In a study published in the Journal of Applied Geography, Jacek Niesterowicz and Professor Tomasz Stepinski used machine vision – a subfield of computer science devoted to analysing and understanding the content of images – to analyse part of the US Geological Survey’s National Land Cover Database 2006. The algorithms they developed allowed them to discover and differentiate 15 distinct landscape types across a section of the map that covered an area of northern Georgia, selected because of its diverse patterns of land cover. They were even able to separate forests by their domination by different plant species.
‘Before now, people would do this mapping by hand, but if you had ten maps drawn by ten people, they would all be different,’ Stepinski said.
‘It’s an entirely new way to conduct geographic research,’ he continued. ‘By leveraging technology developed in the field of computer science, it’s possible to make geography searchable by content. Using this technique, for example, we can quickly discover (using web-based applications on our website) that farms in Minnesota are, on average, larger than farms in Ohio, and ask why that is.
This story was published in the December 2013 edition of Geographical Magazine