AI-powered decision support for the re-evaluation of pipe inspections

An infrastructure operator operates an extensive drainage system on a landfill, the pipelines of which are checked annually for leaks. A KI-supported camera solution is already used for this purpose, which creates recordings of the pipe interiors and automatically categorizes the captured images. For those cases where the KI predicts potential future damage or abnormalities, a professional re-evaluation is currently carried out by an external engineering firm. This makes the final assessment of the findings. In order to ensure the high quality and traceability of this established evaluation process in the long term and to support knowledge transfer, the HPA is looking for a digital solution that meaningfully complements the existing process. Solution approaches already implemented: - Annual camera inspection of the pipelines - Analysis of image data by an engineering firm - Classification of potential damage by the service provider for camera inspections - Partial use of KI-based pattern recognition in the initial assessment - Professional re-evaluation of conspicuous or prognostically relevant findings by an external engineering firm - Independent review, damage classification and documentation according to the 4-eyes principle

Possible solution approaches: - Assistance system to support specialists in reassessment - AI-supported provision of relevant information and previous comparison cases - Comparison of previous cases with current findings, similar to a historical version or comparison system - Improvement of the traceability of AI pre-assessments - Structured preparation of decision-making bases - Support for the documentation of expert assessments - Promotion of consistent assessments across different processors - Innovative approaches from the fields of AI, Explainable AI, Computer Vision, Knowledge Management, Semantic Search, or Digital Assistance Systems Requirements for the solution - The solution must support specialists in the reassessment of pipe inspections. - The assessment process should be made more efficient, transparent, and traceable. - Relevant information and previous comparison cases should be provided and made comparable. - AI can be used to provide relevant information and previous comparison cases. - The solution should improve the traceability of the AI pre-assessment. - Decision-making bases should be structured. - The documentation of the expert assessment should be supported. - The consistency of assessments across different processors should be promoted. - The final expert assessment must remain with the human. - The solution must support, not replace, the expertise of the inspectors. - Decisions of the inspectors should be prepared and documented as best as possible. - The technological implementation is open. - Approaches from the fields of AI, Explainable AI, Computer Vision, Knowledge Management, Semantic Search, or Digital Assistance Systems are welcome.