A Geometric Observer for Scene Reconstruction Using Plenoptic Cameras

TL;DR

This paper introduces a gradient-descent based observer for scene depth reconstruction using plenoptic cameras, leveraging motion to improve depth estimation in low-texture scenes, with proven stability and simulation results.

Contribution

It presents a novel observer method that utilizes motion and gradient descent for depth map estimation from light-field data, with stability analysis and simulation validation.

Findings

Observer converges to accurate depth maps in simulations

Motion enhances depth estimation in low-texture scenes

Stability of the observer error is rigorously proven

Abstract

This paper proposes an observer for generating depth maps of a scene from a sequence of measurements acquired by a two-plane light-field (plenoptic) camera. The observer is based on a gradient-descent methodology. The use of motion allows for estimation of depth maps where the scene contains insufficient texture for static estimation methods to work. A rigourous analysis of stability of the observer error is provided, and the observer is tested in simulation, demonstrating convergence behaviour.