ParkGaussian: Surround-view 3D Gaussian Splatting for Autonomous Parking

2026-01-04

arXiv.org

AI Summary

Existing autonomous parking systems lack high-quality 3D reconstruction tailored to parking scenarios, hindering robust spatial geometry perception and downstream parking spot detection. To address this gap, this work introduces ParkRecon3D, the first 3D reconstruction benchmark specifically designed for automated parking, and proposes ParkGaussian—a novel framework that adapts 3D Gaussian Splatting to surround-view fisheye images for the first time. ParkGaussian incorporates a parking-spot-aware guidance mechanism to enhance reconstruction fidelity in critical regions. Evaluated on ParkRecon3D, the proposed method achieves state-of-the-art reconstruction accuracy and significantly improves perceptual consistency between the reconstructed scene and parking spot detection tasks.

Technology Category

• Computer Vision: 3D Computer Vision
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Parking is a critical task for autonomous driving systems (ADS), with unique challenges in crowded parking slots and GPS-denied environments. However, existing works focus on 2D parking slot perception, mapping, and localization, 3D reconstruction remains underexplored, which is crucial for capturing complex spatial geometry in parking scenarios. Naively improving the visual quality of reconstructed parking scenes does not directly benefit autonomous parking, as the key entry point for parking is the slots perception module. To address these limitations, we curate the first benchmark named ParkRecon3D, specifically designed for parking scene reconstruction. It includes sensor data from four surround-view fisheye cameras with calibrated extrinsics and dense parking slot annotations. We then propose ParkGaussian, the first framework that integrates 3D Gaussian Splatting (3DGS) for parking scene reconstruction. To further improve the alignment between reconstruction and downstream parking slot detection, we introduce a slot-aware reconstruction strategy that leverages existing parking perception methods to enhance the synthesis quality of slot regions. Experiments on ParkRecon3D demonstrate that ParkGaussian achieves state-of-the-art reconstruction quality and better preserves perception consistency for downstream tasks. The code and dataset will be released at: https://github.com/wm-research/ParkGaussian

Problem

• autonomous parking
• 3D reconstruction
• parking slot perception
• surround-view
• GPS-denied environments

Innovation

• 3D Gaussian Splatting
• autonomous parking
• slot-aware reconstruction
• surround-view reconstruction
• ParkRecon3D