A Blooming-Region-Aware Directional Weighting Filter for Correcting Blooming-Induced Background Depth Distortion in AMCW-ToF LiDAR Imaging of a High-Reflectance Object
Open Access
Abstract
When measuring a high-reflectance target using time-of-flight LiDAR (ToF LiDAR), strong reflected optical signals are generated. Pixels receiving such strong reflected optical signals become saturated, and intensity and depth distortions occur even around the saturated pixels owing to crosstalk, internal lens scattering, and related effects. Previous studies on depth distortion in ToF cameras and LiDAR caused by such blooming have mainly focused on removing distorted background pixels, which limits their ability to directly correct background depth distortion around high-reflectance targets. In this paper, we propose a blooming-region-aware directional weighting filter method that corrects blooming-induced background depth distortion around a high-reflectance target using only the intensity map and depth map acquired from a single amplitude-modulated continuous-wave time-of-flight (AMCW-ToF) LiDAR. The proposed method detects the blooming-affected background region based on the intensity map and the gradient information of the intensity map. Subsequently, a window kernel is applied to each correction target pixel within the detected region, and the distorted depth of the correction target pixel is corrected through multi-weight depth filtering that considers the directional relation among pixels inside the kernel. In the validation process, the correction performance was evaluated in terms of mean squared error (MSE) and root mean squared error (RMSE) according to changes in the filtering method, target distance, target direction, and background SNR. The validation results demonstrate that the proposed method reduced the RMSE by an average of 49.811% compared with the blooming-distorted depth under various experimental conditions, indicating that it can mitigate background depth distortion around high-reflectance targets.Keywords
Funding Information
- Ministry of Education (2026-ANCHOR-06-002)
- National Research Foundation of Korea (RS-2025-16043000)
- Korea Research Institute of Standards and Science (KRISS-2026-KP2026-0002)
- Institute of Information & Communications Technology Planning & Evaluation (IITP)-ICAN (IITP-2026-RS-2022-00156212)
This publication has 39 references indexed in Scilit:
- Novel method for robust bilateral filtering point cloud denoisingAlexandria Engineering Journal, 2025
- AdverseNet: A LiDAR Point Cloud Denoising Network for Autonomous Driving in Rainy, Snowy, and Foggy WeatherIEEE Sensors Journal, 2025
- SelfredepthJournal of Real-Time Image Processing, 2024
- Denoising Framework Based on Multiframe Continuous Point Clouds for Autonomous Driving LiDAR in Snowy WeatherIEEE Sensors Journal, 2024
- Simultaneous Multifrequency Demodulation for Single-Shot Multiple-Path ToF ImagingIEEE Transactions on Computational Imaging, 2024
- Multipath Interference Suppression of Amplitude-Modulated Continuous Wave Coaxial-Scanning LiDAR Using Model-Based Synthetic Data LearningIEEE Sensors Journal, 2023
- Highly precise AMCW time-of-flight scanning sensor based on parallel-phase demodulationMeasurement, 2022
- A Review of Indirect Time-of-Flight TechnologiesIEEE Transactions on Electron Devices, 2022
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016
- A model of saliency-based visual attention for rapid scene analysisIEEE Transactions on Pattern Analysis and Machine Intelligence, 1998