Preliminary analysis of RGB-NIR Image Registration techniques for off-road forestry environments

TL;DR

This paper evaluates classical and deep learning RGB-NIR image registration methods for off-road forestry environments, highlighting their strengths and limitations in handling complex natural scenes.

Contribution

It provides a preliminary comparison of traditional and DL-based registration techniques specifically tailored for off-road forestry imagery, identifying key challenges.

Findings

NeMAR shows partial success but GAN loss instability affects geometric consistency.

MURF aligns large-scale features well but struggles with fine details in dense vegetation.

Further refinements are needed for robust multi-scale registration in forestry applications.

Abstract

RGB-NIR image registration plays an important role in sensor-fusion, image enhancement and off-road autonomy. In this work, we evaluate both classical and Deep Learning (DL) based image registration techniques to access their suitability for off-road forestry applications. NeMAR, trained under 6 different configurations, demonstrates partial success however, its GAN loss instability suggests challenges in preserving geometric consistency. MURF, when tested on off-road forestry data shows promising large scale feature alignment during shared information extraction but struggles with fine details in dense vegetation. Even though this is just a preliminary evaluation, our study necessitates further refinements for robust, multi-scale registration for off-road forest applications.