Internship experience report at mundialis
In the following we present a report by our last intern Alicja Balfanz, who illustrates how her internship at mundialis has been taken place. We would like to thank Alicja very much for her great efforts during the last six months and look forward to seeing her again.
“My master’s degree in geoinformatics at the University of Muenster (WWU) offered the opportunity of a practical semester, which is meant to be used to extend the contents of the studies with practical experience and to deepen theoretical knowledge as well. Due to my personal interest in environmental topics and satellite imagery my search led me to mundialis GmbH & Co. KG in Bonn.
It was not only the first impression that mundialis would meet my expectations – the internship of half a year duration was filled with satellite images, geodata, programming – and occasionally some cake when it was someone’s birthday. Together with the friendly sister company terrestris located in the same building and partly in the same offices, the two companies represent a large store of knowledge from which I profited as much as I could.
One of the fields of activity of mundialis in which I could immerse myself was the land use classification of satellite data. I got to know different methods of classification and learned how to prepare satellite data and then automatically classify them with machine learning algorithms.
For the land use classification project, I automated many steps with Python, consolidating my knowledge of the programming language and learning about many packages which can be included in Python for processing geodata. I also learned about the limitations of certain applications and how to avoid faulty data processing during the internship.
Fig. Example of a land cover classification: On the left is a true color representation of a section of a Landsat-8 satellite image from 21.07.2013 and on the right is a section of a classification of the image with 6 land cover classes: Dense vegetation, medium dense vegetation, no vegetation, wood vegetation, water and urban areas. (Classification with Random Forest, with a maximum of 500 trees and a minimum split of nodes set to two)
One highlight was the opportunity to participate in the Global Landscape Forum conference in Bonn and to discover other areas where the processing of geo-data is elementary and further fields of application for mundialis. Since the company consists of a team that likes to learn continuously, new topics are very welcome and participation in conferences such as the Global Landscape Forum is not uncommon. This is a great aspect of mundialis, because you always get influences from outside and everyday life does not get boring.
Not only the diversified work areas, but also a good atmosphere among the employees of mundialis and terrestris, made my practical semester a very good experience full of knowledge accumulation and getting to know interesting people. I am very curious to see how mundialis develops and would like to thank you for the wonderful time. — Alicja Balfanz“
Comments are currently closed.