The normal distributions transform: a new approach to laser scan matching

By Peter Biber, W. Straßer. IEEE/RJS International Conference on Intelligent RObots and Systems, 2003

1,550 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Electrical and Electronic Engineering, Geology

View Paper: https://www.semanticscholar.org/paper/48c3e6c50afa0d74882b412ac23c14a27dbc6065

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Electrical and Electronic Engineering

In Geology

Citations in the corpus

Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registration (2016, methodology): “Biber and Straßer [34] developed the Normal Distributions Transform (NDT) method, where Gaussian models were deﬁned for uniform cells in a spatial grid.”
Localization in highly dynamic environments using dual-timescale NDT-MCL (2014, methodology): “It employs NDT [2, 3] to represent both, the map and the measurements.”
3DmFV: Three-Dimensional Point Cloud Classification in Real-Time Using Convolutional Neural Networks (2018, methodology): “The Normal Distribution Transform (NDT) [33] is a piecewise continuous representation which models the distribution of points by a collection of local normal distributions on several overlapping grid cells.”
A 3-D Scan Matching using Improved 3-D Normal Distributions Transform for Mobile Robotic Mapping (2006, methodology): “Peter’s NDT[1] divides euclid space into lattice cells, and converts reference scan points in a cell into normal distribution.”
L3-Net: Towards Learning Based LiDAR Localization for Autonomous Driving (2019, background): “Iterative Closest Point (ICP) [2], G-ICP [30] and Normal Distributions Transform (NDT) [3] are common registration algorithms that can be considered.”
Autonomous driving based on accurate localization using multilayer LiDAR and dead reckoning (2017, methodology): “In particular, NDT scan matching is considered an efficient method in terms of its memory and computational costs [9], [10], [11].”
Autoware on Board: Enabling Autonomous Vehicles with Embedded Systems (2018, methodology): “This is because the computational cost of the NDT algorithm is not dominated by the map size, thus enabling the deployment of high-definition and high-resolution 3D maps at larger scales.”
Robust localization using 3D NDT scan matching with experimentally determined uncertainty and road marker matching (2017, background): “The normal distribution transform (NDT) can cope with slight environmental changes because an environmental map is represented by a set of normal distributions [10], [11].”
The H3D Dataset for Full-Surround 3D Multi-Object Detection and Tracking in Crowded Urban Scenes (2019, methodology): “We registered a sequence of point clouds at 2Hz from the odometry computed using NDT. Bounding boxes and track IDs are annotated on the registered point cloud.”
Evaluation of 3D registration reliability and speed - A comparison of ICP and NDT (2009, methodology): “This has a smoothing effect similar to the approach of Biber and Straßer [3] without the need to compute more PDFs (see Fig.”
Scan registration for autonomous mining vehicles using 3D‐NDT (2007, methodology): “The normal distributions transform (NDT) is a more recent method for registration developed for two-dimensional scan data (Biber and Straßer, 2003).”
FAST-LIO2: Fast Direct LiDAR-Inertial Odometry (2021, neutral): “Other works based on Normal Distribution Transformation (NDT) [34]– [36] also register raw points, but NDT has lower stability compared to ICP and may diverge in some scenes [36].”
An Open Approach to Autonomous Vehicles (2015, methodology): “We use the Normal Distributions Transform (NDT) algorithm to solve this localization problem.(2) To be precise, we use the 3D version of NDT to perform scan matching over 3D point-cloud data and 3D map data.”
The three-dimensional normal-distributions transform : an efficient representation for registration, surface analysis, and loop detection (2009, methodology): “It was first proposed by Biber and Straßer in 2003 [7] as a method for 2D scan registration.”
A flexible and scalable SLAM system with full 3D motion estimation (2011, methodology): “Normal Distribution Transform (NDT) [14] based scan matching aligns scans to a mixture of normal distributions representing preceding scans.”