POTHOLES, cracks, gnarly tree roots 鈥� the streets are full of nasty obstacles when they fall into disrepair. But an army of online workers could soon fix that, by whizzing virtually through neighbourhoods and earmarking encumbrances for a quick response from the local council.
Jon Froehlich and colleagues at the University of Maryland in College Park have developed software that allows untrained crowdsourced workers from Amazon鈥檚 Mechanical Turk service (AMT) to zoom through the streets using Google Street View and find and label the potholes, obstructed pathways, broken kerbs and missing ramps that can stymie street access, particularly for those using a wheelchair or a walking aid. A report that includes images of highlighted problem areas can then be generated for the local council鈥檚 road-mending crews to act on.
The researchers built a prototype following interest in the idea from officials at the US Department of Transportation, who said that the project could save municipalities precious time and resources by telling repair teams the precise nature of the problem before they head into the field. This would aid scheduling and ensure that crews take the right amount of materials 鈥� such as concrete or flagstones 鈥� to the site.
Advertisement
The team tested the labelling software using six volunteers 鈥� three members of the research team and three wheelchair users 鈥� and then set the task to 400 AMT workers. After viewing , which showed how to identify problem areas by labelling them with coloured shapes, the workers accurately spotted access issues 93 per cent of the time. Froehlich will present in April at the Conference on Human Factors in Computing Systems in Paris, France.
聯Volunteers accurately spotted problem areas in streets 93 per cent of the time聰
In the next version of the system, the team wants to automate the process using computer vision algorithms. The researchers also hope to access the precise laser range-finding data that the Street View camera cars acquire as they comb the streets. Such information would increase accuracy when assessing the sizes and shapes of obstructions.
鈥淚t sounds like this project has a lot of parallels with FixMyStreet,鈥� says Myfanwy Nixon of mySociety, a non-profit organisation in London that runs , where people can report street problems. 鈥淥nline technologies are very good for people with limited mobility or a disability that prevents them from getting their voice out via other means.鈥�
This article appeared in print under the headline 鈥淥nline eyes see the way to an easier amble鈥�
