Karabulut, Navdar
Supervisor:Prof. Gudrun Klinker
Advisor:Rudolph, Linda (@ge29tuw)
Submission Date:[created]


Today most buildings are still designed with two-dimensional CAD software. In addition, the vast majority of building plans are represented in vectorized two-dimensional, such as AutoDesk DWG and DXF files. Although the representation is already vectorized, there is no semantics added to the symbols and elements of the plans. Even worse, connections between objects are often derived by experience instead of providing a symbolic representation of the building systems, such as the fire safety system with interconnected smoke detectors or Heating, ventilation, and air conditioning (HVAC) systems with a connection of pipes. In this master thesis, a rule-based semantical enhancement method is proposed to assist the extraction of objects and systems from two-dimensional CAD’s in the context of vectorized building plans. An extensive review of existing solutions for symbol detection is presented by explaining why some methods are relevant or non-relevant for vector-based input and providing their advantages and disadvantages for two- dimensional CAD’s in building plans. Following a comprehensive examination of existing methods for symbol detection, in this work, a complete framework for symbol detection based on a hybrid approach for vector-based input employing vectorial signatures is presented. The proposed framework is based on the relationship between geometrical primitives that line segments, circles, and arcs. The proposed framework improves the relationship between line segments, between circles, and between line segments and circles based on an existing solution. A novel method is proposed for determining the relationship between geometrical primitives, including between arcs, between line segments and arcs, and between circles and arcs. The proposed method formulates new relation features with different quality factors. Experiments are performed on vector-based inputs. Preliminary results demonstrate the potential of the proposed approach.

Results/Implementation/Project Description


[ PDF (optional) ] 

[ Slides Kickoff/Final (optional)]