SPVSoAP3D: A Second-order Average Pooling Approach to enhance 3D Place Recognition in Horticultural Environments

TL;DR

This paper introduces SPVSoAP3D, a novel 3D place recognition method tailored for horticultural environments, utilizing second-order pooling and descriptor enhancement to improve accuracy in sparse, overlapping LiDAR scans.

Contribution

The work presents a new modeling approach combining voxel-based features with second-order pooling and descriptor enhancement, specifically addressing horticultural environment challenges.

Findings

Second-order pooling outperforms max and first-order methods in horticultural settings.

Descriptor enhancement significantly improves recognition accuracy.

SPVSoAP3D outperforms state-of-the-art models on new and existing datasets.

Abstract

3D LiDAR-based place recognition has been extensively researched in urban environments, yet it remains underexplored in agricultural settings. Unlike urban contexts, horticultural environments, characterized by their permeability to laser beams, result in sparse and overlapping LiDAR scans with suboptimal geometries. This phenomenon leads to intra- and inter-row descriptor ambiguity. In this work, we address this challenge by introducing SPVSoAP3D, a novel modeling approach that combines a voxel-based feature extraction network with an aggregation technique based on a second-order average pooling operator, complemented by a descriptor enhancement stage. Furthermore, we augment the existing HORTO-3DLM dataset by introducing two new sequences derived from horticultural environments. We evaluate the performance of SPVSoAP3D against state-of-the-art (SOTA) models, including OverlapTransformer, PointNetVLAD, and LOGG3D-Net, utilizing a cross-validation protocol on both the newly introduced sequences and the existing HORTO-3DLM dataset. The findings indicate that the average operator is more suitable for horticultural environments compared to the max operator and other first-order pooling techniques. Additionally, the results highlight the improvements brought by the descriptor enhancement stage.