This project focused on applying machine vision and object detection to architectural and construction drawings in order to automate object counting and classification tasks commonly performed manually by designers and construction teams. The system was trained to recognize and classify key architectural elements such as doors, windows, and walls from floor plans and technical sketches.

The workflow involved annotating architectural drawings to create a labeled dataset, training a detection model to identify object types, and generating output images with color-coded and marked elements. In addition to visual verification, the system produced structured counts of detected objects, enabling automatic quantity extraction for downstream use in planning, estimation, and documentation.

By automating object recognition and counting, the solution reduces repetitive manual work, minimizes human error, and accelerates early-stage design and construction workflows. The resulting annotated images provide clear, interpretable feedback for users while supporting integration into broader automated pipelines for architectural analysis.