Machine Vision for Object Detection in Architectural Drawings

Year

2026

Industry

Industrial Design and Construction

Client

AI Investlab

Overview

I developed a machine vision system to automatically detect and count objects in architectural drawings. The goal was to reduce manual work for designers and construction teams by automating object recognition in floor plans and technical sketches.

The challenge

Architects and construction teams often need to count elements such as doors, windows, walls. This process is usually done manually and can take a lot of time. It is also easy to make mistakes, especially in large or complex plans.
The challenge was to create a system that could automatically recognize and count these elements from architectural drawings.

My solution

I created a workflow that includes:
  • Annotating architectural drawings to build a labeled dataset
  • Training an object detection model to recognize specific elements
  • Generating output images with color-coded and marked objects
The system detects doors, windows, walls, and other selected elements in floor plans.
In addition to visual output, the system generates structured object counts. These counts can be used for:
  • Quantity estimation
  • Project planning
  • Documentation
  • Cost calculation
The annotated images provide clear visual feedback so users can easily verify the results.

The result

The system automates object counting in architectural drawings and reduces manual work.
Main results:
  • Automatic detection of architectural elements
  • Structured quantity extraction
  • Reduced human error
  • Faster early-stage project analysis
  • Clear visual output for validation
The solution supports more efficient design and construction planning workflows.
System type :

Machine Vision • Object Detection • Design Automation

Role :

Data Annotation • Model Training • Detection Pipeline Implementation

Related cases

Clear architecture and opeerational impact. Discover how intelligent systems improve efficiency, visibility and decision-making across industries.

< Previous

‹ Offline Tennis Player Tracking and Video Annotation

Next >

AI-Assisted Fiber-Reinforced Concrete Floor Design Tool ›

Key features

  • Clearly machine vision, not generic AI
  • Shows full pipeline understanding (annotation → training → inference)
  • Direct productivity impact for AEC industry
  • Visual outputs make it instantly understandable
  • Scales well conceptually (BIM, quantity takeoffs, QA)
Tags :

Computer Vision,Object Detection,Architecture Technology,Construction Automation,Design Analytics

Request a demo