EPFL develops algorithm to rate scenic landscapes

Researchers behind an algorithm to measure how scenic a region is say it could have benefits for environmental and conservation efforts.

The research by the Swiss Federal Technology Institute Lausanne EPFL and Wageningen University in the Netherlands aimed “to incorporate people’s aesthetic enjoyment of a landscape” into more traditional environmental appraisals, EPFL said on Friday.

Why? Human enjoyment of outdoor activities, especially in beautiful areas (e.g. “rolling hills covered in yellow and lavender”) can be good for our physical and mental health, which is why it should be a factor in ecosystem assessments, EPFL said.

Researchers first trained an algorithm on 200,000 photos of UK landscapes, which were rated according to “scenicness” in a crowd-sourced survey. This algorithm then appraised some 9 million Flickr pictures. The results were then compared with those given by a more conventional rating model based on environmental indicators.

Colour coding a map of the UK according to scenicness, the study found that at a resolution of 5 km2, both models came up with roughly the same results: national parks, highland areas etc. were seen as aesthetic, while urban areas like Glasgow or London were not.

At a closer resolution of around 500 m2 however, the new model was “more accurate”: for example, the area around Heathrow airport, viewed as scenic by the traditional model, was “decidedly un-scenic” according to the crowd-sourced model.

“The use of social media-based data provides a combination of information about the state of the environment and how people interact with it,” said Devis Tuia, associate professor at EPFL’s Environmental Computational Science and Earth Observation Laboratory. 

“Such information has never before been obtained with such a high degree of accuracy.”

The researchers want to extend their project to other countries and explore how it can help “support environmental conservation policies across Europe”.

The full study can be found in Nature magazines “scientific reports”.