DLDDS: Deep Learning Sewer Defects Detection

¿What is it?

An automatized system that analyses network sewer videos and searches for defects.

Product functionalities

  • Totally unattended Cloud analysis.
  • Results with images, statistics, reports, Excel, etc.
  • Private Access by user for network / site, permissions for operator, manager, etc.
  • Easy to use and totally automatized without any need of configuration by the end user.
  • Possibility of personalization of results so that the system can report directly to end users (logo, messages, etc.)
  • Shareable results via static URLs.
  • Possibility of viewing the video piece that triggered a defect detection.
  • Distance estimate for each of the found defects.
  • Possibility of exporting the results to CSV format.

¿How it works?

With three easy steps the system analyses the videos and provides results:


The user logs into the application;

uploads a new video for analysis;


and visualizes the results!


On a global scale, asset investment is huge and optimization in CAPEX and OPEX is crucial.

In particular, Sewer network assets are quite old and are prone to constant deterioration.

Images of the sewer network are currently obtained by robots or network operators.

These images are analysed manually, adding subjectivity and additional time for report setup.

Our goal is that this analysis should be totally automated, delivering objective results quickly and reliably.


  • Faster video analysis in search of defects.

  • Change in the way which operators work:

    • It is no longer necessary to revise during the filming
    • Video can be acquired at higher speeds

  • Automatization of report generation and results visualization


  • Cost savings in personnel that currently revise videos manually.

  • Objective analysis - a quick way to certify

  • Image of modernity, possible savings for the end user that receives the defect detection service.