CMOS图像传感器预处理数字电路的设计与实现
发布时间：2018-08-02 12:45
【摘要】：随着半导体工艺的提升和革新,对采用CMOS工艺的图像传感器有更加深入的研发。国内外研究机构和商业公司对CMOS图像传感器的兴趣越来越大,投入的资金也逐年增加。CMOS图像传感器芯片作为重要影像采集部件,有着巨大的商业市场,每年增加的市场需求是其广泛使用的重要原因。CMOS图像传感器在在电路集成度和功耗等方面的优势,是CCD图像传感器无法比拟的。这使得CMOS图像传感器在移动设备,数码相机,行车安全,安防监控,航天航空领域的应用越来越多,但是CMOS图像传感器在成像质量和噪声控制方面还有天生的弱势。本论文正是针对上述问题,将设计与实现的对象锁定为CMOS图像传感器预处理数字电路。首先使用一个基于FPGA开发板的平台和MIS100X系列的130W像素的CMOS图像传感器进行图像采集,采集到的RAW格式图像数据作为算法验证对象,同时研究了主流的CMOS图像传感器的像素单元,及MIS100X系列图像传感器的时序和驱动,以便准确的采集所需的图像数据。在深入研究分析了CMOS图像传感器的像素单元的重要参数,主要噪声来源,以及推导了相关的噪声公式后,重点针对影响成像质量最大因素:暗电流,坏点,行噪声,列方向条纹噪声问题。在分析已有的抑制方法的优缺点的基础上,分别提出基于数字电路方式抑制噪声的方法,并使用芯片采集的RAW数据,在Matlab软件中对方法进行了图像验证,分模块设计了暗电流抑制算法,坏点去除算法,行噪声抑制算法,列方向条纹噪声算法,最终结果均达到预期效果。本论文将设计好的算法使用Verilog语言实现了硬件电路的描述,按照数字电路前端设计的流程,分模块形成了暗电流抑制数字电路,坏点去除数字电路,行噪声抑制电路,列方向条纹噪声抑制电路的RTL级代码,并详细说明了每个模块的输入和输出端口,数据流向,功能框图。最后将写好的CMOS预处理电路RTL级代码按照流程进行了仿真验证,验证平台是基于MIS100X系列的芯片的顶层验证平台。针对预处理电路的功能和时序制定了验证方案,分模块和Testcase对RTL代码进行了仿真,并使用FPGA进行了测试,结果均达到了设计预期。
[Abstract]:With the improvement and innovation of semiconductor technology, the image sensor based on CMOS process has been developed more deeply. Domestic and foreign research institutions and commercial companies are more and more interested in CMOS image sensors, and the funds invested are increasing year by year. As an important part of image acquisition, there is a huge commercial market. The increasing market demand every year is an important reason for its wide use. The advantages of CMOS image sensors in circuit integration and power consumption are unparalleled by CCD image sensors. This makes CMOS image sensors more and more used in mobile devices, digital cameras, vehicle safety, security monitoring, aerospace applications, but CMOS image sensors in imaging quality and noise control are inherently weak. In order to solve the above problems, this paper locks the designed and implemented object into a preprocessing digital circuit of CMOS image sensor. Firstly, a platform based on FPGA development board and a 130W pixel CMOS image sensor of MIS100X series are used for image acquisition. The collected image data in RAW format are used as the verification object of the algorithm. At the same time, the pixel unit of the mainstream CMOS image sensor is studied. And MIS100X series image sensor timing and drive, in order to accurately collect the required image data. After deeply studying and analyzing the important parameters of pixel unit of CMOS image sensor, the main noise sources, and deducing the related noise formula, we focus on the most important factors affecting the imaging quality: dark current, bad point, line noise, etc. The problem of directional stripe noise. Based on the analysis of the advantages and disadvantages of the existing suppression methods, the noise suppression methods based on the digital circuit are proposed, and the method is verified in the Matlab software by using the RAW data collected by the chip. The dark current suppression algorithm, the bad point removal algorithm, the row noise suppression algorithm and the column direction stripe noise algorithm are designed in each module, and the final results reach the expected results. In this paper, we use Verilog language to describe the hardware circuit. According to the design flow of the front end of the digital circuit, the dark current suppression digital circuit is formed, the bad point is removed, and the line noise suppression circuit is implemented. The RTL level code of the column direction stripe noise suppression circuit is described in detail. The input and output ports of each module, the direction of the data flow, and the function block diagram are described in detail. Finally, the RTL level code of the CMOS preprocessing circuit is simulated and verified according to the flow chart. The verification platform is the top level verification platform based on the MIS100X series chips. According to the function and timing of the preprocessing circuit, the verification scheme is established. The RTL code is simulated by module and Testcase, and tested by FPGA. The results reach the design expectation.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212
本文编号：2159482
[Abstract]:With the improvement and innovation of semiconductor technology, the image sensor based on CMOS process has been developed more deeply. Domestic and foreign research institutions and commercial companies are more and more interested in CMOS image sensors, and the funds invested are increasing year by year. As an important part of image acquisition, there is a huge commercial market. The increasing market demand every year is an important reason for its wide use. The advantages of CMOS image sensors in circuit integration and power consumption are unparalleled by CCD image sensors. This makes CMOS image sensors more and more used in mobile devices, digital cameras, vehicle safety, security monitoring, aerospace applications, but CMOS image sensors in imaging quality and noise control are inherently weak. In order to solve the above problems, this paper locks the designed and implemented object into a preprocessing digital circuit of CMOS image sensor. Firstly, a platform based on FPGA development board and a 130W pixel CMOS image sensor of MIS100X series are used for image acquisition. The collected image data in RAW format are used as the verification object of the algorithm. At the same time, the pixel unit of the mainstream CMOS image sensor is studied. And MIS100X series image sensor timing and drive, in order to accurately collect the required image data. After deeply studying and analyzing the important parameters of pixel unit of CMOS image sensor, the main noise sources, and deducing the related noise formula, we focus on the most important factors affecting the imaging quality: dark current, bad point, line noise, etc. The problem of directional stripe noise. Based on the analysis of the advantages and disadvantages of the existing suppression methods, the noise suppression methods based on the digital circuit are proposed, and the method is verified in the Matlab software by using the RAW data collected by the chip. The dark current suppression algorithm, the bad point removal algorithm, the row noise suppression algorithm and the column direction stripe noise algorithm are designed in each module, and the final results reach the expected results. In this paper, we use Verilog language to describe the hardware circuit. According to the design flow of the front end of the digital circuit, the dark current suppression digital circuit is formed, the bad point is removed, and the line noise suppression circuit is implemented. The RTL level code of the column direction stripe noise suppression circuit is described in detail. The input and output ports of each module, the direction of the data flow, and the function block diagram are described in detail. Finally, the RTL level code of the CMOS preprocessing circuit is simulated and verified according to the flow chart. The verification platform is the top level verification platform based on the MIS100X series chips. According to the function and timing of the preprocessing circuit, the verification scheme is established. The RTL code is simulated by module and Testcase, and tested by FPGA. The results reach the design expectation.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212
【参考文献】
相关期刊论文 前10条
1 王华;曹剑中;王华伟;郭惠楠;赵晓冬;刘广森;;基于估计方法的CMOS图像传感器列固定模式噪声校正方法[J];红外与激光工程;2013年07期
2 董建婷;杨小乐;董杰;石志城;;面阵CMOS图像传感器响应非均匀性校正方法研究[J];半导体光电;2012年04期
3 王艳;李斌桥;徐江涛;;一种消除CMOS图像传感器行噪声的时序[J];传感技术学报;2011年10期
4 张钰;;一种新的片上抑制CMOS图像传感器暂态噪声方法[J];传感技术学报;2011年06期
5 周津;姚素英;徐江涛;;列固定模式噪声的校正方法[J];天津大学学报;2008年01期
6 李芹;蔡理;吴刚;王森;;纳米CMOS图像传感器及其在航空航天中的应用[J];微纳电子技术;2007年11期
7 徐江涛;姚素英;朱天成;;CMOS有源像素传感器光响应分析及实验模型建立[J];传感技术学报;2007年02期
8 刘宇;王国裕;;CMOS图像传感器固定模式噪声抑制新技术。[J];固体电子学研究与进展;2006年03期
9 张生才;董博彦;徐江涛;;光电二极管有源像素CMOS图像传感器固定模式噪声分析[J];传感技术学报;2005年04期
10 裴志军,国澄明,姚素英,赵毅强;一种CMOS图像传感器列FPN抑制技术[J];传感技术学报;2004年01期
相关博士学位论文 前1条
1 谭鑫;Bayer图像序列噪声抑制技术研究[D];国防科学技术大学;2014年
相关硕士学位论文 前2条
1 邹义平;CMOS图像传感器的图像降噪技术的研究[D];北京邮电大学;2009年
2 崔博;CMOS图像传感器的噪声研究与抑制电路设计[D];华中科技大学;2007年
，本文编号：2159482
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2159482.html
