激光雷达
限制
遥感
压缩(物理)
闪光灯(摄影)
数据压缩
计算机科学
环境科学
脉冲压缩
数据采集
压缩比
光学
作者
Torben Neumann,Franz Kallage
标识
DOI:10.1109/jsen.2026.3720559
摘要
Flash LiDAR systems based on single-photon avalanche diode (SPAD) arrays face significant data compression challenges due to high-resolution pixel matrices and the need for real-time processing. This article presents a comprehensive overview of data compression methods for SPAD-based Flash LiDAR systems, focusing on the specific challenges introduced by photon-statistical effects, especially “pile-up.” We provide an empirical demonstration of how object surface color significantly influences this effect and explain why this is less pronounced in scanning LiDAR sensors. We evaluate and expand three compression approaches: the modified Shifted inter-frame Histogram (SifH) method, sparse coding using Gray-based Fourier encoding, and Golomb–Rice (GR) coding as a reference. Our analysis considers critical system parameters including measurement distances up to 50 m and a 100-ps time-to-digital converter resolution. A detailed comparative study of compression performance across distances and illumination conditions is given. We suggest a concept for the extension of both SifH and sparse coding to support multiobject detection in compressed data streams. Results show that the modified SifH method provides a hardware-oriented solution suitable for multiple peaks with minimal processing overhead, while sparse coding offers greater compression potential (up to$32\times $) but faces challenges in correcting pile-up effects. Both methods can maintain acceptable distance measurement accuracy while significantly reducing data volume, with the optimal choice depending on background light conditions and required precision.
科研通智能强力驱动
Strongly Powered by AbleSci AI