Optimal DLT-based Solutions for the Perspective-n-Point

TL;DR

This paper introduces an improved, non-iterative DLT-based algorithm for the PnP problem that outperforms existing methods in accuracy and speed, approaching the true optimal solution with minimal computational overhead.

Contribution

The paper presents a modified normalized DLT algorithm with analytical measurement weighting, significantly enhancing performance and efficiency over traditional PnP solutions.

Findings

Outperforms popular PnP methods like EPnP, CPnP, RPnP, and OPnP in accuracy and runtime.

Approaches the optimal solution obtained by Gauss-Newton optimization.

Open source implementation provided for broader use and validation.

Abstract

We propose a modified normalized direct linear transform (DLT) algorithm for solving the perspective-n-point (PnP) problem with much better behavior than the conventional DLT. The modification consists of analytically weighting the different measurements in the linear system with a negligible increase in computational load. Our approach exhibits clear improvements -- in both performance and runtime -- when compared to popular methods such as EPnP, CPnP, RPnP, and OPnP. Our new non-iterative solution approaches that of the true optimal found via Gauss-Newton optimization, but at a fraction of the computational cost. Our optimal DLT (oDLT) implementation, as well as the experiments, are released in open source.