Crack Detection And Measurement In Tennis Courts: An Edge Detection Approach
Date
2022-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
The studies provide strong evidence for the use of Computer Vision and UAV
(Unmanned Aerial Vehicle) to count the number of people in parks, but no one has
previously examined utilizing computer vision to evaluate park amenity quality
[5][6][7]. Current methods to evaluate park quality require a team to manually go
through the park and record the incivilities. Specially a task like measuring the length
of cracks could take hours of bunch of people. For this paper, we propose a method to
detect and quantify the cracks in tennis courts using computer vision and image processing techniques. The algorithm can detect and quantify the cracks in tennis
courts with 99% precision in matter of only a few minutes. However, it fails to detect
the cracks in rare cases. The traditional method of manually measuring the dimensions
of cracks is time consuming and prone to human error. Automating the crack detection
process using drones and computer vision has proven to be beneficial [1]. Deep
learning and image processing solutions have been developed to detect cracks in civil
infrastructures [1]. Yuan et. al. developed a method to compute the length of cracks on
the metal surfaces using deep learning [2]. Deep learning might give better results in
crack detection, but the quantification of cracks fails with deep learning methods in
different settings. Yuan. et. al. uses the pixel-per-metric ratio to calculate the length of
cracks. It works only if the image is taken at an angle of 90 degrees to the crack and
fails when the angle is changed. Furthermore, deep learning methods are
unexplainable when they fail on detection.
Description
Keywords
Civil infrastructure, Crack detection, Tennis courts