CN110519535A - A kind of Novel pixel structure of high sensitivity Larger Dynamic range - Google Patents

The present invention discloses a kind of Novel pixel structure of highly sensitive Larger Dynamic range, CMI module including the output image current using the feedback arrangement based on current mirror, connect with the CMI module be used to control the image current replicated reach output end selecting module up and down, with it is described above and below the transmission module that connect of selecting module；Photogenerated current can be reflected into integrated capacitance simultaneously by the CMI module；When current integration, voltage increases with the increase of the time of integration, and slope is directly proportional to pixel photoelectric current.The present invention substitutes traditional CMOS active pixel sensor using the dot structure based on current mirror mirror integration module, so that supply voltage is swung in a certain range, the actual maximum storage charge of pixel capacitance is expanded, in reconstruction stage according to the amplitude of oscillation and number of oscillations of supply voltage, so that the virtual enhancing of the operation voltage of circuit, to provide Larger Dynamic range mode for pixel.

一种高灵敏度大动态范围的新型像素结构A new pixel structure with high sensitivity and large dynamic range

技术领域technical field

本发明涉及图像传感器技术领域，特别是涉及一种高灵敏度大动态范围的新型像素结构。The invention relates to the technical field of image sensors, in particular to a novel pixel structure with high sensitivity and large dynamic range.

背景技术Background technique

大动态范围(High-Dynamic Range，HDR)和高灵敏度是大多数图像传感器无法提供的理想特性。大动态范围传感器能够在一幅图像中同时记录极亮和极暗信号。当场景的光照对比度较大时，高灵敏度对于图像传感器成像的性能同样重要。然而，大多数具有HDR操作的图像传感器往往缺乏灵敏度。Large dynamic range (High-Dynamic Range, HDR) and high sensitivity are ideal characteristics that most image sensors cannot provide. The large dynamic range sensor is capable of recording extremely bright and extremely dark signals simultaneously in one image. High sensitivity is also important for the imaging performance of the image sensor when the lighting contrast of the scene is large. However, most image sensors with HDR operation tend to lack sensitivity.

图像传感器是一组像素。像素电路通常与光电二极管连接，光电二极管将光转换成光电流，常用的图像传感器是有源像素传感器(Active pixel sensor,APS)，APS的响应几乎为线性，当它工作在离饱和较远的区域是，其输出电压与输入光电流成比例，其动态范围有限。在像素饱和之前可以被收集的最大电流I_sat由像素电路的最大电荷存储容量决定，而像素电路的最大电荷存储容量Q_max又取决于电路集成电容C_int和芯片中的最大供给电压VDD，由Q_max＝I_sat·ΔT，可得出I_sat＝C_int·VDD/ΔT。灵敏度可通过给定输入电流变化时像素电路输出端相应产生的电压差来测量。灵敏度越大，光学功率接近的两个输入光强越容易区分。由于输出电压与输入电流有关，它也可以解释为电路的直流过电压增益。对于作为积分器的线性像素电路，其灵敏度S可推导为， An image sensor is a collection of pixels. The pixel circuit is usually connected with a photodiode, which converts light into a photocurrent. A commonly used image sensor is an active pixel sensor (APS), and the response of the APS is almost linear. When it works far from saturation The region is, its output voltage is proportional to the input photocurrent, its dynamic range is limited. The maximum current I _sat that can be collected before the pixel is saturated is determined by the maximum charge storage capacity of the pixel circuit, and the maximum charge storage capacity Q _max of the pixel circuit depends on the circuit integration capacitance C _int and the maximum supply voltage VDD in the chip, determined by Q _max =I _sat ·ΔT, it can be obtained that I _sat =C _int ·VDD/ΔT. Sensitivity is measured by the voltage difference across the output of the pixel circuit for a given change in input current. The greater the sensitivity, the easier it is to distinguish between two input light intensities with similar optical powers. Since the output voltage is related to the input current, it can also be interpreted as the DC overvoltage gain of the circuit. For a linear pixel circuit as an integrator, its sensitivity S can be derived as,

为了改进APS的动态范围特性，主要采用对数或压缩传感器法、剪切传感器法、控制积分时间传感器法。第一种方法是压缩输出响应的增益，这是通过在整个输入电流范围内渐进地降低像素的跨导增益来实现的，从而降低了灵敏度。第二种类型调整输出节点电容，允许使用更大的电容来检测更大的输入光电流，因此也导致灵敏度下降。最后一种采用多次曝光的算法，减少了较大光电流的积分时间，允许HDR操作，但也降低了灵敏度。In order to improve the dynamic range characteristics of APS, the logarithmic or compression sensor method, the shear sensor method, and the control integration time sensor method are mainly used. The first method is to compress the gain of the output response, which is achieved by progressively reducing the transconductance gain of the pixel over the entire input current range, thereby reducing sensitivity. The second type adjusts the output node capacitance, allowing a larger capacitance to be used to detect a larger input photocurrent, thus also resulting in reduced sensitivity. The last algorithm, which uses multiple exposures, reduces the integration time for larger photocurrents, allowing HDR operation, but also reduces sensitivity.

为了重建图像，电压模式的CMOS图像传感器必须为每个像素提供关于时间和电压的信息。对于线性积分结构APS来说，有I_sat＝VDD/S。在不损害灵敏度的前提，可采用提高VDD的方式提高最大非饱和电流，在实际应用中，集成电路技术越新，VDD越低，使得实现HDR更加困难，并且倾向于降低像素的灵敏度。To reconstruct an image, a voltage-mode CMOS image sensor must provide information about time and voltage for each pixel. For the linear integral structure APS, there is I _sat =VDD/S. On the premise of not damaging the sensitivity, the maximum non-saturation current can be increased by increasing VDD. In practical applications, the newer the integrated circuit technology, the lower the VDD, which makes it more difficult to achieve HDR and tends to reduce the sensitivity of the pixel.

发明内容Contents of the invention

本发明的目的是针对现有技术中存在的技术缺陷，而提供一种高灵敏度大动态范围的新型像素结构。The purpose of the present invention is to provide a novel pixel structure with high sensitivity and large dynamic range aiming at the technical defects existing in the prior art.

为实现本发明的目的所采用的技术方案是：The technical scheme adopted for realizing the purpose of the present invention is:

一种高灵敏度大动态范围的新型像素结构，包括采用基于电流反射镜的反馈结构的输出镜像电流的CMI模块、与所述CMI模块连接的用于控制复制的镜像电流达到输出端的上下选择模块、与所述上下选择模块连接的传输模块；所述CMI模块能同时将光生电流反射到集成电容；当电流积分时，电压随积分时间的增加而增加，斜率与像素光电流成正比。A new pixel structure with high sensitivity and large dynamic range, including a CMI module that uses a feedback structure based on a current mirror to output a mirror current, an up and down selection module connected to the CMI module for controlling the copied mirror current to reach an output terminal, The transmission module connected with the upper and lower selection modules; the CMI module can simultaneously reflect the photogenerated current to the integrated capacitor; when the current is integrated, the voltage increases with the integration time, and the slope is proportional to the pixel photocurrent.

其中，所述CMI模块中包括四个晶体管构成的双电流镜、与双电流镜的输出连接的镜像电流复制模块以及光电探测器；其中所述光电探测器与读取电流的双电流镜连接；Wherein, the CMI module includes a dual current mirror composed of four transistors, a mirror current replication module connected to the output of the dual current mirror, and a photodetector; wherein the photodetector is connected to the dual current mirror for reading current;

所述镜像电流复制模块包括两个选择晶体管，分别对应的复制镜像电流，分别输出绝对值相同、方向相反的电流I_UP和I_DOWN到上下选择模块。The mirror current duplication module includes two selection transistors, which respectively correspond to duplicate mirror currents and output currents I _UP and I _DOWN with the same absolute value but opposite directions to the upper and lower selection modules.

其中，所述上下选择模块包括两个选择晶体管，构成上下开关，通过SW信号控制复制的镜像电流是否达到输出端。Wherein, the upper and lower selection module includes two selection transistors, which form an up and down switch, and control whether the replicated mirror current reaches the output terminal through the SW signal.

其中，所述传输模块包括通过一对差分的选择信号控制的传输门，所述传输门提供采样保持功能，并选定特定的像素模块的信号输出。Wherein, the transmission module includes a transmission gate controlled by a pair of differential selection signals, the transmission gate provides a sample and hold function, and selects a signal output of a specific pixel module.

本发明采用基于电流镜镜像积分(Current Mirroring Integration,CMI)模块的像素结构替代传统的有源像素传感器，使得电源电压在一定范围内摆动，像素电容实际的最大存储电荷得到扩展，在重构阶段根据电源电压的摆幅和摆动次数，使得电路的操作电压的虚拟增强，从而为像素提供大动态范围模式。The present invention adopts the pixel structure based on the Current Mirroring Integration (CMI) module to replace the traditional active pixel sensor, so that the power supply voltage swings within a certain range, the actual maximum stored charge of the pixel capacitor is expanded, and in the reconstruction stage Depending on the swing and the number of swings of the supply voltage, a virtual enhancement of the operating voltage of the circuit is made, thereby providing a large dynamic range mode for the pixel.

为防止集成电压达到饱和，本发明的像素结构的信号在参考信号上限处反弹，并继续向下积分，保持斜率，直到最终达到参考下限并反弹向上。每当信号在规定的电压差范围内饱和时，这个过程就会重复。通过这种反复改变的方式可以建立一个虚拟的电压信号，而虚拟电压信号比芯片实际的电压源高很多倍。在保证像素灵敏度的同时，极大地提高了像素结构的动态范围。In order to prevent the integration voltage from reaching saturation, the signal of the pixel structure of the present invention bounces at the upper limit of the reference signal, and continues to integrate downward, maintaining the slope until it finally reaches the lower limit of the reference and bounces upward. This process repeats every time the signal saturates within the specified voltage difference. A virtual voltage signal can be established through this repeated change method, and the virtual voltage signal is many times higher than the actual voltage source of the chip. While ensuring the pixel sensitivity, the dynamic range of the pixel structure is greatly improved.

本发明提出的像素结构在整个光照范围内具有恒定的灵敏度，由于其仅需要一次采样就可以达到较大的动态范围，像素的帧速率更高，并提供线性响应，允许在场景中进行均匀的对比度再现。此外，整个传感器直接将电路的输出值处理后得到最终图像，无需额外的行控制，也不需要任何二次输入，可以直接得出HDR图像。The pixel structure proposed by the present invention has a constant sensitivity over the entire illumination range, and since it only requires one sampling to achieve a large dynamic range, the frame rate of the pixel is higher, and provides a linear response, allowing uniform illumination in the scene. Contrast reproduction. In addition, the entire sensor directly processes the output value of the circuit to obtain the final image, without additional line control or any secondary input, and can directly obtain an HDR image.

附图说明Description of drawings

图1为本发明的高灵敏度大动态范围的新型像素结构图；Fig. 1 is the novel pixel structure diagram of the present invention with high sensitivity and large dynamic range;

图2为本发明的像素的电压信号时序图。FIG. 2 is a timing diagram of the voltage signal of the pixel of the present invention.

具体实施方式Detailed ways

以下结合附图和具体实施例对本发明作进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。The present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

如图1所示，本发明高灵敏度大动态范围的新型像素结构，包括CMI模块101、上下选择模块102和传输模块103，CMI模块采用基于电流反射镜的反馈结构，具有稳定的偏置电压和低输入阻抗，同时将光生电流反射到集成电容。当电流积分时，电压随积分时间的增加而增加，其斜率与像素光电流成正比。在CMI模块中，四个晶体管(104、105、106、107)构成双电流镜，在光电探测器115中读取电流，其中，I1＝I2＝I_DC。As shown in Figure 1, the new pixel structure with high sensitivity and large dynamic range of the present invention includes a CMI module 101, an up and down selection module 102, and a transmission module 103. The CMI module adopts a feedback structure based on a current mirror, and has a stable bias voltage and Low input impedance while reflecting the photogenerated current to the integrated capacitor. When the current is integrated, the voltage increases with the integration time, and its slope is proportional to the pixel photocurrent. In the CMI module, four transistors ( 104 , 105 , 106 , 107 ) form a double current mirror to read the current in the photodetector 115 , where I1 = I2 = I _DC .

晶体管108和晶体管109复制镜像电流，分别输出IUP和IDOWN，二者绝对值相同，方向相反。Transistor 108 and transistor 109 copy the mirror current and output IUP and IDOWN respectively, both of which have the same absolute value and opposite directions.

晶体管111、晶体管110构成上下开关，通过SW信号112控制这些电流是否达到输出端。The transistor 111 and the transistor 110 form an up and down switch, and the SW signal 112 is used to control whether these currents reach the output terminal.

最后，所选电流通过一对差分的select信号(113、114)控制的传输门，传输门提供采样保持功能，并选定特定的像素模块的信号输出。Finally, the selected current passes through a pair of transmission gates controlled by a pair of differential select signals (113, 114). The transmission gates provide a sample-and-hold function and select the signal output of a specific pixel module.

当积分电压达到低阈值或高阈值时，以正比于I_DC的绝对斜率反复弹跳，使之避免达到饱和。信号总是在ΔV＝V_max-V_min的任意范围内，并根据强度出现高峰和低谷。When the integrated voltage reaches the low or high threshold, it bounces repeatedly with an absolute slope proportional to I _DC to keep it from reaching saturation. The signal is always in an arbitrary range of ΔV = V _max - V _min with peaks and valleys according to intensity.

本发明中，积分时间ΔT内弹跳次数N＝5，积分时间结束后，电压信号V_CAP被采样。根据V_CAP和N，HDR信号即可重建。重建的虚拟信号V_R＝N·ΔV+(V_CAP-V_MIN)，可以比VDD高几倍，限制最大虚拟电压(V_{R_max})的唯一参数是最大允许摆动次数(N_max)。使用比特数为b_c的计数器计算信号弹跳，对于该像素，当计数器即将溢出时发生饱和电流，即，当时，最大饱和电流发生。在计数完成时，则由于I_sat与b_c成指数关系，像素灵敏度较高时也可以保持较大的最大非饱和电流。In the present invention, the number of bounces within the integration time ΔT is N=5, and after the integration time ends, the voltage signal V _CAP is sampled. According to V _CAP and N, the HDR signal can be reconstructed. The reconstructed dummy signal V _R =N·ΔV+(V _CAP -V _MIN ), can be several times higher than VDD, the only parameter limiting the maximum dummy voltage (VR — _max ) is the maximum number of allowed swings (N _max ). Count signal bounces using a counter of bits b _c , For this pixel, the saturation current occurs when the counter is about to overflow, that is, when , the maximum saturation current occurs. When counting is complete, then Since I _sat is exponentially related to b _c , a larger maximum non-saturation current can also be maintained when the pixel sensitivity is higher.

现有技术中提高CMOS图像传感器动态范围的方法是调整或修改APS，这总是导致强光下灵敏度下降，特别是当图像传感器需要在光照较强的环境下呈现一个良好的对比度时，这一缺点尤为明显。The way to improve the dynamic range of CMOS image sensors in the prior art is to adjust or modify the APS, which always leads to a decrease in sensitivity under strong light, especially when the image sensor needs to present a good contrast in a brightly lit environment. The disadvantages are particularly obvious.

本发明提出的基于电源电压自适应法的像素结构可避免了上述的限制，通过在相同积分时间内避免信号饱和，从而扩大最大的非饱和电流。通过计算可得，数字计数器位数为5位时，这种结构的像素的灵敏度相比常规APS可提高25倍。数字计数器的比特值更高时，灵敏度可以提高数百倍，综合传感器电路的复杂度增加和动态范围、灵敏度等性能的提高，可使用6～8位的计数器，即像素的最大弹跳次数为2⁶～2⁸。The pixel structure based on the power supply voltage self-adaptive method proposed by the present invention can avoid the above limitation, and can expand the maximum non-saturated current by avoiding signal saturation within the same integration time. It can be obtained by calculation that when the number of digits of the digital counter is 5, the sensitivity of the pixel with this structure can be increased by 25 times compared with the conventional APS. When the bit value of the digital counter is higher, the sensitivity can be increased hundreds of times, and the complexity of the sensor circuit is increased and the performance of the dynamic range and sensitivity is improved. A 6-8-bit counter can be used, that is, the maximum number of bounces of the pixel is 2 ⁶ ~ 2 ⁸ .

以上所述仅是本发明的优选实施方式，应当指出的是，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。The above is only a preferred embodiment of the present invention, it should be pointed out that, for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, these improvements and Retouching should also be regarded as the protection scope of the present invention.