This month I attended the Geopython 2023 conference in Basel, Switzerland. This was a well-organized conference with between 100-200 in-person attendees. The focus of most of the talks was on geospatial applications, particularly the rise of machine learning algorithms is geospatial data processing.
For example, Mikolaj Czerkawski, PhD student at the University of Strathclyde in Glasgow, gave an interesting talk entitled, “Generating Clouds in Satellite Images with SatelliteCloudGenerator.” Why would you want to generate clouds to obscure your satellite images, particularly in Scotland where there are plenty of clouds already? You need lots of training examples to teach machine learning algorithms how to remove them. It turns out generating clouds isn’t easy because their impact on satellite imagery depends on the spectral band. Very interesting!
Another talk I found fascinating was “Automated weeding of railway tracks” presented by Sven Schmitz-Leuffen. Weeding railways may not sound interesting at first, but the system involves a specially designed train engine equipped with GPS sensors, trained AI plant detectors and a cartographic computer controller that ensures herbicide is not applied in areas spraying is not allowed, such as populated areas, over waterways, etc. All of this is done in real time as the train is cruising along the tracks — a very clever application of python and geospatial processing!
There were many more fascinating talks and workshops. You can find the full list here.
Of course, Switzerland is such a beautiful country, and the trains make it so easy to get around that I combined the conference with a vacation, including a trip to Zermatt to view the Alps, and to tour the Lindt Chocolate factory outside Zurich. Overall, a wonderful trip!
