Horizon3D: Sparse Radar-Camera Fusion for Long-Range 3D Perception in Autonomous Driving

2026-06-29

AI Summary

Existing radar-camera fusion methods struggle to simultaneously capture scene-level context and fine-grained object details for long-range 3D object detection in high-speed autonomous driving scenarios. This work proposes a sparse fusion framework that innovatively integrates Gaussian primitives with sparse bird’s-eye-view (BEV) features, initializing object representations through keypoint-guided proposals and employing an object-centric sparse fusion strategy. Furthermore, a dual-path temporal fusion mechanism—comprising a BEV path and a Gaussian path—is introduced to jointly model scene-level feature accumulation and object-level motion dynamics. Evaluated on the TruckScenes validation set, the proposed method achieves state-of-the-art performance, surpassing the previous best approach by +3.0 NDS and +1.6 mAP while maintaining efficient inference speed.

Abstract

Long-range 3D object detection is critical for safe autonomous driving at highway speeds, yet existing radar-camera fusion methods remain limited at extended ranges. BEV-based methods capture scene-level context but incur rapidly growing computation and often lose fine-grained object detail, while query-based methods are efficient but provide limited scene-level context. Temporal fusion further requires both multi-frame accumulation for sparse distant observations and object-level motion modeling for fast-moving objects. We propose Horizon3D, a sparse radar-camera fusion framework for long-range 3D object detection that combines Gaussian primitives with sparse BEV features. Horizon3D initializes Gaussian primitives at radar- and camera-estimated object keypoints using Keypoint-Guided Gaussian Initialization, refines them through Object-Centric Sparse Fusion, and splats them onto the BEV plane to fuse object-level detail with sparse radar BEV context. It further introduces Dual-Path Temporal Fusion, which aggregates temporal cues through a BEV path for scene-level accumulation and a Gaussian path for object-level motion propagation. Experiments on TruckScenes show that Horizon3D achieves state-of-the-art radar-camera 3D detection performance. On the validation set, it outperforms the previous best method by +3.0 NDS and +1.6 mAP while maintaining competitive inference speed.

Problem

• long-range 3D object detection
• radar-camera fusion
• sparse perception
• temporal fusion
• autonomous driving

Innovation

• sparse fusion
• Gaussian primitives
• BEV representation
• temporal fusion
• radar-camera fusion