Deep learning for 3D point cloud processing in construction: PRISMA-based systematic review (2016–2025)

Caochenyu Zhou

doi:10.1016/j.autcon.2026.107197

Deep learning for 3D point cloud processing in construction: PRISMA-based systematic review (2016–2025)

2026-08-06

0

PRE

AI

C

Caochenyu Zhou *

J

Junshan Liu

S

Salman Azhar

DOI：10.1016/j.autcon.2026.107197

Original

Original request for help

Save

Abstract

En 中文

• Reviews 94 studies on deep learning for construction point-cloud processing. • Links construction goals, processing tasks, and expected deliverables. • Semantic segmentation is the main entry point for engineering-ready outputs. • Compares four model families under construction data constraints. • Identifies barriers to robust and workflow-ready automation.

Keywords:

Systematic literature review

Deep learning

Point cloud

Construction management

As-built BIM

Deep learning for 3D point cloud processing in construction: PRISMA-based systematic review (2016–2025)

Abstract

Journal

Organization

Cited Papers

Citing Papers