The world of wastewater infrastructure has been abuzz over the past year as sewer services contractors, utilities, and consulting engineers have now begun leveraging Artificial Intelligence (AI) computer vision tools in their CCTV pipeline inspection workflows. This AI functionality, also known as AutoCode™, identifies and labels the conditions in a pipe by analyzing each frame in an inspection video. These tools were developed based on hundreds of thousands of “training” images, extracted from tens of millions of feet of pipe data, each manually reviewed and annotated by experienced subject matter experts.
Combine this rigorous training with the nearly infinite scalability of remote computers working together synchronously in a Cloud, the computer vision can automatically analyze tens of thousands of feet of inspection data per hour, 24 hours a day. The outputs are then reviewed by industry-certified technicians who finalize the condition assessment deliverables.
Overall, combining AI computer vision and Cloud technologies enable an approximate doubling of production for CCTV crews in the field, while reducing by approximately two-thirds the amount of time for review in the office.
These technologies are badly needed. As any casual observer of recent news events can attest, our infrastructure is crumbling above, beneath and all around us. Wastewater utilities are now under more strain than ever to extend the remaining useful life of their collection systems, eliminate sewer overflows, and do it with increasingly limited resources. Unfortunately, the process of obtaining the type of reliable and accurate condition assessment data to support sound decision-making has historically been expensive, slow, and fragmented.
As our industry continues to evolve beyond antiquated workflows and data management practices, one of the primary positive impacts of AI and Cloud-based technologies in our industry is to make inspection operations scalable. These technologies allow us to quickly and cost-effectively onboard new employees to immediately become successful CCTV operators.
In the past, inspection operators needed to memorize hundreds of different pipe condition “codes”, as well as the numerous deterioration mechanisms affecting sewer collection systems, all while having a solid understanding of underground construction practices and materials, current and historic. Developing these unique skill sets takes years, and requires lots of “lessons learned,” aka “mistakes”. Once individuals garner this sewer inspection expertise, they are in demand and their employers are faced with retention challenges in today’s strong job market.
The introduction of AutoCode™ enables a simplified workflow, where the CCTV operator no longer manually enters the condition codes while the camera traverses through the pipe, but rather focuses only on recording high quality video. This expedites the operator’s day and enables less experienced technicians to become as productive as their more-seasoned counterparts.
According to Jordan Camacho, VP of Operations at Pipe & Plant Solutions Inc, based in Northern California:
“By implementing the AutoCode workflow in the field, we’ve seen new operators produce more than double the footage they would’ve captured entering data manually. Since the majority of our projects are awarded based on the lowest unit-based price, this is a huge win for our company and the utilities we work for, who are obligated to spend wisely from their budgets.”
SOURCE – SewerAI