United States Department of Transportation
i
A novel image database analysis system maintenance of transportation facility.
-
2009-01-01
-
[PDF-309.78 KB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final report.
-
Corporate Publisher:
-
Abstract:The current project was funded by MIOH-UTC in the Spring of 2008 to investigate efficient
maintenance methods for transportation facilities. To achieve the objectives of the project, the
PIs undertook the research of various technologies of image analysis and storage. Therefore,
initially a database was developed to store various information and images obtained from the
regional roads. In this direction, a number of methods for storage of images were investigated
and compared such as storage of images into database directly versus the storage of images
into database after compression; or storage of the location of the images only in database.
In another direction, the PIs investigated various imaging technologies for the identification
of the state of various roads, since the demand for automated inspection, monitoring, and
pattern recognition for transportation applications are ever increasing. This increasing
demand is partly driven by the decreasing costs of software and hardware technologies which
allow faster access to the required information for these applications. The structured storage
provided by the use of modern database technologies provides proper mechanism to store a
variety of information artifacts which can be merged from different perspectives for
investigation. For example, combined image and textual information regarding different
attributes of a typical inspection would identify accurately the region and the proper
maintenance path. A typical inspection process, including pavement distress inspection, can
be divided into three stages: preprocessing, segmentation, classification and measurements.
Preprocessing is used to improve the quality of the input image in order to facilitate the
analysis and interpretation at subsequent stages. Important tasks in preprocessing can
include filtering for noise removal, deblurring the image, and the highlighting of specific
features, e.g., cracks on the pavement. Image segmentation is the process of dividing an
image into meaningful regions, such as objects of interest and background.
This report looks at the imaging technologies to enhance the management of pavements. The
main parameters of interest for pavement management are the pattern classification and
measurement of various parameters of crack features. The first section explains the imaging
technologies used for processing images. In section 2 the algorithm is explained and section 3
shows the simulation result and our progress in achieving the goals set in this project.
-
Format:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
