Semantic Segmentation of Terrestrial LIDAR Data Using Co-Registered RGB Data

To develop a method for semantic segmentation of 3D point cloud using 2D Convoluted Neural Networks (CNNs).

28 July 2021

Research Team

Erick Sanchez Castillo | David Griffiths | Jan Boehm

Technology Areas

A.I., Machine Learning and Deep Learning

Application Areas

Construction

Abstract

This project proposes a semantic segmentation pipeline for terrestrial laser scanning data. We achieve this by combining co-registered RGB and 3D point cloud information. Semantic segmentation is performed by applying a pre-trained off-the-shelf 2D convolutional neural network over a set of projected images extracted from a panoramic photograph.

Publication: Publisher