Trains delayed by 'leaves on the line' might soon be a thing of the past as an AI system is trialled to predict build ups on the line and warn of encroaching plants.
The artificial intelligence studies footage of plants near the line taken from trains and attempts to spot when leaves change colour, indicating that they might fall.
It can also warn of fallen trees and when vegetation growth might soon obstruct the path of trains and lead to delays.
The project is one of 24 high-tech schemes that have today been funded a total of £7.8 million ($9.9 million) by the UK government to improve the nation's railways.
Scroll down for video
Trains delayed by 'leaves on the line' might soon be a thing of the past as an AI system is trialled to predict build ups on the line and warn of encroaching plants (stock image)
Slippery rails — commonly referred to as 'leaves on the line' — result when build ups on the track led to trains not being able to grip the rails properly.
This can cause locomotive wheels to slip when trying to accelerate and slide when trying to brake.
The most common cause of build up comes from from moist leaves, which cling to the tops of the rails.
The wake of trains as they pass through the air actually helps pull in leaves towards the tracks.
Leaves build up incrementally, as they are not worn away fast enough by the passage of trains over them.
The problem has become worse since the introduction of disc brakes, which replaced the brake shoes which would help remove leaves by scraping them from train wheels.
Developed by London-based tech firm Hack Partners, the artificial intelligence system works through a camera in the driver's cab which captures footage of any vegetation growing near the railway line along the train's journey.
The algorithm analyses this video to predict where trees might be about to shed their leaves, where branches could grow into the path of the train and where trees might be about to fall across the line.
It does this by analysing the colour of leaves to determine if they're likely to fall onto the tracks, uses a boundary system to check if vegetation is encroaching into the path of the trains.
The AI also compares current and past footage to spot any trees that appear to have moved — for example, during storms — that might need to be cut down.
This advance warning will then allow Network Rail to tackle the problematic vegetation before it leads to delayed trains.
'Encroaching vegetation and fallen leaves on the line can lead to significant delays for the millions of passengers who use and rely on the railway every day,' a spokesperson for Network Rail told The Times.
Nearly 19,000 incidents of trees and branches falling onto UK railways lines were logged from 2017–18, leading to the cancellation of more than 1,750 services in the former year alone.
Meanwhile, it is believed that leaves on the line resulted in around 3,261 hours' worth of delays in 2017.
'We work hard to reduce these delays while also making better use of environmental data to improve net biodiversity and contribute to the government’s targets on habitat and woodland creation,' the Network Rail spokesperson added.
It is intended that the system will not lead to more trees being chopped down by Network Rail but instead will allow for a more targeted approach to managing line-side vegetation, said Hack Partners founder and CEO River Tamoor Baig.
The AI system will begin a