Step On The Box

Asset Recognition, one image can say more than a thousand words (Artificial Intelligence & Data Science Hackathon)

We are looking for:
HBO & WO students in all fields of study.

Status
Running
Spots
8 available
Registration deadline
30 Oct 2025
Sprintday
6 Nov 2025
Location
Kattenburgerstraat 7, 1018 JA Amsterdam, Netherlands, Onsite
1. Description of the challenge

Develop a working proof-of-concept that is capable of:

  1. Recognizing symbols, components, and connections in drawings and/or photos.
  2. Unambiguously identifying these visual elements and linking them to relevant data in a dataset/database.
  3. Enabling automated actions based on this, such as:
  • Asset inventories
  • Validation of drawings
  • Harmonization and standardization of drawings

Requirements for the solution

  • Recognition rate of at least 70% reliability on a representative dataset.
  • Results must be unambiguous and reproducible: fewer reliable matches are preferable, above matches that requires additional validation.
  • The solution should be scalable towards: an autonomous component that can act independently within a workflow (e.g., signaling missing assets or deviations).

Why is this challenge important for our organisation?
Unlocking Efficiency and Insight with AI-Powered Image Recognition at EQUANS
As a leading technical service provider, EQUANS manages and maintains a vast and diverse portfolio of customer assets. Implementing AI-driven image recognition transforms how we can interact with these assets—turning photos, drawings, and symbols into actionable data. By automatically identifying equipment, infrastructure elements, and schematic symbols, we can enrich our master data contextually, giving visual elements meaning and traceability.

This capability reduces manual labor and site visits, enabling remote asset verification and condition monitoring. It enhances operational efficiency by accelerating asset registration, maintenance planning, and compliance checks.

From a CSR perspective, image recognition supports safer, more sustainable operations—minimizing unnecessary travel, improving worker safety, and ensuring accurate documentation for environmental and regulatory reporting.

Ultimately, this technology empowers EQUANS to scale smarter, respond faster, and manage assets with greater precision—delivering value across labor optimization, cost control, data quality, and customer satisfaction.

2. Expected outcome

EQUANS maintains and develops a vast range of technical assets across customer sites. This PoC challenges students to demonstrate how AI-powered image recognition can extract meaningful data from photos and/or technical drawings—without focusing on the image ingestion pipeline itself. 

The goal is to build a working prototype that can detect and classify visual elements (e.g. assets in site photos or symbols in schematics) and link them to relevant data contexts. This could mean identifying a pump in a photo and connecting it to its maintenance record,or recognizing a symbol in a drawing and mapping it to its function or asset type.

Students are free to choose their tools and approach. Success is defined by demonstrating the core value: turning visual input into structured, contextualized data that supports smarter asset management.

About Equans

Worldwide, EQUANS is the market leader in technical services, with 5.500 employees in 29 locations in the Netherlands. With innovative technical, digital and sustainable solutions, EQUANS helps clients within the utility, industry and government sectors. Constant improvement is paramount. In this way we help our customers to keep up with the latest developments. And we assist you in the three transitions of our time: in the field of energy, digital and industrial.