Author:
Li Qin,Zhu Huifeng,Huang Guyue,Yu Zijie,Qiao Fei,Wei Qi,Liu Xinjun,Yang Huazhong
Abstract
Purpose
The smart image sensor (SIS) which integrated with both sensor and smart processor has been widely applied in vision-based intelligent perception. In these applications, the linearity of the image sensor is crucial for better processing performance. However, the simple source-follower based readout circuit in the conventional SIS introduces significant nonlinearity. This paper aims to design a low-power in-pixel buffer circuit applied in the high-linearity SIS for the smart perception applications.
Design/methodology/approach
The linearity of the SIS is improved by eliminating the non-ideal effects of transistors and cancelling dynamic threshold voltage that changes with the process variation, voltage and temperature. A low parasitic capacitance low leakage switch is proposed to further improve the linearity of the buffer. Moreover, an area-efficient SIS architecture with a sharing mechanism is presented to further reduce the number of in-pixel transistors.
Findings
A low parasitic capacitance low leakage switch and a gate-source voltage pre-storage method are proposed to further improve the linearity of the buffer. Nonlinear effects introduced by parasitic capacitance switching leakage, etc., have been investigated and solved by proposing low-parasitic and low-leakage switches. The linearity is improved without a power-hungry operational amplifier-based calibration circuit and a noticeable power consumption increment.
Originality/value
The proposed design is implemented using a standard 0.18-µm CMOS process with the active area of 102 µm2. At the power consumption of 5.6 µW, the measured linearity is −63 dB, which is nearly 27 dB better than conventional active pixel sensor (APS) implementation. The proposed low-power buffer circuit increase not only the performance of the SIS but also the lifetime of the smart perception system.
Subject
Electrical and Electronic Engineering,Industrial and Manufacturing Engineering
Reference15 articles.
1. Low-power CMOS laser doppler imaging using non-CDS pixel readout and 13.6-bit saradc;IEEE Transactions on Biomedical Circuits and Systems,2016
2. Processing near sensor architecture in mixed-signal domain with CMOS image sensor of Convolutional-Kernel-Readout method;IEEE Transactions on Circuits and Systems I: Regular Papers,2019
3. CMOS image sensor with area-efficient block-based compressive sensing;IEEE Sensors Journal,2015
4. CMOS photodetector systems for low-level light applications;Journal of Materials Science: Materials in Electronics,2009
5. Advances on CMOS image sensors;Sensor Review,2016
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献