Share this article:
How Smartphones Are Smoothing Out Bumps in the Road
A new Android-based application automatically identifies road surface distress in real-time.
Potholes and cracks are annoying, driving road users to distraction, but they also have serious implications for road safety across the globe. In 2015, the US Department of Transportation reported that 35,092 people died in motor vehicle traffic accidents. In the same year in Europe, more than 26,000 people were killed on the road. While driver behaviour and error is the major cause, distressed (cracked, rutted or otherwise distorted) road pavements (or surfaces) are also contributing factors. As a 2014 study by the European Parliament points out, “there are accidents caused directly by the poor condition of the road network, but there are also accidents caused by drivers’ behaviour in reaction to the condition or the design of the road.”
Fortunately, teams of certified inspectors and structural engineers regularly check civil infrastructure, such as roads and bridges, in an attempt to improve road safety. The teams report on the size and severity of road surface distress, suggesting causes and remedies. This information is then used to plan urban road maintenance and rehabilitation (M&R) projects and pavement management systems (PMS).
Traditionally, the reports have been generated manually, and are based on visual inspections. This work is very time-consuming, and road agencies are increasingly looking to Computer Vision (CV) methods to automate, or semi-automate, the process using mobile devices. CV acquires and analyses digital images to produce information about objects. Mobile-based CV solutions have collected data about pavements, but were not able to analyse information in real time until now.
A paper published in the journal Advanced Engineering Informatics, describes how researchers from the University of Roma Tre developed an Android-based application that automatically identifies road surface distress in real time. The new Automatic Pavement Distress Recognition (APDR) system uses a smartphone or tablet video camera to record potholes and cracks. It then automatically identifies the distress category for each image using cascading classifiers – computerised stages of object recognition – which are defined in an OpenCV (Open Source Computer Vision) library embedded in the application.
These classifiers allow the application to quickly and accurately identify the three most common distress types: potholes, longitudinal and transversal cracks, and fatigue cracks. The system not only recognises distress incident types and numbers, it also records the date and time of detection, and provides images and a GPS location. All results are outputted to the end user in real time. Researchers can also feed the information into a database, and use it to monitor road surface conditions over time, filtering results by type and location.
Francesco Benedetto, a co-author on the paper, says the new system will improve inspection efficiency and accuracy, and shows how communications technologies can be exploited in surprising ways. “It demonstrates how a common object of our everyday life (our smartphone or our tablet PC) can be used for several purposes,” he says. In this case, to improve our road safety.
Read Article free online until 25th May 2018
Tedeschi, A., Benedetto, F.: “A real-time automatic pavement crack and pothole recognition system for mobile Android-based devices,” Advanced Engineering Informatics (2017)
You can view Elsevier’s Engineering books and journals on Elsevier.com
Engineering brings science and technology out of the lab and into the real world. Often without thinking about it, we engage every day with technology that is the product of careful, precise design and execution by engineers in electronics, optics, and communications; embedded systems; automotive, aerospace, and marine; mechanical; and many other disciplines. For decades, Elsevier has maintained and grown extensive collections in these and other cutting-edge areas, like biomechanics and nanotechnology, through our trusted imprints: Newnes, Academic Press, and Woodhead Publishing. In addition, our powerful online platforms like Knovel and Engineering Village help streamline research and development processes for users around the world.