Navigation und Service

PhD position: remote sensing and machine learning

21.06.2017 - Monash University, Melbourne (Australia)

The PhD is about developing machine learning methods into time series classification, and apply this to series of satellite images (Landsat-8, Sentinel-2).

Where: Monash University (Australia's largest University), located in Melbourne. The student will be part of the Monash Centre for Data Science, which counts 60+ academic in Machine Learning and Data Mining.

Starting date: Now and up to Jan 2018

Skills: If the student is amazing, then it doesn't matter, we'll train him or her on-site. If the student is solid, then he or she could either be from the Remote Sensing community with good maths and an ability to code, or from the Machine Learning community with good maths and coding skills as well as willing to learn about remote sensing.

Scholarship: The PhD student receives $AUD 2,150 per month (tax-free) for 3 years + PhD fees covered

A bit more detail about the topic:
Earth Observation (EO) satellites have long provided indispensable perspective on the status of large-scale agricultural, environmental, and climate problems to inform our decisions. We are entering a new era in Earth Observation, with latest-generation satellites starting to image Earth frequently, completely, in high-resolution, and at no
charge to end-users; introducing unprecedented opportunities to monitor the dynamics of any region of our planet over time and revealing the constant flux that underpins the bigger picture of our world. These satellites produce vast streams of unprecedentedly rich data in the form of time series, enabling the creation of nuanced, temporal land-cover maps that describe the evolution of an area over time.

The challenge:
It is not yet possible to tap the full value of this data, as existing machine learning methods for classifying time series cannot scale to such vast volumes
of data. Temporal land-cover maps assign unique labels to geographic areas, describing their evolution over time. One of today’s key challenges is how to
automatically produce these maps from the growing torrent of satellite data, to monitor Earth’s highly dynamic systems. Presently, state-of-the-art research into time
series classification lags behind the demands of the latest space missions, which produce terabytes of data each day. Why? Most of the research into time series
classification has been done with datasets that hold no more than 10 thousand time series [a]. In contrast, the Sentinel-2 satellite gathers over 10 trillion time series, capturing Earth’s land surfaces and coastal waters at resolutions of 10–60m. This Project aims to create the machine learning technologies necessary to analyse series of satellite images, and to produce accurate temporal land-cover maps of our planet. Potential high-value applications for Australia include fire prevention, agricultural planning, and mining site monitoring and rehabilitation.

The student will be co-supervised by Christoph Rudiger on the Remote Sensing side and by Geoff Webb on the Machine Learning side.

Contact: Chris Rüdiger,

[a] Y. Chen, E. Keogh, B. Hu, N. Begum, A. Bagnall, A. Mueen and G. Batista (2015). The UCR Time Series Classification Archive. URL