Fast 3D Extended Target Tracking using NURBS Surfaces

TL;DR

This paper introduces fast, novel methods for jointly estimating the 3D shape and motion of extended targets using NURBS surfaces from sparse LiDAR data, enabling real-time processing and accurate shape reconstruction.

Contribution

The paper presents new algorithms that efficiently estimate 3D target shapes and dynamics using NURBS surfaces, with two approaches focusing on surface scaling and weights, achieving fast processing times.

Findings

Effective shape and motion estimation from sparse LiDAR data.

Real-world evaluations demonstrate improved speed and accuracy.

Capable of reconstructing detailed 3D shapes of targets.

Abstract

This paper proposes fast and novel methods to jointly estimate the target's unknown 3D shape and dynamics. Measurements are noisy and sparsely distributed 3D points from a light detection and ranging (LiDAR) sensor. The methods utilize non-uniform rational B-splines (NURBS) surfaces to approximate the target's shape. One method estimates Cartesian scaling parameters of a NURBS surface, whereas the second method estimates the corresponding NURBS weights, too. Major advantages are the capability of estimating a fully 3D shape as well as the fast processing time. Real-world evaluations with a static and dynamic vehicle show promising results compared to state-of-the-art 3D extended target tracking algorithms.