研究目的
To present a novel method based on adaptive photon coincidence detection to suppress background light and improve dynamic range in SPAD-based LiDAR systems for outdoor applications like autonomous driving.
研究成果
Adaptive photon coincidence detection effectively suppresses background light and improves dynamic range by over 40 dB in SPAD-based LiDAR systems, enabling reliable distance measurements in high ambient light conditions. The method allows pixel-wise parameter adjustment for capturing high dynamic scenes, with potential for further enhancements through reduced filter bandwidth, shorter laser pulses, and advanced data processing.
研究不足
The theoretical model does not fully account for effects of photon coincidence detection circuits, leading to discrepancies at high photon rates. The adjustment algorithm may be slow in fast-changing environments, and the system's range is limited by factors like dead time and afterpulsing. Improvements are needed in histogram processing algorithms for better distance extraction.
1:Experimental Design and Method Selection:
The study uses a statistical model for photon coincidence detection, incorporating dead time effects, and implements adaptive parameter adjustment based on ambient light intensity.
2:Sample Selection and Data Sources:
Outdoor measurements with Lambertian targets of varying reflectivity (e.g., 8% and 60%) at 100 klx ambient sunlight.
3:List of Experimental Equipment and Materials:
A 192 × 2 pixel CMOS SPAD-based LiDAR sensor fabricated in a
4:35 μm CMOS process, flash LiDAR camera 'Owl' with FPGA, lens, pulsed laser sources (905 nm, 75 W peak power), and optical bandpass filters. Experimental Procedures and Operational Workflow:
The sensor measures time-of-flight using direct TOF technique, accumulates timestamps in histograms, and uses event counting mode to measure ambient event rates. Adaptive algorithm adjusts coincidence parameters (depth, time, number of SPADs) pixel-wise.
5:Data Analysis Methods:
Success probability calculated from histograms, standard deviation analysis, and comparison with theoretical models using statistical calculations.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容