Optics and Photonics Journal, 2013, 3, 99-102
doi:10.4236/opj.2013.32B025 Published Online June 2013 (http://www.scirp.org/journal/opj)
Image Preprocessing Methods to Identify Micro-cracks
of Road Pavement
Hui Wang, Zhang Chen*, Lijun Sun
Key Laboratory of Road and Traffic Engineering of the Ministry of Education, Tongji University, Shanghai, China
Received 2013
ABSTRACT
Standards of highway conservation and maintenance are improved gradually following the improvement of require-
ments of road service. Before obvious damage such as obvious cracking (block，transverse, longitudinal ) and rutting
emerge, inconspicuous distress (micro-cracks, polishing, pockmarked) is generated previously. These inconspicuous
distresses may provide basis and criteria for pavement preventive maintenance. Currently most of preventive conserva-
tion measures are determined by experienced experts in maintenance and repair of road after site visits. Thus method is
difficult in operation, and has a certain amount of instability as it is based on experience and personal knowledge. In
this paper, camera and laser were used for automated high-speed acquisition images. Methods to preprocess pavement
image are compared. The pretreatment method suitable for analyze micro-cracks picture is elected, an effective way to
remove shadow is also proposed.
Keywords: Pavement Distress Automatic Detection; Inconspicuous Distress; Micro-crack; Laser Light Image;
Image-preprocessing
1. Introduction
Preventive maintenance should be integrated into pave-
ment manage system to determine the proper time for
operation. The U.S. scholar Zimmennan K.A had pros-
pected (2003), “transportation agencies want to combine
preventive maintenance and pavement management, which
requires pavement condition surveys and calculation of
indicators”.
Inconspicuous distresses (micro-cracks, polishing,
pockmarked) are generated previously before obvious
distress generates. These distresses are difficult to be
found and recognized, especially micro-cracks (millime-
ter level), developed from isolation of aggregates of road
surface. Polishing and pockmarked are difficult to be
quantified itself, traditional recogn ized human vision, but
a manual surface distress survey is subjected to many
limitations, such as not repeatable, too subjective and
high human cost.
Line laser light and camera has been used for detection
rutting by calculating the curve changes of the line
(Yuntao Wei etc. 20 09).
In this paper, a laser line and a camera is used to detect
micro-cracks and its width. Image processing method is
descripted b elow.
2. Background Correction
The composition of the road pavement image is mainly
generated by three aspects. One is background lighting
signal, including shadow information. One is noise signal,
mainly caused by the road surface texture. Last one is
distress signal, which consists of pavement distress,
stains, marking.
The purpose of image background correction is to ex-
clude the backgrou nd interference br ought by light inten -
sity uneven. Pavement distress information should be
assured during this process. The whole image matrix is
made into small matrix of the same size, using matrix
grayscale characteristics of the image background, image
relative ideal background is calculated and analyzed.
Reference background method is used to correct image
background, results are shown in Figure 2-1, Figure 2-2.
3. Image Noise Reduction and Image
Enhance
Noise caused by background non unifo rm disappeared by
background correction, but still contains a small amount
of random noise, which needs further eliminate. Noise
reduction effect by 10 different noise-reduction methods
is shown in Figure2-3.
1, Original image; 2, Low-pass convolution template,
suppose original image , is impulse re-
sponse of low-pass filtering( matrix),
,

H. WANG ET AL.
100
Original image
After background-correction
Figure 2-1 Road pavement image.
Figure 2-2 Gray values and Gray histogram of road pavement image.
the template is “Box template”, ; 3, High-pass filtering, ; 4,

H. WANG ET AL. 101
High-pass filtering, ; 5,
High-pass filtering, ; 6,
High-pass filter, ; 7, High-pass
filtering, ; 8, Mean Filtering; 9,
Threshold filtering; 10, Median filtering It can be seen
through Figure2-4, method 2 and method 7 are better.
However, road distress signal is very weak, therefore
distress information should be assured, enhancing dis-
tress information, weakening background information.
4. Distress Image Segmentation
There are lots of methods to obtain optimal threshold,
such as range method, the largest category of variance
method. In this paper, region without distress is separated
from calculation, then range method is used, result is
shown in Figure2-5.
5. Conclusions
By analyzing the gradation characteristics of the pave-
ment micro-cracks image, background correction method
suitable for this type of image is selected, also effect of
the shadow is effectively eliminated.
Figure 2-3. Images after different enhance methods.
Figure 2-4. Gray values of road pavement image.

H. WANG ET AL.
102
Figure 2-5. Image segmentation of road pavement image.
On the basis of the background correction, several
noise deduction methods are compared. Low-pass con-
volution template method is better.
Finally, distress image is segmented.
REFERENCES
[1] K. A. Zimmerman, “Pavement Management Perspective
and Integrating Preventive Maintenance into A Pavement
Management System,” Transportation Research Record,
No.1827, 2003, pp. 3-9.
[2] J. N. Meegoda, S.Y. Gao, S. Liu and N. C. Gephart,
“Pavement Texture from High-Speed Laser for Pavement
Management System,” International Journal of Pavement
Engineering, 2012, pp.1-9.
[3] M.-T. Do, Z. Z. Tang, M. Kane, Francois de Larrard,
“Pavement Polishing—Development of A Dedicated
Laboratory Test and Its Correlation with Road Results,”
Wear, No. 263, 2007, pp. 36-42.
[4] Y.T. Wei, H.Y. Hong, X. H. Zhang and J. Y. Yu, “A New
Method for Automatic Detection of Rut Feature Based on
Road Laser Images,” Multispectral Image Acquisition
and Processing, Proceedings of SPIE, Vol. 7494, 2009.