Abstract

Classifying point cloud of urban landscapes plays essential roles in many urban applications. However, automating such a task is challenging due to irregular point distribution and complex urban scenes. Incorporating contextual information is crucial in improving classification accuracy of point clouds. In this article, we propose a hierarchical approach for point cloud classification with 3-D contextual features, which comprises three steps:segment-based classification with primitive features and a random forest classifier; extracting novel 3-D contextual features from the initial labels considering spatial relationships between neighboring segments and semantic dependencies; and refining classification with a combination of primitive features and spatial contextual features, and a hierarchical multilayer perceptron classifier that considers primitive features and spatial contextual features at different levels. The proposed method was tested on two point cloud datasets:the National University of Singapore (NUS) dataset and the Vaihingen benchmark dataset of the International Society of Photogrammetry and Remote Sensing. The evaluation results showed that the proposed method achieved an overall accuracy of 92.51% and 82.34% for the NUS dataset and Vaihingen dataset, respectively. The feature importance evaluation showed that 3-D spatial contextual features contributed useful information for discriminating different classes, such as roof, facade, grassland, tree, and ground. Quantitative comparisons further showed that the proposed method is more advantageous, especially in the detection of class roof and facade.

Highlights

  • T HE drive toward smart cities around the world has necessitated the development of 3-D spatial data infrastructures given that they provide precise descriptions of the man-made structures and representations of natural resources in 3-D, and quantifiable pieces of evidence into urban dynamics when armed with Internet-of-Things and social media feeds

  • Two experiments were performed with a terrestrial laser scanning based National University of Singapore (NUS) dataset and an aerial laser scanning based Vaihingen dataset provided by International Society of Photogrammetry and Remote Sensing (ISPRS)

  • The quantitative evaluation showed that the additional 3-D spatial contextual features improve the classification accuracy significantly

Read more

Summary

Introduction

T HE drive toward smart cities around the world has necessitated the development of 3-D spatial data infrastructures given that they provide precise descriptions of the man-made structures and representations of natural resources in 3-D, and quantifiable pieces of evidence into urban dynamics when armed with Internet-of-Things and social media feeds. Most of these applications require point cloud classification as a basic LiDAR processing step, which is to assign each point a semantic label such as ground and grassland [7]. The automation of point cloud classification in urban areas is challenging because of the complexity of urban scenes and a high level of heterogeneity

Objectives
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call