GridNet-HD: A High-Resolution Multi-Modal Dataset for LiDAR-Image Fusion on Power Line Infrastructure

TL;DR

GridNet-HD introduces a comprehensive high-resolution multi-modal dataset combining LiDAR and imagery for 3D semantic segmentation of power line infrastructure, enabling improved fusion models and benchmarking.

Contribution

The paper provides the first public dataset with high-density LiDAR and high-resolution imagery for power-line asset segmentation, along with baseline models and evaluation metrics.

Findings

Fusion models outperform unimodal baselines by +5.55 mIoU.

The dataset contains 7,694 images and 2.5 billion points with 11 annotated classes.

High complementarity between geometry and appearance improves segmentation accuracy.

Abstract

This paper presents GridNet-HD, a multi-modal dataset for 3D semantic segmentation of overhead electrical infrastructures, pairing high-density LiDAR with high-resolution oblique imagery. The dataset comprises 7,694 images and 2.5 billion points annotated into 11 classes, with predefined splits and mIoU metrics. Unimodal (LiDAR-only, image-only) and multi-modal fusion baselines are provided. On GridNet-HD, fusion models outperform the best unimodal baseline by +5.55 mIoU, highlighting the complementarity of geometry and appearance. As reviewed in Sec. 2, no public dataset jointly provides high-density LiDAR and high-resolution oblique imagery with 3D semantic labels for power-line assets. Dataset, baselines, and codes are available: https://huggingface.co/collections/heig-vd-geo/gridnet-hd.