Voxel-Based Neighborhood for Spatial Shape Pattern Classification of Lidar Point Clouds with Supervised Learning

Abstract

Improving the effectiveness of spatial shape features classification from 3D lidar data is very relevant because it is largely used as a fundamental step towards higher level scene understanding challenges of autonomous vehicles and terrestrial robots. In this sense, computing neighborhood for points in dense scans becomes a costly process for both training and classification. This paper proposes a new general framework for implementing and comparing different supervised learning classifiers with a simple voxel-based neighborhood computation where points in each non-overlapping voxel in a regular grid are assigned to the same class by considering features within a support region defined by the voxel itself. The contribution provides offline training and online classification procedures as well as five alternative feature vector definitions based on principal component analysis for scatter, tubular and planar shapes. Moreover, the feasibility of this approach is evaluated by implementing a neural network (NN) method previously proposed by the authors as well as three other supervised learning classifiers found in scene processing methods: support vector machines (SVM), Gaussian processes (GP), and Gaussian mixture models (GMM). A comparative performance analysis is presented using real point clouds from both natural and urban environments and two different 3D rangefinders (a tilting Hokuyo UTM-30LX and a Riegl). Classification performance metrics and processing time measurements confirm the benefits of the NN classifier and the feasibility of voxel-based neighborhood.

Field Navigation Using Fuzzy Elevation Maps Built with Local 3D Laser Scans

“…To increase the output rate, the calculation of the FEM and the FRM has been parallelized into two threads inside the Map node by separating the calculation of Equation (8) to that of Equation (9). For further improvement, the Map node can be launched two or more times working on different 3D point clouds.…”

“…In urban areas and indoors, processing of laser information can benefit from the recognition of common geometric features, such as planes [7]. Natural terrain classification can be performed by computing saliency features that capture the local spatial distribution of 3D points [8,9].…”

Generalized Sparse Convolutional Neural Networks for Semantic Segmentation of Point Clouds Derived from Tri-Stereo Satellite Imagery

“…Hence, we compare the model to a representation learning technique, and once to a non-representation learning technique. In doing so, we report the average precision [99], average recall [99], average F1 score [99,100], average accuracy, overall accuracy [101], kappa score [102], Matthew's Correlation Coefficient (MCC) [103,104], and confusion matrices. We use the Minkowski Engine [91], which is built on top of PyTorch [105], for training the GSCNN.…”

A Comparative Study of Object Classification Methods Using 3D Zernike Moment on 3D Point Clouds

“…A variety of machine learning techniques have been applied to classify 3D point clouds [9]. In addition, supervised machine learning is utilized by presenting prelabeled examples to obtain useful predictive models that can be applied to new data [10,11]. Especially in the last few years, neural networks have been the basis of the methods used by advanced computer vision algorithms in many areas such as classification [12], segmentation [13] and target detection [14].…”