Context-Aware Mapping of 2D Drawing Annotations to 3D CAD Features Using LLM-Assisted Reasoning for Manufacturing Automation

Oct 29, 2026

Upload by Fans

Manufacturing automation in process planning, inspection planning, and digital-thread integration requires a unified specification linking geometric features in 3D CAD to manufacturing intent expressed in 2D engineering drawings. This work proposes a deterministic-first, context-aware framework that enriches drawing callouts semantically, then scores candidates using interpretable metrics covering type compatibility, tolerance-aware agreement, and conservative context consistency. When deterministic scoring cannot resolve ambiguity, the method escalates to multimodal and constrained LLM reasoning, followed by a single human-in-the-loop review. Experiments on 20 real CAD-drawing pairs report mean precision 83.67%, recall 90.46%, and F1 86.29%, with ablation confirming each component’s contribution.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

Problem motivation and challenges

Proposed context-aware mapping framework

Contributions and evaluation overview

Context-Aware Mapping of 2D Drawing Annotations to 3D CAD Features Using LLM-Assisted Reasoning for Manufacturing Automation arXiv :2602 . 18296v2 [ cs .CE] 24 Feb 2026 Muhammad Tayyab Khana, c* , Lequn Chenb , Wenhe Fenga , Seung Ki Moonc a Singapore Institute of Manufacturing Technology (SIMTech), A*STAR, Singapore 636732, Republic of Singapore b Advanced Remanufacturing and Technology Centre (ARTC), A*STAR, Singapore 637143, Republic of Singapore c School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore 639798 *[Email: KHAN0022@e.ntu.edu.sg](Email: KHAN0022@e.ntu.edu.sg) Abstract—Manufacturing automation in process planning, inspection planning, and digital-thread integration depends on a unified specification that binds the geometric features of a 3D CAD model to the geometric dimensioning and tolerancing (GD&T) callouts, datum definitions, and surface requirements carried by the corresponding 2D engineering drawing. Although Model-Based Definition (MBD) allows such specifications to be embedded directly in 3D models, 2D drawings remain the primary carrier of manufacturing intent in automotive, aerospace, shipbuilding, and heavy-machinery industries. Correctly linking drawing annotations to the corresponding 3D features is difficult because of contextual ambiguity, repeated feature patterns, and the need for transparent and traceable decisions. This paper presents a deterministic-first, context-aware framework that maps 2D drawing entities to 3D CAD features to produce a unified manufacturing specification. Drawing callouts are first semantically enriched and then scored against candidate features using an interpretable metric that combines type compatibility, tolerance-aware dimensional agreement, and conservative context consistency, along with engineering-domain heuristics. When deterministic scoring cannot resolve an ambiguity, the system escalates to multimodal and constrained large-language-model reasoning, followed by a single human-in-the-loop (HITL) review step. Experiments on 20 real CAD–drawing pairs achieve a mean precision of 83.67%, recall of 90.46%, and F1 score of 86.29% . Anablation study shows that each pipeline component contributes to overall accuracy, with the full system outperforming all reduced variants. By prioritizing deterministic rules, clear decision tracking, and retaining unresolved cases for human review, the framework provides a practical foundation for downstream manufacturing automation in real-world industrial environments. Index Terms—Manufacturing context mapping, LLM-assisted reasoning, 2D–3D CAD correspondence, human-in-the-loop, digital thread integration. I. INTRODUCTION Modern manufacturing automation relies on the accurate and machine-actionable interpretation of design intent [1] . While 3D CAD models provide precise geometric descriptionsand enable automatic feature recognition (AFR), they do not fully capture manufacturing requirements [2] . Critical information, such as geometric dimensioning and tolerancing (GD&T) constraints, datum references, surface finish requirements, and thread specifications are typically specified in 2D drawings [3],[4] . Downstream tasks, including process planning, inspection planning, tolerance analysis, and digital-thread integration, require these heterogeneous representations to be unified into a consistent manufacturing specification. Notably, despite advances in Model-Based Definition (MBD) [5], 2D drawings remain the authoritative source of symbolic manufacturing semantics in many industrial workflows [6], [7] . Aligning 2D drawing intent with 3D geometry remains challenging because drawing annotations are inherently context dependent. A numeric value such as “10” may denote a hole diameter, a slot width, a fillet radius, or a depth depending on drafting conventions, leader placement, and view association. GD&T symbols further complicate interpretation because visually similar annotat

FAQs

Why is mapping 2D drawing annotations to 3D CAD features difficult in manufacturing automation?

Drawing annotations are context dependent, and similar numeric or GD&T symbols can denote different constraints depending on conventions, leader placement, view association, symbol modifiers, and datum references. Misinterpretation can propagate errors into manufacturing and inspection planning.

How does the proposed framework resolve ambiguity between drawing callouts and 3D CAD features?

It first semantically enriches drawing callouts and applies deterministic scoring with an interpretable metric that combines type compatibility, tolerance-aware dimensional agreement, and conservative context consistency. If deterministic scoring cannot decide, it escalates to multimodal and constrained LLM reasoning and then performs a single human-in-the-loop review.

What performance results does the paper report, and what do the ablation studies show?

On 20 real CAD–drawing pairs, the method achieves mean precision 83.67%, recall 90.46%, and F1 score 86.29%. The ablation study indicates each pipeline component contributes to overall accuracy, and the full system outperforms reduced variants.

Oct 29, 2026

Upload by Fans