超宽动态范围图像传感器的研究

发布时间：2018-06-15 22:23

本文选题：CMOS图像传感器 + 宽动态范围　；参考：《吉林大学》2016年博士论文

【摘要】：本论文的主要内容为设计了一款使用电荷补偿型光电晶体管的超宽动态范围图像传感器,并基于传感器芯片设计制成了完整的成像系统,性能与成像测试的结果都显示该系统为高动态范围成像提供了一种很好的解决方案。本论文首先在分析已有的高动态范围成像的优缺点的基础上,提出一种应用于CMOS图像传感器的新型光电器件----电荷补偿型光电晶体管。该晶体管设计完全基于标准的商用CMOS工艺,在实现过程中无需任何额外的工艺结构,具有良好的通用性。所设计的电荷补偿型光电晶体管在传统的光电二极管的基础之上,增加了一个被称之为补偿二极管的光电二极管,其正极接至可调电压源上,负极与主光电二极管的负极直接相连。这样将其应用于有源像素结构(APS)时,当入射光强比较弱或积分时间比较短时,两个光电二极管均处于反偏压工作状态,共同作为光生电流源,从像素电容抽取电荷。这种方式没有因为引入新结构而降低实际的像素的填充因子,因此保持了其在弱光下的响应率。而当入射光强较强或积分时间较长时,补偿二极管转入正向导通状态,通过电压源向像素电容快速地补充电荷直至与主光电二极管的光生电流达到平衡状态。平衡点的高低只与入射光强有关,且与入射光强的大小呈对数响应,像素输出不易饱和,从而大大地拓展了在强光下的动态范围。器件仿真和测试结果显示,这种新型光电器件可探测光强的动态范围达到了167.1d B,已经接近了自然场景的动态范围。本文基于电荷补偿型光电晶体管设计了对应的图像传感器芯片,并针对器件测试中发现的问题,尤其是在极强光照下的温度和寄生电阻等所带来的影响进行了深入分析,并提出了相应的补偿手段。在成像系统的设计过程中,我们针对传感器特殊的光电响应曲线提出了一种近似还原算法,在保证暗场信号细节不丢失的前提下更多地容纳了强光下物体的成像细节,实际成像结果显示其具有很好的高动态成像能力。此外,针对原型系统成像过程中的噪声等问题,我们也提出了相应的优化手段,并设计了一种可应用于CMOS图像传感器的12bit Cyclic ADC,仿真结果显示其有效精度超过10bit,可以很好地应用于图像传感器系统。
[Abstract]:The main content of this thesis is to design an ultra-wide dynamic range image sensor using charge compensation phototransistor, and design a complete imaging system based on sensor chip. Both performance and imaging test results show that the system provides a good solution for high dynamic range imaging. Based on the analysis of the advantages and disadvantages of the existing high dynamic range imaging, this paper presents a novel photoelectric device, charge compensated phototransistor, which is used in CMOS image sensor. The design of the transistor is completely based on standard commercial CMOS process, and no extra process structure is needed in the process of implementation, so it has good versatility. On the basis of the traditional photodiode, the designed charge compensated phototransistor adds a photodiode called the compensation diode, whose positive pole is connected to the adjustable voltage source. The negative electrode is directly connected to the negative electrode of the main photodiode. When the incident light intensity is weak or the integration time is short, the two photodiodes are in the state of reverse bias and are used as photo-generated current sources to extract charges from pixel capacitors. This method does not reduce the pixel filling factor because of the introduction of the new structure, so it maintains its response rate under weak light. When the incident light intensity is strong or the integration time is longer, the compensation diode is transferred to the positive pass state, and the voltage source is used to rapidly replenish the charge to the pixel capacitance until the photo-generated current of the main photodiode reaches equilibrium state. The balance point is only related to the intensity of the incident light, and is logarithmic response to the intensity of the incident light. The pixel output is not easy to saturate, thus greatly expanding the dynamic range under the strong light. The simulation and test results show that the dynamic range of the new optoelectronic device detectable light intensity reaches 167.1 dB, which is close to the dynamic range of natural scene. In this paper, the corresponding image sensor chip is designed based on the charge compensated phototransistor, and the problems found in the device test, especially the influence of temperature and parasitic resistance under extremely strong illumination, are analyzed deeply. And put forward the corresponding compensation means. In the design of the imaging system, we propose an approximate reduction algorithm for the special photoelectric response curve of the sensor, which can accommodate the imaging details of the object under strong light without losing the details of the dark field signal. The actual imaging results show that it has good dynamic imaging capability. In addition, in view of the noise in the imaging process of the prototype system, we also put forward the corresponding optimization methods. A novel 12bit cyclic ADC for CMOS image sensor is designed. The simulation results show that the precision is over 10 bits, and it can be applied to the image sensor system.
【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP212;TP391.41

【相似文献】