AECOM’s PipeInsights™ improves the accuracy, speed, and efficiency of sewer inspection programs using Artificial Intelligence (AI), including Machine Learning (ML) and Computer Vision (CV). This made by AECOM cloud-based solution assists in the detection, localization, validation, and diagnosis of pipe problems, and geo-locates the area for inspection through a built-in GIS interface.
Two of the key challenges to sewer inspection programs are finding a cost-effective and accurate way to produce condition assessments. Reviewing and validating inspection data is a critical step toward selecting the right restoration technique. It is often a time-consuming and labour-intensive effort that requires multiple focus-driven personnel to review hours of data to identify and classify issues and defects, validate them with their team, and then geo-locate the problem areas. Generally, this manual process is prone to errors due to subjectivity and is a very resource intensive activity requiring extensive costs and lengthy schedules.
PipeInsights™ was developed as a cloud-based tool that performs fully automated and accurate defect recognition and is designed to utilize non-working hours to process sewer inspections through the use of a suite of AI algorithms.
The movement from a highly manual inspection workflow with multiple validations to the use of Artificial Intelligence for inspection review is beginning to take hold in the industry. However, there is still not a standard that can validate AI accuracy. AECOM is focused on improving methods and workflow to validate the results of AI solutions that can be used in the future as a standard to validate and certify similar solutions developed for the sewer inspection industry.
Central to AECOM’s solution is the use of Machine Learning Algorithms to quickly analyze vast quantities of inspection footage, identify areas of concern and then classify them. The system is trained on hundreds of thousands of defects and continues to learn and improve as new inspections are added. PipeInsights™ ensures CCTV inspections are of the highest quality, enabling the development of accurate pipe maintenance or rehabilitation plans and strategies to manage sewer systems with confidence.
Training the AI to recognize and classify defects meant an initial classification effort to create a standard of defects for ease of identify. The results of those efforts mean the Algorithms can now sort and identify defects into classification families based on visual cues. Ultimately, the inspection validation is cut in half through the use of AI – saving time and resources that can be re-focused on other project tasks.
PipeInsights™ also uses geospatial data exchange standards that allow organizations to leverage their enterprise GIS data. Connected to a GIS interface users can quickly locate inspections on site for maintenance teams to diagnose problems.