Transportation departments are required to survey for these bats before conducting repair or replacement projects.

Hundreds of thousands of bats from about 40 species routinely nest under US bridges, and many of them are endangered or threatened. Transportation departments must mandatorily check for them before commencing bridge repairs. It’s a painstaking survey, with clues to the animals’ presence being extracted from their guano, droppings and stains. But AI is making the task easier. (TechXplore)

A recently published study by the University of Virginia’s MOB Lab in collaboration with the Virginia Transportation Research Council found that photographs and computer vision techniques using artificial intelligence can detect the presence of bats on bridges automatically and accurately.

That’s a huge improvement from visual inspection from humans, who also need comprehensive training. Besides, guano may be lodged in inaccessible places or may fall to the water from the underside of the bridge.

Threatened animals, bridges and AI

However, AI models can now be trained to tell the difference between stains produced by bats or other sources, such as water seeps, rust staining, asphalt leaching or other types of structural deterioration.

The researchers trained a model on a database of images of bridges with and without the tell-tale signs of the presence of bats. Once the model was trained to pick up the signs that indicated bats, it could identify their presence under a bridge from images with 90% accuracy.

Users can also upload images of stains, on or under a bridge or structure, to a web application and generate results from the AI.

Related Story: Artificial Intelligence Will Monitor Cities’ Aging Infrastructure

Image: Urban bats take flight from the Congress Street Bridge in Austin, Texas (Credit: Flickr)