PipeReport has been designed to utilises deep learning techniques in the computer vision field. The advent of artificial intelligence as a disruptive technology in most industries lends itself almost seemlessly to automated reporting of visually captured mediums.The combination of well curated image dataset, professional CCTV coders and accurate artificial intelligence mean the hard work of sitting through hours of CCTV survey footage can be considerably reduced by optimising and automating the initial stages.

Consequently, we have designed a system that provides objective, automated reports that are available for editing before release to the client. Not only is the reporting experience made easier for the user but the client will have access to rich, dynamic reports relatively quickly.

The main objectives of the project:

  • Creation of an automated, objective reporting system – The application takes video files and creates a report using automated deep learning techniques. Pipereport is designed specifically for reporting on the condition of drain and sewer pipes.
  • Reports remain editable for moderation – Automated reports are returned to the user for moderation and editing as a final quality process. Dynamic reports can be viewed via link.
  • Works with any camera system – The system works on the basis that the user uploads video files of drainage surveys. We then take the information provided and turn it into a report.
  • Works on any platform – PipeReport is a web application and just requires a browser. Dynamic reports can  be viewed via link. Exports to national standards for easy interoperability with external applications.
  •  High accuracy – Over 95% accuracy in model accuracy, reduces moderation and checking time.

Example of PipeReport deep learned inference

The objective was to create an automated reporting system using deep learning models built from a comprehensive dataset. Information on condition is inferred via automation including structural, operational condition, defect labelling, distance gathering and header information.

The automated report is then available for checking and adjusting where required. The application is based around screening all surveys and reporting in detail where required. Of course, all automated reports still need a final check before use.

What are the benefits?

  • Increased volumes of surveying on site. The Pipereport system only requires the video file, no need to stop and report on each defect. 
  • No special camera system or software reporting application required. Pipereport will work with any camera system and on any platform.
  • Quick and easy turnaround from survey to report. Reduced workload as initial reporting is outsourced prior to final quality checks.
  • Clients can have access to dynamic reports.
  • Interoperability of reporting. Whether you want a richer report using existing reporting software or want to add your data to a GIS.

Let automation take the strain.