Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.

Crack detection flood defences

Detecting cracks in embankments using AI

Last change on 30th of October 2024, at 7:41 (CET) | Publication Standard 1.0
Publication category
Other algorithms
Impact assessment
Field not filled in.
Status
In development

General information

Theme

Space and Infrastructure

Begin date

11-2023

Contact information

https://www.hhnk.nl/form/contactformulier/

Link to publication website

https://www.hhnk.nl

Responsible use

Goal and impact

  • Inspecting the condition of our flood defences more efficiently and cost-effectively.
  • The consequences are minimal for residents, citizens and businesses (ingelanden).

Considerations

Advantages:

  • Reduced inspection burden
  • More targeted deployment of personnel
  • Uniform working method
  • More detailed inspection

Disadvantages:

  • Technical limitations, weather-dependent
  • Dam with trees more difficult to inspect
  • Editable process (no real-time editing of drone images yet)

Human intervention

  • This digital tool remains parallel to the current work process.
  • The algorithm is not autonomous, is a tool for administrators.
  • Training data is carefully reviewed by in-house experts.

Risk management

  • Situations labelled as a crack are always checked afterwards by an experienced inspector (fault positive labelled).
  • All flood defences are periodically physically checked by an experienced inspector to prevent falsely negatively labelled cracks.
  • Based on new insights, the algorithm is continuously improved and monitored.

Legal basis

The Water Act sets safety standards for flood defences, both primary and regional.

Operations

Data

Drone imagery, water mapping

Technical design

The area to be scanned is first mapped using a drone. The drone takes photos, which are then merged into an ortho-mosaic. This orthomozaic is divided into smaller segments, each of which is analysed by a segmentation algorithm to detect cracks.

To monitor the growth, size, length and width of cracks, each pixel is recorded whether it is part of a crack. These pixels can be related to coordinates making it possible to calculate all these monitoring values. Detecting the pixels of cracks is done specifically using a U-net segmentation algorithm. This algorithm is fed with images where the cracks are labelled by polygons drawn around the cracks.

External provider

Internally developed

Similar algorithm descriptions

  • This algorithm helps Customs to select goods for inspection based on risk. It uses declaration data from companies and considers whether or not there are risks of inaccuracies in the declarations.

    Last change on 28th of February 2024, at 12:29 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • Inspection of provincial roads using LiDAR or laser technology. Using these techniques, the road section can be scanned in great detail and all unevenness highlighted. From this, the algorithm selects damage images on the basis of which the province will see what maintenance is needed here.

    Last change on 5th of February 2025, at 9:11 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • RISM uses a LiDAR camera which scans a road and its immediate surroundings. Based on the resulting 3D models, obstacles in the obstacle-free zone can be identified and removed where necessary.

    Last change on 5th of February 2025, at 9:10 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • Detection loops in the road register when there is traffic in front of a traffic light. Green times can be adjusted accordingly.

    Last change on 27th of June 2024, at 13:06 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use