B&N has developed an AI software platform called PipeAId(TM) to support utilities with processing condition assessment CCTV videos related to sewer pipes by providing CCTV defect codes and defect mapping. PipeAId is a Computer Vision and Machine Learning AI system that mimics the process that people use when observing and coding pipe inspection videos. Think of PipeAId as a person and the Computer Vision component as the eyes that can look at the video and extract data such as if taps, cracks, roots, etc. are present. The Machine Learning component is the brain that gets smarter and more accurate as more data is processed. PipeAId has a supervised learning environment that allows us to add images to the AI model training datasets effectively increasing the accuracy of the coding output. To get a full look at your pipelines, PipeAId utilizes GIS to geospatially locate each defect/ observation. The GIS database will show you exactly where each root, tap, break, sag, etc. is located and provide more detailed information for field staff, contractors, and decision makers when deciding whether to maintain/rehab/repair/replace. PipeAId can assign probability ratings to each code as an additional QA/QC measure.
GIS, Public Engineering