Image sensor with pixels

Abstract

An image sensor having pixels, in a case where a photodiode is in a saturated state, may transition from a saturated state to an unsaturated state during supply of a knee pulse from a knee pulse supply unit. The knee pulse supply unit preferably supplies knee pulses multiple times in one frame period.

Description

Technical Field

[0001] The present disclosure relates to an image sensor including a plurality of pixels, in particular, each pixel operates in a linear mode and a photovoltaic mode. Background Art

[0002] An image sensor includes a plurality of pixels, and each pixel outputs a signal corresponding to incident light to obtain a two-dimensional image. Each pixel generally obtains a signal corresponding to the amount of incident light per frame.

[0003] An image sensor is desired to have as wide a dynamic range as possible. One example discloses a technique in which a high-sensitivity signal with a long accumulation period and a low-sensitivity signal with a short accumulation period are added together to obtain a signal with a wide dynamic range.

[0004] The light amount of a light-emitting diode is controlled by PWM (Pulse-width modulation). Therefore, turning on / off the light emission is repeated at a predetermined frequency. Thus, when the image sensor acquires an image, the image when the light emission is on and the image when the light emission is off are mixed together, resulting in flicker.

[0005] Another example discloses a technique in which a peak-holding circuit is provided in each pixel to hold the peak of the signal, thereby eliminating flicker. Summary of the Invention

[0006] An image sensor according to the present disclosure includes a plurality of pixels. Each pixel includes a photodiode configured to operate in a linear mode, in which the photodiode exhibits a linear response to incident light, and also operates in a photovoltaic mode, in which the photodiode exhibits a logarithmic response to incident light, a reset transistor configured to reset the photodiode, a peak-holding circuit configured to integrate the output of the photodiode, wherein the peak-holding circuit includes (1) a peak-holding transistor configured to receive the output of the photodiode at the gate, (2) a switching transistor configured to turn on / off the output of the peak-holding transistor, and (3) a holding capacitor configured to accumulate the output of the switching transistor, and a knee pulse supply unit connected to the output terminal of the photodiode through a knee capacitor and configured to supply a knee pulse within a predetermined period to shift the voltage of the output terminal of the photodiode. In the case where the photodiode is in a saturated state, the photodiode can be switched from the saturated state to a non-saturated state within the period when the knee pulse supply unit supplies the knee pulse.

[0007] The knee pulse supply unit preferably supplies the knee pulse a plurality of times within one frame period.

[0008] Through the knee pulse, the saturated state of the photodiode may be temporarily cancelled, and the dynamic range may be broadened. In addition, providing the peak-holding circuit may make it possible to suppress the occurrence of flicker. Description of the Drawings

[0009] Embodiments of the present disclosure will be described based on the following diagrams, where:

[0010] Figure 1 is a diagram illustrating a pixel circuit configuration according to the present disclosure;

[0011] Figure 2 is a diagram illustrating the change over time of the output potential of the photodiode PD in the case of applying a knee pulse;

[0012] Figure 3 is a diagram where the horizontal axis represents the incident light amount and the vertical axis represents the output signal of the photodiode PD;

[0013] Figure 4 is a diagram illustrating the change over time of the output of the photodiode PD in the case where the turn-on period of the knee pulse and the incident time of the incident light are staggered;

[0014] Figure 5 is a diagram illustrating a pixel circuit configuration using multiple knee pulses;

[0015] Figure 6 is an illustration of Figure 5 the timing diagram of the circuit operation in, and is a diagram illustrating the change over time of the output of the photodiode PD;

[0016] Figure 7 is a diagram illustrating the change in the signal-to-noise ratio with respect to the incident light amount;

[0017] Figure 8 is a diagram illustrating an image sensor with a two-dimensional arrangement of pixels according to an embodiment;

[0018] Figure 9 is a timing diagram illustrating the operation of the image sensor;

[0019] Figure 10 is a diagram illustrating an example curve of the output voltage Vsig of the peak hold circuit in the case where the horizontal axis represents the logarithm of the incident light (Log(Isig)); and

[0020] Figure 11 is a diagram illustrating the output voltage characteristics with respect to the incident light in the case of using a two-step pulse as the knee pulse.

[0021] Explanation of the reference numerals in the drawings

[0022] 40: Output line;

[0023] 200: Image sensor.

[0024] 210: Pixel array

[0025] 212: Vertical scanning circuit

[0026] 214: Analog-to-digital converter

[0027] 216: Horizontal scanning circuit

[0028] Cc: Knee capacitor

[0029] Csig: Holding capacitor

[0030] Isig: Logarithm (Log)

[0031] P: Pixel

[0032] PD: Photodiode

[0033] PH: Peak-holding transistor

[0034] RESET: Reset potential

[0035] RST: Reset transistor

[0036] SEL: Selection transistor

[0037] SF: Source-follower transistor

[0038] SWsig: Switching transistor

[0039] t0, t1, t2: Time

[0040] Vsig: Output voltage Detailed implementation manners

[0041] An embodiment of the present disclosure will be described below with reference to the accompanying drawings. The following embodiments do not limit the present disclosure, and configurations obtained by selectively combining multiple diagrams are also included in the present disclosure.

[0042] Configuration of the pixel circuit

[0043] Figure 1 is a diagram illustrating the configuration of the pixel circuit according to the present disclosure. The photodiode PD accumulates charges (electrons in this example) based on incident light. The photodiode PD can operate in both linear mode and photovoltaic mode. In linear mode, the photodiode PD shows a linear response to incident light, and in photovoltaic mode, the photodiode PD shows a logarithmic response to incident light. The cathode of the photodiode PD serves as the output terminal, and the anode is connected to a power supply (e.g., ground).

[0044] One end (drain) of the reset transistor RST is connected to the output terminal of the photodiode PD, and the other end (source) of the reset transistor RST is connected to a predetermined power supply. In this example, the reset transistor RST is an n-channel transistor.

[0045] The output terminal of the photodiode PD is connected to the gate of the peak-holding transistor PH. The peak-holding transistor PH is a p-channel transistor. The source of the peak-holding transistor PH is connected to the injection power supply.

[0046] The source of the switch transistor SWsig is connected to the drain of the peak-holding transistor PH. One end of the holding capacitor Csig is connected to the drain of the switch transistor SWsig. The other end of the holding capacitor Csig is connected to the power supply (e.g., ground). The switch transistor SWsig is a p-channel transistor.

[0047] The peak-holding transistor PH, the switch transistor SWsig, and the holding capacitor Csig are configured as a peak-holding circuit.

[0048] The drain of the switch transistor SWsig is connected to the gate of the source-follower transistor SF. The drain of the source-follower transistor SF is connected to the power supply, and the source is connected to the drain of the selection transistor SEL. The source of the selection transistor SEL is connected to the output line 40.

[0049] The row selection signal is supplied to the gate of the selection transistor SEL. When the row selection signal becomes high level, the signal corresponding to the gate voltage of the source-follower transistor SF is output to the output line 40.

[0050] After the reset transistor RST is turned on to reset the photodiode PD, the switch transistor SWsig is turned on, and an injection pulse is applied to the source of the peak-holding transistor PH, thereby injecting charge (holes) into the holding capacitor Csig and discharging the excess charge through the peak-holding transistor PH. As a result, the holding capacitor Csig is placed in the reset state.

[0051] In this state, within one frame period, the signal supplied from the peak-holding transistor PH accumulates in the holding capacitor Csig through the switch transistor SWsig. As described above, the output voltage corresponding to the accumulated charge of the photodiode PD is supplied to the gate of the peak-holding transistor PH. Therefore, the charge corresponding to the incident light amount of the photodiode PD within one frame period accumulates in the holding capacitor Csig.

[0052] In addition, when the switch transistor SWsig is turned off and the selection transistor SEL is turned on, the voltage signal corresponding to the charge accumulated in the holding capacitor Csig is read out to the output line 40.

[0053] The knee pulse supply unit is connected to the output terminal of the photodiode PD through the knee capacitor Cc. The knee pulse supply unit is a power supply that generates a knee pulse, and the generated knee pulse is applied to the output terminal of the photodiode PD through the knee capacitor Cc.

[0054] The knee pulse is a short-period positive pulse. When the knee pulse is applied, the voltage at the output terminal of the photodiode PD shifts to the lower side. In other words, the voltage shifts only during a predetermined period when the knee pulse is at the H level. Therefore, in the case where the photodiode PD is in the saturation state, the photodiode PD can be switched to the non-saturation state. The amplitude of the voltage shift is the height of the knee pulse.

[0055] Figure 2 is a graph showing the time variation of the output potential of the photodiode PD in the case where the knee pulse is applied. In this example, the knee pulse is applied during the period from t1 to t2 at the end of the accumulation period from t0 to t2 corresponding to one frame. Figure 2 The vertical axis in indicates the potential of the photodiode PD (positive on the lower side) and the potential change caused by the incident light. Here, the intensity of the incident light is constant within one frame period. In the case of weak incident light, the potential gradually decreases from the reset potential RESET (in the figure, since electrons are negative charges as signals, the potential changes in the upward direction).

[0056] In the case of medium incident light, the potential reaches the potential of the linear saturation region in the middle of the accumulation period. When the knee pulse is applied at the start time t1 of the knee pulse, the potential decreases again. At the time t2 when the accumulation period ends, the potential becomes the potential corresponding to the incident light.

[0057] In the case of strong incident light, the potential reaches saturation in the linear region in the middle of the accumulation period and also reaches the knee potential during the knee period. After the potential reaches the knee potential, the potential enters the logarithmic region.

[0058] In Figure 2 the potential in the case of weak incident light is indicated by a dotted line, the potential in the case of medium incident light is indicated by a dashed line, and the potential in the case of strong incident light is indicated by a solid line.

[0059] Figure 3 is a graph whose horizontal axis represents the incident light amount and whose vertical axis represents the output signal of the photodiode PD. The sensitivity in the knee region relative to the sensitivity in the linear region is (t1 - t2) / (t1 - t0).

[0060] As described above, the knee region is provided, which enables the photodiode PD that has once entered the saturation state to be placed in the non-saturation state and accumulate charges.

[0061] Figure 2 and Figure 3 each illustrate the case where the incident light is fixed, that is, the intensity of the incident light is constant within one frame period. On the other hand, the light-emitting diode is used for signals and repeats turning on and off. As Figure 4As shown, when the turn-on period of the knee pulse and the incident time of the incident light are staggered from each other, the knee output shown by the thin solid line in the figure cannot be obtained, and the photodiode PD remains in the saturated state. Therefore, the advantages of applying the knee pulse cannot be obtained. If an incident light pulse exists within the period from t1 to t2, the knee output is obtained for PDout and PHout, that is, PDout and PHout increase within the period from t1 to t2. This means that LED flicker occurs because the magnitudes of PDout and PHout change depending on whether an incident light pulse exists within the period from t1 to t2.

[0062] Pixel circuit configuration using multiple pulses

[0063] Figure 5 is a diagram illustrating a pixel circuit configuration using multiple knee pulses. Different from Figure 1 multiple knee pulses in which a predetermined number of knee pulses are input within one frame are used.

[0064] In Figure 5 the example, within the LED turn-on period, at least one knee pulse is generated to suppress the flicker phenomenon through the knee region.

[0065] The turn-on time of the LED cannot be known in advance. Therefore, it is necessary to generate the knee pulse a sufficient number of times within the accumulation period (one frame period) so that at least one knee pulse is generated within the LED turn-on period.

[0066] In addition, regarding the output of the photodiode PD, when the knee pulse returns to the off state, the signal (signal having a knee characteristic) that has changed through charge accumulation within the knee pulse period returns to the saturation level in the linear region again. In the present embodiment, the peak hold circuit is connected to the output terminal of the photodiode PD. Therefore, the output of the peak hold circuit can be maintained at the level of the signal having a knee characteristic.

[0067] Figure 6 is a timing diagram illustrating the circuit operation in Figure 5 and is a diagram illustrating the time change of the output of the photodiode PD. When an incident light exists, one knee pulse is generated, and thus, charges (electrons) are further accumulated from the saturation level in the linear region. In addition, even after the knee pulse is turned off, the signal level is maintained by the peak hold circuit. In other words, the peak hold circuit in the subsequent stage has the characteristic of holding the peak of the output of the photodiode PD. Therefore, flicker occurs in the output of the photodiode PD, while flicker does not occur in the output of the peak hold circuit.

[0068] The number of times the LED is turned on in the signal varies depending on the signal manufacturer, etc. However, even in the case of LEDs from any manufacturer, the human eye does not see flicker. Therefore, it is estimated that the number of on / off times of the LED is 100 times or more per second. Considering the ease of control, the number of on / off times of the LED does not greatly exceed 100 times per second. When the accumulation period (one frame period) of the sensor is 33 milliseconds (30 times / second), the number of times the LED is turned on within this period is approximately three times. Therefore, when the number of knee pulses generated is 10 to 100 times the number of times the LED is turned on (30 to 300 times per frame), the knee pulses generated at any time correspond to the on period. Therefore, in this embodiment, 30 to 300 knee pulses are generated within one frame period.

[0069] The output of the photodiode PD and the output of the peak hold circuit were simulated in the case where the frame period is 10 milliseconds and the incident light pulse enters three times.

[0070] In the output of the photodiode PD, there are pulses (flicker) based on the voltage change of the knee pulse. When light enters, the voltage value of the pulse output by the photodiode PD shifts in response to the incident light.

[0071] The peak hold circuit shows a slow response and holds the peak of the output voltage of the photodiode PD. Therefore, in the output voltage of the peak hold circuit, the flicker caused by the knee pulse disappears. In addition, it was found that the output voltage of the photodiode PD is integrated. Therefore, at the end of one frame, the same output is obtained even if light enters at any time. As described above, flicker can be suppressed.

[0072] Figure 7 It is a graph showing the change of S / N with respect to the incident light amount. As shown in the figure, in the logarithmic region, S / N is the ratio of kTC thermal noise (kT / C) 1 / 2 and the thermal voltage kT / q, and is basically constant. In contrast, in the knee region, S / N is improved compared to S / N in the logarithmic region. Therefore, using the knee pulse can improve S / N in the knee region beyond the linear region.

[0073] Configuration of the image sensor

[0074] Figure 8It is a diagram showing an image sensor 200 with a two-dimensional pixel arrangement according to this embodiment. The pixel array 210 includes pixels P arranged in m columns * n rows (m * n) as described above, that is, including m pixels in the horizontal direction and n pixels in the vertical direction. The vertical scanning circuit (V-Scan) 212 sequentially selects the rows of the pixels P in the vertical direction. The pixels in each column are connected to the analog-to-digital converter (ADC) 214 through readout lines in the vertical direction. The horizontal scanning circuit (H-Scan) 216 is connected to the analog-to-digital converter (ADC) 214, and the image signals of the respective pixels are sequentially output from the horizontal scanning circuit (H-Scan) 216.

[0075] Figure 9 It is a timing diagram showing the operation of the image sensor 200. In the (k - 1)th row, after the two readouts are completed, the reset transistor RST is turned on to reset the photodiode PD. The reset is performed for each vertical period (1V = 1 frame). The exposure starts after one reset. In addition, during the exposure period, a plurality of knee pulses (denoted by Cc in the figure) are applied through the knee capacitor Cc. The number of knee pulses is about 30 to about 300, and the number of knee pulses is illustrated after being reduced in the figure. Here, the SWsig pulse is omitted.

[0076] The first readout of the signal is performed shortly before the next reset. The readout is performed in the above-described manner, that is, the selection transistor SEL is turned on, and the signal accumulated in the holding capacitor Csig is read out to the analog-to-digital converter (ADC) 214 through the output line 40. The analog-to-digital converter (ADC) 214 converts a plurality of analog P signals input from a plurality of pixels into a plurality of digital signals.

[0077] As described above, when preparing the signals of the respective pixels, the horizontal scanning circuit (H-Scan) 216 sequentially outputs the signals of m pixels.

[0078] Next, by shifting the horizontal period by 1H, the same operation is performed in the kth row. By repeating this operation n times, the signals of all m * n pixels can be read out.

[0079] Figure 10 It is a diagram showing an example curve of the output voltage Vsig of the peak holding circuit when the logarithm of the incident light (Log(Isig)) is represented on the horizontal axis.

[0080] A rectangular wave is used as the knee pulse. In the linear region, the output of the photodiode PD is not affected by the knee pulse. When the output of the photodiode PD saturates and enters the logarithmic region, a knee region is formed, and the output is improved in the knee region. In the logarithmic region, as described above, the output of the photodiode PD is not affected by the knee pulse. Therefore, using the knee pulse can increase the output in the knee region as shown by the dashed line in the figure.

[0081] Figure 11 It is a graph showing the characteristics of the output voltage with respect to the incident light when a two-step pulse is used as the knee pulse. As shown in the figure, when a two-step pulse is used, two knee regions appear, which allows the knee region to be expanded. Therefore, the dynamic range in the expanded knee region can be further extended. When two knee pulses are used, the photodiode PD can return to the unsaturated state after being saturated once, and the dynamic range can be extended compared to the case of increasing the voltage of one knee pulse.

Claims

1. An image sensor comprising a plurality of pixels, wherein: Each of the pixels comprises: A photodiode configured to operate in a linear mode, in which the photodiode exhibits a linear response to incident light, and also to operate in a photovoltaic mode, in which the photodiode exhibits a logarithmic response to the incident light, a reset transistor configured to reset the photodiode, A peak hold circuit configured to integrate the output of the photodiode, wherein the peak hold circuit comprises: a peak hold transistor configured to receive the output of the photodiode at a gate, a switch transistor configured to turn on / off the output of the peak hold transistor, and a hold capacitor configured to accumulate the output of the switch transistor, and a knee pulse supply unit connected to the output terminal of the photodiode through a knee capacitor and configured to supply a knee pulse for shifting the voltage of the output terminal of the photodiode within a predetermined period, and In a case where the photodiode is in a saturation state, the photodiode switches from the saturation state to a non-saturation state within a period in which the knee pulse is supplied from the knee pulse supply unit. 2 . The image sensor according to claim 1 , wherein the knee pulse supply unit supplies the knee pulse a plurality of times within one frame period. 3 . The image sensor according to claim 1 , wherein the knee pulse supply unit supplies the knee pulse 30 to 300 times within one frame period. 4 . The image sensor according to claim 1 , wherein the knee pulse is a two-step pulse that changes in a stepwise manner.

5. The image sensor according to claim 1, further comprising: A vertical scanning circuit; Multiple analog-to-digital converters; as well as A horizontal scanning circuit, wherein The plurality of pixels are arranged in a matrix manner, The vertical scanning circuit sequentially selects and drives the plurality of pixels arranged in the horizontal direction in the vertical direction, The analog-to-digital converter converts signals from the plurality of pixels selected by the vertical scanning circuit into a plurality of digital signals, and The horizontal scanning circuit outputs the plurality of digital signals in series. 6 . The image sensor according to claim 5 , wherein the knee pulse supply unit supplies the knee pulse a plurality of times within one frame period. 7 . The image sensor according to claim 5 , wherein the knee pulse supply unit supplies the knee pulse 30 to 300 times within one frame period.

8. The image sensor according to claim 5, wherein the knee pulse is a two-step pulse that changes in a stepwise manner.

Images