Share Email Print
cover

Proceedings Paper

Visual surveying platform for the automated detection of road surface distresses
Author(s): Thegaran Naidoo; Deon Joubert; Tapiwa Chiwewe; Ayanda Tyatyantsi; Bruno Rancati; Asanda Mbizeni
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Road distresses, such as potholes and edge cracks, are not only a source of frustration to drivers but also negatively impact the economy due to damage to motor vehicles and costly ro6ad repairs. Regular and rapid pavement inspection and maintenance is vital to preventing pothole formation and growth. To improve the efficiency of maintenance and reduce the cost thereof, the Visual Surveying Platform (VSP) is being developed that will automatically detect and analyse road distresses. The VSP consists of a vehicle mounted sensor system, consisting of a high speed camera and a Global Positioning System (GPS) receiver, and an analysis and visualization software suite. The system extracts both a visual image and the coordinates of a detected road defect from recorded video and presents it in an interactive interface for use by technical experts and maintenance schedulers. The VSP automatically detects and classifies road distresses using a two-stage artificial neural network framework. Video frames first undergo hue, saturation and value (HSV) colour space conversion as well as a spatial frequency transformation before being used as inputs to the neural networks. A road detector neural network first classifies which section of the image contains the road, after which a distress detector neural network identifies those road regions containing defects. Although the VSP can be adapted to detect any type of road distress it has been trained to specifically detect potholes. An initial prototype of the VSP was designed and constructed. The prototype was also trained and tested on real-world data collected from provincial roads.

Paper Details

Date Published: 23 June 2014
PDF: 10 pages
Proc. SPIE 9257, Sensors, MEMS and Electro-Optical Systems, 92570D (23 June 2014); doi: 10.1117/12.2066116
Show Author Affiliations
Thegaran Naidoo, Council for Scientific and Industrial Research (South Africa)
Deon Joubert, Council for Scientific and Industrial Research (South Africa)
Tapiwa Chiwewe, Council for Scientific and Industrial Research (South Africa)
Ayanda Tyatyantsi, Council for Scientific and Industrial Research (South Africa)
Bruno Rancati, Council for Scientific and Industrial Research (South Africa)
Asanda Mbizeni, Council for Scientific and Industrial Research (South Africa)


Published in SPIE Proceedings Vol. 9257:
Sensors, MEMS and Electro-Optical Systems
Monuko du Plessis, Editor(s)

© SPIE. Terms of Use
Back to Top