ARTIFICIAL intelligence (AI) and machine learning is being used by Network Rail to better understand and monitor what species are living along the UK’s railway.
The initiative comes in partnership with the Zoological Society of London (ZSL), with the organisations working alongside Google to develop methods for monitoring species on line-side land.
In partnership with the People’s Trust for Endangered Species (PTES), equipment was set up in woodland near Calke Abbey in Derbyshire in June. Images, videos, and acoustic recordings were gathered and will now be used to develop new AI and machine learning algorithms for identifying dormouse presence remotely.
The data is stored and analysed on Google Cloud, offering the potential for Network Rail to make ‘significant’ savings of time and money compared with manually checking dormouse nest boxes for their presence throughout the season.
Network Rail ecologist, Aline Gomes, said, “Understanding the plant and animal life around our railway plays a big part in helping us manage our railway sustainably. Our regional ecology team wanted to find innovative ways of monitoring the wildlife on our line-side while maintaining a safe distance, and working with ZSL has allowed us to achieve that. The knowledge we are gaining is already improving the way we manage our land to benefit the different species that live along our railway, including dormice.
“The railway can provide a great habitat for dormice and by gaining a deeper understanding of the species, we can help them by improving how and when we manage the lineside habitat to contribute towards the conservation of the species.”
ZSL’s project manager, Rachael Kemp, commented, “Unlocking the power of cutting-edge technology, our remote, automated methods provide an exciting, cost-efficient opportunity to study dormice living near railways. We are using images, videos and audio files of dormice collected at Calke Abbey and Cowden to train machine learning algorithms to help Network Rail understand which nest boxes along the tracks are being used by this beloved mammal and easily monitor them over time. This will help Network Rail understand how lineside maintenance and conservation strategies are affecting local populations”.
Network Rail’s biodiversity strategy manager, Dr Neil Strong, added, “Managing our railway lineside is important not just because we have to keep trains safe but also because those linesides provide wildlife corridors for species such as dormice, which would struggle to move between habitats otherwise.
“We have to manage them though – it’s not about letting them grow wild. Trees can fall in front of trains and grow in front of signals, so we work to cut back those where we have to, and make the right habitat, ideally away from the tracks, to give good cover for creatures who need it.”
Google Cloud’s executive sponsor, Omer Mahmood, said, “As part of our ongoing commitment to help our customers use Google Cloud’s AI and ML to drive important climate and conservation efforts, it has been our privilege to be a key technology partner with ZSL and Network Rail on this project.
“We look forward to new opportunities for collaboration, enabling Network Rail to meet their ambitions for a net gain in lineside biodiversity, by improving the efficiency of their data pipelines and monitoring at scale, across their rail network in the years to come”.
