Crack detection in concrete pipes using deep learning assisted computer vision

Start year: 2025

Summary: The accuracy of existing CCTV technology is not suitable for concrete pipes crack detection in the field. CCTV camera manufacturers have advised that measurement of crack width below 1mm is not possible. However, AS/NZS4058 (Precast concrete pipes) specifies acceptable crack widths ranging from 0.1 to 0.25 mm for new pipes. Hence CCTV reports overstate crack widths resulting in costly and unnecessary lining, replacement or repair of installed pipelines. In addition, the crack width limitation in AS/NZS4058 applies to newly manufactured concrete pipes tested under standard loads. However, for in-service pipes, crack widths significantly greater than 0.25 mm are not cause for alarm. The product is still within its working load range, able to sustain load and the corrosion of reinforcement is such that the durability is not a concern. This is because, in the field, the time-dependent autogenous healing of concrete contributes to reduce the initial crack width significantly, preventing reinforcement corrosion. This project will develop a deep learning assisted computer vision as an alternative to current CCTV monitoring of in-service pipes. The technology will be able to achieve 0.1 mm resolution in crack width estimation. The development and commercialization of equipment, process and inspection guidelines related to this new technology would provide the asset owner with appropriate data to make improved informed assessments and reduce unnecessary costs in construction. This project will also investigate the corrosion of reinforcement of cracked concrete specimens with different crack widths, exposed to diverse environmental conditions. The self-healing of the cracks will be monitored together with the reinforcement corrosion. This investigation will determine suitable values of acceptable crack width depending on the type of exposure to be implemented in AS/NZS4058 for installed concrete pipes, which are currently missing.

Acknowledgement of Country

UTS acknowledges the Gadigal people of the Eora Nation, the Boorooberongal people of the Dharug Nation, the Bidiagal people and the Gamaygal people, upon whose ancestral lands our university stands. We would also like to pay respect to the Elders both past and present, acknowledging them as the traditional custodians of knowledge for these lands.