A Matlab toolbox for Road Crack Detection and Characterization
Roads are important man-made infrastructures playing a very efficient role in populations’ development, allowing the easy mobility of people, goods and merchandises. However, pavement surface exhibits distresses due to their constant usage. Hence, the maintenance of road networks is an essential task to ensure the correct pavement performance. To establish a proper maintenance policy of road networks, it is necessary to implement an adequate information system that allows supporting the management of their maintenance, capable of dealing with several interdependent data. The type and extension of pavement surface distresses are considered the most important data about pavement surface condition, necessarily to be collected during periodic road surveys. Images of road pavement surface (taken during periodic road surveys) are considered an important source of information for the quantitative and qualitative distress evaluation on road pavement surfaces, allowing the adequate identification and quantification of road pavement surface distresses. The proposed toolbox (developed under the MatLab environment) allows the automatic processing of images of pavement surface for the detection and characterization of road cracks (considered the most common pavement surface degradation found by road inspectors).
This toolbox is the result of research work developed at the Multimedia signal Processing Group of Instituto de Telecomunicações, on the detection and characterization of cracks in flexible road pavements.
· Includes a tool to help a human operator label blocks of road pavement surface images as either containing crack pixels or not, thus creating the ground-truth needed for a quantitative analysis of the results obtained;
· Capable to handle images acquired from different types of imaging sensors (active and non-active remote seining sensors);
· Allows the detection of cracks using two types of strategies: block-based (images divided into non-overlapping image blocks) and pixel based;
· Allows the implementation of a pattern recognition system to detect cracks using a two-dimensional block-based feature space (block-based approach);
· Allows the implementation of a segmentation by thresholding technique to detect crack at pixel level (pixel-based approach)
· Includes a strategy for the automatic selection of images for training a system, when pattern recognition techniques are used to detect cracks;
· Includes a crack linking strategy, allowing to group different connected components belonging to the same crack at pixel-based;
· Includes a crack type classification algorithm, to automatically characterize the detected cracks as longitudinal, transversal or miscellaneous;
· Includes a crack severity labeling procedure, based on the crack’s width computations.
· Software (version v1.5) – 114MB (Supports MatLab versions 2015b and 2016a and includes a small image database of the surface of flexible road pavements acquired during a traditional road survey and using a non-active remote sensor).
· To use the toolbox please fill in the license agreement and send a scanned copy of the signed form by e-mail to: firstname.lastname@example.org
A list of references dedicated to the road crack detection and characterization, some of them mentioned on several toolbox help files:
· Oliveira, H.; Correia, P.L.; "CrackIT – An image processing toolbox for crack detection and characterization", Proc. IEEE International Conf. on Image Processing - ICIP, Paris, France, October, 2014.
· Oliveira, H.; "CRACK DETECTION AND CHARACTERIZATION IN FLEXIBLE ROAD PAVEMENTS USING DIGITAL IMAGE PROCESSING", PhD Thesis, Instituto Superior Técnico, July, 2013.
· Oliveira, H.; Correia, P.L.; "Automatic Road Crack Detection and Characterization", IEEE Trans. on Intelligent Transportation Systems, Vol. 14, No. 1, pp. 155 - 168, March, 2013.
· Oliveira, H.; Correia, P.L.; "Supervised Crack Detection and Classification in Images of Road Pavement Flexible Surfaces" - Chapter in Recent Advances in Signal Processing, In-Tech, In-Tech, Austria, 2009.
How to reference the toolbox:
· Oliveira, H.; Correia, P.L.; "CrackIT – An image processing toolbox for crack detection and characterization", Proc IEEE International Conf. on Image Processing - ICIP, Paris, France, October 2014
How to reference the techniques included in the toolbox:
· Oliveira, H.; Correia, P.L.; "Automatic Road Crack Detection and Characterization", IEEE Trans. on Intelligent Transportation Systems, Vol. 14, No. 1, pp. 155 - 168, March, 2013
Feedback and contributions are welcome. Please provide your comments and suggestions to:
· Henrique Oliveira: email@example.com
· Paulo Lobato Correia: firstname.lastname@example.org