like land-use classification, radar image segmentation, machine learning, data fusion with spectral/optical satellites, image processing, validation with ground truth and advanced masker and filter design.


Potential candidates should have adequate knowledge of Matlab and it us beneficial to have some experience in Python.

If you are interested and/or have any questions regarding this opportunity, please don’t hesitate to contact:

Sven van Haver

-          Matlab/Python

-          Signal/data processing and analysis

-          Synthetic Aperture Radar (SAR)

-          Change Detection

-          False alarm reduction

-          Filter design

-          Image processing

-          Landuse classification (unsupervised)

-          Segmentation

-          Machine learning

-          Radar scattering cross section

-          INSAR, DINSAR, CCD

-          Level 1 SLC (complex)

-          Earth Observation (EO), Sentinel, Copernicus

We are a fun and intelligent team that provides both flexibility and room to grow. If you are interested in this role, please apply below or send an email to - we look forward to hearing from you!

Apply for this job

We will handle your information with care. Please see our Privacy Statement for more details.