Learning to Suppress SPAD-based LiDAR Flare
2026-07-03 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors study a type of LiDAR sensor that uses Single-Photon Avalanche Diodes (SPAD) for detecting distances based on light. They focus on a problem called flare, where too many returned photons cause wrong depth readings and distorted 3D images. Instead of relying on traditional hardware fixes, the authors treat flare removal like a picture-labeling task and use machine learning to recognize flare patterns in the sensor data. They propose a new method (PILF) that looks at different light echoes separately and combines this information to better spot flare. Their tests show this method works better than existing ones in real-world settings.
Single-Photon Avalanche DiodeLiDARflarepile-up effectdepth estimationsemantic segmentationmulti-echomachine learningmIoU
Authors
Xuanya Zhu, Linghao Shen
Abstract
Single-Photon Avalanche Diode (SPAD)-based Light Detection and Ranging (LiDAR) is emerging for autonomous vehicles due to its high sensitivity and precise depth sensing capabilities. However, flare caused by excessive photon returns or pile-up effects can lead to incorrect depth estimation and exaggerated boundaries in point clouds, resulting in severe distortions of geometric measurements, making flare suppression essential for safety-critical applications. Existing flare mitigation methods primarily operate at the hardware or signal-processing levels. While effective under specific configurations, they are largely rule-based and configuration-dependent, lacking learnable representations that generalize across diverse sensing scenarios. In this work, we reformulate flare suppression as a semantic segmentation problem, enabling data-driven learning of geometric and photometric cues directly from SPAD measurements. We first benchmark representative segmentation models on the newly introduced SPAD flare dataset and observe that they struggle to exploit the intrinsic multi-echo characteristics of SPAD signals. Motivated by this observation, we propose Physically-Informed segmentation for LiDAR Flare (PILF), a learning-based approach that treats the first and second echoes, together with ambient illumination, as distinct modalities, aggregating cross-echo information while jointly encoding geometric and photometric features. Experiments across multiple real-world scenes demonstrate that PILF significantly outperforms compared segmentation models, achieving up to 79.32% mIoU, and providing an effective solution for SPAD-based LiDAR flare suppression.