动态范围
噪音(视频)
功率(物理)
计算机科学
补偿(心理学)
像素
航程(航空)
信号(编程语言)
光学
声学
电气工程
电子工程
物理
材料科学
工程类
人工智能
复合材料
精神分析
程序设计语言
图像(数学)
量子力学
心理学
作者
Canxing Piao,Yeonsoo Ahn,Donguk Kim,Jihoon Park,Jubin Kang,Min‐Seok Shin,Kangbong Seo,Seong‐Jin Kim,Jung-Hoon Chun,Jaehyuk Choi
标识
DOI:10.1109/a-sscc53895.2021.9634817
摘要
Indirect time-of-flight (iTOF) is a 3D depth-sensing technology that provides the distance to the object by measuring the phase difference of emitted and reflected waves of light. Usually, iTOF sensors consume power over 200 mW owing to high frequency modulation, which prevents them from the application to the energy limited wearable devices for VR/AR. To minimize column fixed depth noise (FDN), the TX driver typically has a double-sided clock tree that provides TX signals from both the left and the right sides of the pixel array [1], [2]. However, the double-sided driver consumes significant power even though the column FDN is still present. Moreover, the clock tree generates a huge peak current to induce depth noise. Some sensors employed a clock chain to distribute the peak current [3], [4]. Because the chain induces inherent row FDN, they used a DLL [3] or two opposite-directional chains [4] for the post compensation. However, these schemes still have column FDN, power consumption from the double-sided driver, and PVT variation. Another important issue is a limited depth dynamic range (DDR) because reflected light power falls along with the squared distance. Saturation occurs for short range (SR) whereas noise overwhelms signal for long range (LR). For the WDDR, two or multiple shots of images should be synthesized while consuming additional power.
科研通智能强力驱动
Strongly Powered by AbleSci AI