Federal Ministry of Economy Research Project

SDaC Smart Design and Construction

Pattern detection in a building model

Intelligent planning and realization of buildings through artificial intelligence AI.

Smart Design and Construction Research Project

„Smart Design and Construction“ (SDaC) is a digital platform that brings together the fragmented construction industry and enables collaboration across data and companies.
On the one hand, artificial intelligence (AI) methods make data from individual companies interoperable, and on the other hand, AI applications can use this data to strengthen medium-sized businesses.

Efficiency, safety and sustainability are at the core of SDaC for a responsible and charitable development of AI in the construction industry. SDaC consists of a strong consortium with access to data from a large number of projects and a first sample of a use case.

Research project SDaC Smart Design and Construction

AI-applications for the construction industry

2 Prototypes

Symbol recognition on construction plans

AI symbol recognition
AI symbol recognition

Classification of construction site images

AI image classification
AI image classification

© digitales bauen & Karlsruhe Research Institute KSRI

Current developments

Detail HVAC fixtures

using pattern recognition

Floor plan generation

using spatial relationships in the room book

Digital modeling

using analog plans

In a first case study, individual elements in building plans were examined using machine learning methods to increase the degree of similarity. © SDaC

Solutions
Research project on the integration of AI (artificial intelligence) into the construction industry.

Project Data

Research Period
2020 – 2023

Handover of the funding certificate to the SDaC Consortium at the Federal Ministry of Economics and Energy Berlin © SDaC

Funding Certificate of the Federal Ministry of Economics and Energy Berlin © SDaC

SDaC Partners

Institute for Technology and Management in Construction TMB

Karlsruhe Service Research Institute KSRI

GOLDBECK US Inc.

digitales bauen

SDaC AI application: Symbol recognition

SDaC AI application:
Symbol recognition

AI recognizes symbols on construction plans with an average probability of 90%

Potentials:

  • Quickly create digital models of many existing buildings with analog plans.
  • Simplify quotation creation by automatically counting symbols.

Manual counting

Combination of AI and user control

AI recognition

Time-consuming and expensive manual symbol recognition on analog construction plans

AI symbol recognition computer vision
and user control

AI recognizes symbols with an average probability of 90%
increase to 99% through continuous training of the AI

Manual counting

Combination of AI and user control

AI recognition

Time-consuming and expensive manual symbol recognition on analog construction plans

AI symbol recognition computer vision
and user control

AI recognizes symbols with an average probability of 90%
increase to 99% through continuous training of the AI

Time-consuming and expensive manual symbol recognition on analog construction plans

AI symbol recognition computer vision
and user control

AI recognizes symbols with an average probability of 90%
increase to 99% through continuous training of the AI

SDaC AI application: Image classification

SDaC AI application:
Image classification

Classify construction site images with AI and make them usable

Construction sites are regularly documented with images. These images lack useful meta information such as location or content information. Automated classification simplifies the inspection and documentation of the construction site.
After classification, the construction site images are transferred to Construction Management.

Site images

Combination of AI and user control

AI classification

Large data set of construction site images

Multi-label classification
Semi-supervised learning

AI classifies construction site images with an average probability of 65%
increase to 95% through continuous training of the AI
SDaC application with search and filter function

Site images

Combination of AI and user control

AI classification

Large data set of construction site images

Multi-label classification
Semi-supervised learning

AI classifies construction site images with an average probability of 65%
increase to 99% through continuous training of the AI
SDaC application with search and filter function

Are you interested in or have questions about our AI applications for the construction industry?
Get in touch with our AI expert:

Are you interested in or have questions about our AI applications for the construction industry?

Get in touch with our AI expert:

Georg Müller
georg.mueller(at)digitales-bauen.com
0721 266 756 35

Georg Müller
georg.mueller(at)digitales-bauen.com
0721 266 756 35

digital planning, constructing
and operating

digitales bauen
Augartenstrasse 1
76137 Karlsruhe
Germany
+49 721 266756 10
info@digitales-bauen.de

© digitales bauen GmbH 2024
Imprint
Privacy policy