crackDetector

######################################################################################### Graph network refining for pavement crack detection based on multi-scale curvilinear structure filter

(paper under review)

#########################################################################################

1.Installation.

a) This released is written for the Matlab interpreter (tested with versions R2017b) and requires the Matlab Image Processing Toolbox. Due to the policy of cooperative enterprises, we cannot release source codes, and encrypted them using ’.p’ files.

b) As geodetic distance transformation and seeded topological watershed transformation are used to generate over-complete potential crack paths, the following two Toolbox are required:

Dirk-Jan Kroon's Accurate Fast Marching Toolbox. It can be downloaded at: https://www.mathworks.com/matlabcentral/fileexchange/24531-accurate-fast-marching

The DIPimage toolbox is requred for the seeded topological watershed transformation . It can be downloaded at: http://www.diplib.org/download

2.Getting Started.

• Make sure to carefully follow the installation instructions above.

• Please see "Demo.m" to run demos and get basic usage information.