Map It Anywhere (MIA): Empowering Bird's Eye View Mapping using Large-scale Public Data

TL;DR

This paper introduces MIA, a scalable data engine leveraging large-scale public map platforms to create extensive datasets for training models that predict Bird's Eye View maps from first-person images, significantly improving zero-shot generalization.

Contribution

The paper presents MIA, a novel data curation framework that enables large-scale, diverse dataset creation from open-source map platforms for BEV map prediction.

Findings

MIA dataset contains 1.2 million FPV-BEV pairs across diverse environments.

Pretraining on MIA data improves zero-shot BEV prediction performance by 35%.

Large-scale public data enhances model generalization for autonomous navigation.

Abstract

Top-down Bird's Eye View (BEV) maps are a popular representation for ground robot navigation due to their richness and flexibility for downstream tasks. While recent methods have shown promise for predicting BEV maps from First-Person View (FPV) images, their generalizability is limited to small regions captured by current autonomous vehicle-based datasets. In this context, we show that a more scalable approach towards generalizable map prediction can be enabled by using two large-scale crowd-sourced mapping platforms, Mapillary for FPV images and OpenStreetMap for BEV semantic maps. We introduce Map It Anywhere (MIA), a data engine that enables seamless curation and modeling of labeled map prediction data from existing open-source map platforms. Using our MIA data engine, we display the ease of automatically collecting a dataset of 1.2 million pairs of FPV images & BEV maps encompassing diverse geographies, landscapes, environmental factors, camera models & capture scenarios. We further train a simple camera model-agnostic model on this data for BEV map prediction. Extensive evaluations using established benchmarks and our dataset show that the data curated by MIA enables effective pretraining for generalizable BEV map prediction, with zero-shot performance far exceeding baselines trained on existing datasets by 35%. Our analysis highlights the promise of using large-scale public maps for developing & testing generalizable BEV perception, paving the way for more robust autonomous navigation. Website: https://mapitanywhere.github.io/