Convolutional neural networks for the crack diagnostics in concrete structures

Abstract

EN: Problem statement. Millions of dollars are spent annually in the world on technical diagnostics of buildings and structures. Natural disasters such as floods and earthquakes, along with numerous negative man-made impacts lead to serious damage to building structures. The problem of diagnostics of the buildings and structures became extremely urgent after the aggression of the russian federation in Ukraine, which led to large-scale damage and destruction of industrial projects, housing stock and infrastructure projects such as roads, bridges, tunnels, etc. An important and urgent problem of Civil Engineering is to automate the processes of diagnostics of buildings and structures and to develop new methods for identifying building defects in building structures that would save human resources and reduce the dependence of survey results on subjective factors.