Search papers, labs, and topics across Lattice.
This paper evaluates classical and deep learning-based RGB-NIR image registration techniques for off-road forestry environments, a challenging domain for sensor fusion and autonomy. The NeMAR method showed partial success but suffered from GAN loss instability, hindering geometric consistency. MURF demonstrated promising large-scale feature alignment but struggled with fine details in dense vegetation, highlighting the need for further refinements in multi-scale registration.
Off-road forestry environments expose critical limitations in current RGB-NIR image registration techniques like NeMAR and MURF, demanding further research into robust, multi-scale approaches.
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.
