Estimating Rigid Transformation Between Two Range Maps Using Expectation Maximization Algorithm

TL;DR

This paper presents a fast EM algorithm for estimating rigid transformations between two LiDAR point sets, crucial for target tracking, with linear complexity in the number of scan points.

Contribution

The paper introduces a computationally efficient EM-based method for rigid transformation estimation between point clouds, optimized for LiDAR data.

Findings

Achieves linear complexity O(N) for transformation estimation

Effective for real-time LiDAR-based target tracking

Provides accurate alignment of point sets in practical scenarios

Abstract

We address the problem of estimating a rigid transformation between two point sets, which is a key module for target tracking system using Light Detection And Ranging (LiDAR). A fast implementation of Expectation-maximization (EM) algorithm is presented whose complexity is O(N) with $N$ the number of scan points.