Computer Vision Support Tool for Assessing Concrete Hydro-Dams Surface Deterioration

Abstract

Hydro-dams failure risks should be continuously assessed since they are critical for
environment/society. A key component of the dam monitoring regulations is visual inspection,
which should allow objective measurements of their structural condition evolution. Towards this
goal, computer tools should be developed, using artificial vision. For concrete hydro-dams, an
important issue is the examination of the downstream surface deterioration. This paper proposes a
computer vision tool to assist experts in generating visual surveillance reports in respect to the
quantitative and linguistic assessment of the deterioration and erosion of concrete. Compliant with
the human observation, the results prove the system's functionality.