Search papers, labs, and topics across Lattice.
This paper addresses the challenge of reliable single-shot high dynamic range (HDR) imaging in extreme illumination conditions by introducing a comprehensive benchmark dataset for evaluating HDR perception in multi-sensor robotic systems. The authors present a novel Depth-guided Multi-view Exposure Bracketing (DMEB) method that utilizes multi-view low-bit-depth cameras to capture varying exposures, which are then fused using depth-guided confidence-aware techniques. Evaluations demonstrate that DMEB significantly enhances HDR imaging performance, establishing it as a strong reference for future research in multi-camera and depth sensor configurations.
DMEB achieves robust HDR perception by effectively fusing exposures from multiple low-bit-depth cameras, setting a new standard for multi-sensor robotic vision.
Achieving reliable single-shot high dynamic range (HDR) imaging under extreme illumination conditions remains a long-standing challenge, yet no comprehensive benchmark exist for evaluating HDR perception in multi-sensor robotic systems. To fill this gap, we introduce a large-scale dataset collected via a custom robotic vision platform and an iPhone 13 Pro: 121 real-world scenes spanning modest and ultra-high dynamic range conditions, alongside 20 synthetic video sequences from the CARLA simulator. As a reference pipeline for this dataset, we propose Depth-guided Multi-view Exposure Bracketing (DMEB), a single-shot HDR method that distributes drastically different exposures across multi-view low-bit-depth cameras and fuses them via depth-guided confidence-aware fusion. Evaluations on our dataset show that DMEB establishes a strong reference point and highlight the promise of this sensor configuration for robust HDR perception in diverse multi-camera and depth sensor system.
