Automatic Recognition and Extraction of Component Information From Construction Drawings Based on Improved Hough Transform

Generating three-dimensional reinforcement models from construction drawings is resource-intensive and requires extensive domain expertise. While building information modeling has advanced automation for architectural elements, reinforcement drawings remain largely overlooked despite their critical role in construction. This study presents an automated framework utilizing an improved Hough transform algorithm to detect and extract reinforcement information directly from construction drawings. The enhancement integrates spatial positioning and inter-element distance features into the classical algorithm, significantly improving segmentation precision and enabling accurate identification of rebar lines at structurally critical locations. A standards-driven annotation matching mechanism further enhances interpretability by leveraging construction norms to decode textual labels, symbols, and graphical cues. Experimental evaluations across diverse datasets consistently achieve accuracy rates exceeding 96%, demonstrating the method's robustness and practical applicability.