Hochschule Burgenland

Academic Affiliation

Hochschule Burgenland – University of Applied Sciences, Austria.

The work is conducted in an academic research context, with a strong focus on applied artificial intelligence, computer vision, and engineering systems.

Mission

PannOnonia OpticAI is a research-driven initiative focused on the digitalization and semantic understanding of technical engineering drawings, with a strong emphasis on:

• P&ID (Piping & Instrumentation Diagrams)
• Process flow diagrams
• Building services engineering drawings
• Strangschemata and technical schematics

Our long-term objective is to transform static technical documents into machine-interpretable, structured representations that can serve as a foundation for intelligent agents in building and industrial environments.

---

What We Are Working On

We develop end-to-end AI pipelines that combine:

• Object detection (e.g. symbols, equipment, instruments)
• Image segmentation (lines, regions, structural elements)
• Text recognition and interpretation
• Graph reconstruction of technical systems

Our approach integrates state-of-the-art computer vision techniques with domain-specific engineering knowledge.

---

Core Research Areas

• Digitalization of technical drawings
  From scanned PDFs and images to structured, analyzable data

• P&ID understanding
  Detection of symbols, valves, instruments, signal lines, and control logic

• Hybrid AI approaches
  Combining classical computer vision, deep learning, and rule-based methods

• Vision Transformers & modern architectures
  Exploration of CNN–Transformer hybrids and attention-based models for technical diagrams

• Dataset creation & curation
  Development of high-quality, domain-specific datasets aligned with ISO and ISA standards

• Model development & evaluation
  Training and benchmarking of detection and segmentation models for engineering diagrams

---

Why This Matters

Technical drawings are a critical but underutilized source of knowledge in engineering, construction, and industrial operations.
By enabling machines to understand these documents, we aim to support:

• Intelligent building agents
• Automated engineering analysis
• Digital twins
• Smarter maintenance and operation workflows

---

Outlook

Our vision is to enable AI agents that can:

• Understand complex technical systems
• Reason over engineering constraints
• Support decision-making in building technology and industrial processes

---

Contact & Collaboration

We are an open, research-oriented group and welcome exchange with:

• Academic partners
• Research groups
• Industrial stakeholders (early-stage, non-commercial)

---

This Space serves as an organizational overview and research status page.
Detailed models, datasets, and experimental results are maintained in dedicated repositories.