【作者】 陈强；

【导师】 刘旭； 徐鹏霄；

【作者基本信息】 东南大学 ， 电子信息（专业学位）， 2024， 硕士

【摘要】 本研究探讨了夜间监控、安防及医学成像等领域低照度CMOS图像传感器的应用和需求,这一类传感器在近年来已逐渐成为图像传感器技术研究的重点。本文详细分析了低照度CMOS图像传感器的成像性能,并对成像系统进行了综合的软硬件设计和优化。此外,本研究还开发了专门针对低照度环境的图像处理算法,以改善成像质量和效率。本文首先综述了CMOS图像传感器及微光夜视技术的基本原理与发展历程,以及这些技术在现代成像系统中的应用。接着,本文设计了一种新型的基于现场可编程门阵列(FPGA)的低照度CMOS图像传感器成像系统。此系统中,FPGA作为核心处理单元,承担了CMOS图像传感器的驱动、读出与成像的任务。系统设计完成后,对硬件电路进行了严格的调试,随后对包括基础成像模块和图像处理模块的成像电路系统逻辑进行设计。基础成像模块涵盖了低照度CMOS图像传感器的驱动模块、系统主控模块、时钟管理模块、DDR3高速缓存模块以及HDMI显示模块,能够显示1936×1280分辨率、30帧每秒、14比特色深的高质量图像。本文在图像处理模块中完成了直方图均衡化算法的应用,还提出了自适应直方图均衡算法,针对微光环境下的成像效果进行了优化,增强了图像的可视性。最后,针对低照度环境下图像噪声水平较高的问题,本研究创新性的基于随机共振原理设计并实现了随机共振图像算法,并完成了在低照度图像传感器中的应用。本文还探索了低照度条件下CMOS图像传感器的成像性能,基于最新的国际标准EMVA1288创新性的构建了应用于低照度CMOS图像传感器的成像性能测试平台,设计新的测试流程,以确保测试结果的准确性和可靠性。通过理论分析与实验研究紧密结合,对饱和信号、全局增益、量子效率、光谱响应度、信噪比、动态范围、暗电流及非均匀性等多个性能参数进行了测量和评估。每一参数的测量都旨在评估成像系统在极端光照条件下的显示能力和稳定性,进一步分析了这些参数与图像质量之间的关联。综上所述,本文开发的低照度CMOS图像传感器成像系统能够在急剧降低光照条件下稳定地显示图像。该系统整合了新型的图像处理算法,显著提升了在低光环境下的成像性能,实现了对采集图像的实时处理和优化,从而提高了图像的清晰度和精确性。本文详尽地描述了系统的硬件配置、软件开发及性能评估方法,并通过一系列控制实验,证实了系统在低照度条件下的高效性与可靠性。研究表明,该成像系统具备高速、高精度及低功耗的特点,预示着其在夜视、安防、机器视觉及医学成像等多个领域的广泛应用潜力。更多还原

【Abstract】 This study explores the application and demand of low-light CMOS image sensors in fields such as nighttime monitoring,security,and medical imaging for these sensors have increasingly become a focus in recent years.The thesis thoroughly analyzes the imaging performance of low-light CMOS image sensors and carries out comprehensive hardware and software design and optimization of the imaging system.Furthermore,image processing algorithms specifically for low-light environments has been developed to improve imaging quality and efficiency.The thesis begins with reviewing the basic principles and history of CMOS image sensors and low-light night vision technology,as well as their applications in modern imaging systems.Subsequently,based on Field-Programmable Gate Array(FPGA)a novel imaging system for low-light CMOS image sensors is designed.In this system,the FPGA serves as the core pro-cessing unit,handling the driving,readout,and imaging tasks of the CMOS image sensor.The hardware circuits are debugged to ensure accuracy,followed by the implementation of the im-aging circuit’s system logic design,which includes basic imaging and image processing mod-ules.The basic imaging module consists of the low-light CMOS image sensor’s drive module,system control module,clock management module,DDR3 high-speed cache module,and HDMI display module,achieving high-quality image display with specifications of 1936×1280resolution,30 frames per second,and 14-bit color depth.Moreover,the imaging quality under low-light conditions is optimized by an adaptive histogram equalization algorithm to enhance the visibility of images.Lastly,this thesis designed and implemented a stochastic resonance algorithm based on the principle of stochastic resonance,and completed its application in low illumination image sensors.This thesis also explores the imaging performance of CMOS image sensors under low-light conditions and specifically constructs testing platform,conceives testing procedures ac-cording to the latest industry standard EMVA1288 to ensure the accuracy and reliability of the test results.Multiple key performance parameters including saturation signal,global gain,quan-tum efficiency,spectral response,signal-to-noise ratio,dynamic range,dark current,and non-uniformity are analyzed theoretically and experimentally focusing on the display capability and stability under extreme lighting conditions.The relationship between these parameters and im-age quality is further investigated.In summary,the proposed low-light CMOS image sensor imaging system can stably dis-play images under heavily reduced illumination conditions.The system integrates advanced image processing algorithms,significantly enhances imaging performance,and achieves real-time processing and optimization of captured images,guaranteeing image clarity and accuracy.The system’s hardware configuration,software development,and performance evaluation methods are comprehensively described.Through a series of elaborated experiments,the high-efficiency and reliability under low-light conditions is obtained.The research demonstrates that the imaging system featured by high speed,high precision,and low power consumption could benefit a broad of application fields such as night vision,security,machine vision,and medical imaging.更多还原