WO2013128817A1 - Dispositif de capture d'image à semi-conducteurs - Google Patents

Dispositif de capture d'image à semi-conducteurs Download PDF

Info

Publication number
WO2013128817A1
WO2013128817A1 PCT/JP2013/000786 JP2013000786W WO2013128817A1 WO 2013128817 A1 WO2013128817 A1 WO 2013128817A1 JP 2013000786 W JP2013000786 W JP 2013000786W WO 2013128817 A1 WO2013128817 A1 WO 2013128817A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
correction data
unit
pixel
characteristic
Prior art date
Application number
PCT/JP2013/000786
Other languages
English (en)
Japanese (ja)
Inventor
将積　直樹
Original Assignee
コニカミノルタ株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by コニカミノルタ株式会社 filed Critical コニカミノルタ株式会社
Priority to JP2013515614A priority Critical patent/JP5299597B1/ja
Publication of WO2013128817A1 publication Critical patent/WO2013128817A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
    • H04N25/50—Control of the SSIS exposure
    • H04N25/57—Control of the dynamic range
    • H04N25/571—Control of the dynamic range involving a non-linear response
    • H04N25/573—Control of the dynamic range involving a non-linear response the logarithmic type
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
    • H04N25/60—Noise processing, e.g. detecting, correcting, reducing or removing noise
    • H04N25/616—Noise processing, e.g. detecting, correcting, reducing or removing noise involving a correlated sampling function, e.g. correlated double sampling [CDS] or triple sampling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
    • H04N25/70—SSIS architectures; Circuits associated therewith
    • H04N25/76—Addressed sensors, e.g. MOS or CMOS sensors
    • H04N25/77—Pixel circuitry, e.g. memories, A/D converters, pixel amplifiers, shared circuits or shared components
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
    • H04N25/70—SSIS architectures; Circuits associated therewith
    • H04N25/76—Addressed sensors, e.g. MOS or CMOS sensors
    • H04N25/78—Readout circuits for addressed sensors, e.g. output amplifiers or A/D converters

Definitions

  • the present invention relates to a solid-state imaging device that includes a pixel portion having two photoelectric conversion characteristics, that is, a linear characteristic and a logarithmic characteristic, and is capable of imaging with a wide dynamic range.
  • CMOS solid-state imaging device in which pixels and peripheral circuits are CMOS type is based on a CMOS / LSI manufacturing process, and it is easy to manufacture pixels and peripheral circuits in the same process. It's being used. In these CMOS type solid-state imaging devices, there is a problem that the accumulated charge is saturated when there is strong incident light, and a wide dynamic range cannot be obtained.
  • CMOS-type solid-state imaging devices aimed at wide dynamic range imaging have been developed.
  • a linear logarithmic conversion type solid-state imaging device as shown in Patent Document 1 has been proposed.
  • This solid-state imaging device outputs a signal that changes linearly with respect to incident light when the amount of incident light is small by setting the potential state of the transfer transistor connected to the photodiode that performs photoelectric conversion to an appropriate state.
  • the photoelectric conversion element leaks charges and outputs a logarithmically changing signal.
  • Patent Document 2 describes a method for suppressing fixed pattern noise caused by variations in inflection points. According to this technique, the photoelectric conversion characteristics of each pixel are measured in advance and stored in a storage device as correction data. Then, at the time of actual imaging, the pixel data is corrected so as to obtain standard photoelectric conversion characteristics by using correction data for each pixel, and fixed pattern noise is suppressed.
  • Patent Document 2 since the photoelectric conversion characteristics of all pixels measured in advance are stored as correction data, there is a problem that the amount of correction data is enormous. In addition, since the photoelectric conversion characteristics change depending on the temperature, there is a problem that the amount of data further increases when correction data for each temperature is stored in the storage unit.
  • Patent Document 3 describes a method for reducing random noise during AD conversion of a pixel signal, but does not describe white reset.
  • An object of the present invention is to provide a solid-state imaging device in which random noise included in a white reset output is reduced without increasing the circuit scale.
  • a solid-state imaging device includes a plurality of pixel units including a photoelectric conversion element, a transfer transistor that flows a photocurrent so that charges accumulated in the photoelectric conversion element have a linear logarithmic characteristic, and a floating diffusion.
  • a pixel control unit that controls the pixel unit; and an image processing unit that processes a signal output from the pixel unit, wherein the pixel control unit applies an intermediate level voltage to a control terminal of the transfer transistor.
  • an exposure process for exposing the subject to the photoelectric conversion element a charge accumulated in the photoelectric conversion element by the exposure process is converted into a voltage by the transfer transistor being transferred to the floating diffusion, and the first signal is After the acquisition of the first acquisition process and the first acquisition process, the photoelectric conversion element is reset to a zero bias state. After the light reset process and the white reset process are completed, a leak process is performed in which the intermediate level voltage is applied to the control terminal of the transfer transistor to leak the charge accumulated in the photoelectric conversion element for a predetermined first leak period.
  • the pixel unit performs a second acquisition process in which the charge remaining in the photoelectric conversion element is converted into a voltage by the transfer transistor being transferred to the floating diffusion and a second signal is acquired.
  • the image processing unit applies the second signal to the second signal from a plurality of correction data in which a representative value of the second signal and a representative characteristic that is a photoelectric conversion characteristic corresponding to the representative value are associated with each other.
  • a correction unit is provided that selects one correction data based on the correction data and corrects the first signal using the selected correction data.
  • FIG. 1 is an overall configuration diagram of a solid-state imaging device according to Embodiment 1 of the present invention. It is a block diagram of the image pick-up element shown in FIG.
  • FIG. 3 is a circuit diagram of a pixel unit shown in FIG. 2.
  • 3 is a timing chart of the pixel portion shown in FIG.
  • FIG. 5 is a potential diagram of a pixel portion at times t0 to t9 in FIG. It is a graph which shows the photoelectric conversion characteristic of a LinLog signal and a WR signal.
  • 3 is a timing chart of the solid-state imaging device according to Embodiment 1 of the present invention. It is a circuit diagram of ADC and a ramp waveform generation circuit. It is a wave form diagram of a VIDEO signal and a ramp waveform.
  • (A) is a graph showing the difference between DLTa of LUTa + 1 and DLTa of LUTa
  • (B) is a graph showing DIp (Ip)
  • (C) is a graph showing LSCa (IP). It is a sequence diagram of the pixel part in Embodiment 5 of this invention.
  • FIG. 1 is an overall configuration diagram of a solid-state imaging device according to Embodiment 1 of the present invention.
  • the solid-state imaging device includes an imaging element control unit 100, an imaging element 200, and an image processing unit 300.
  • the image sensor control unit 100 outputs a system clock signal (sysclk) and a register control signal to the image sensor 200.
  • the system clock signal is a clock signal for synchronizing circuit elements constituting the solid-state imaging device.
  • the register control signal is a signal for writing a register value for defining the waveform of the pixel control signal for controlling the pixel unit 220 (see FIG. 2) to a register included in the image sensor 200.
  • the image sensor control unit 100 outputs transfer gate voltage data to the image processing unit 300.
  • the transfer gate voltage data is data for defining an intermediate level voltage applied to the gate of the transfer transistor, and is used when generating correction data, which will be described later, and used as an index when specifying the correction data.
  • the image sensor 200 outputs a Linlog signal, a WR signal, and temperature data, which are image signals, to the image processing unit 300.
  • the LinLog signal is data indicating luminance information of the subject, and is data to be corrected.
  • the WR signal is a white reset signal described later, and is used as an index when specifying correction data.
  • the temperature data is data indicating the environmental temperature of the image sensor 200, and is used as an index when specifying correction data.
  • the image processing unit 300 includes a correction data storage unit 301 and a correction unit 302, and performs image processing on a signal output from the image sensor 200.
  • the correction data storage unit 301 stores correction data in advance. Details of the correction data and the correction unit 302 will be described later.
  • FIG. 2 is a configuration diagram of the image sensor 200 shown in FIG.
  • the image sensor 200 includes a row decoder 210 (an example of a pixel control unit), a pixel unit 220, a timing control unit (an example of a pixel control unit) 230, a column ADC array 240, a column decoder 250, a ramp waveform generation circuit 260, and a temperature measurement unit. 270 and an output circuit 280.
  • a plurality of pixel units 220 are arranged in a matrix form of n rows ⁇ m columns to form a pixel array unit.
  • the timing control unit 230 includes a PLL, a timing generator (TG), and a register.
  • the PLL multiplies the system clock signal in order to set the frequency of the system clock signal (sysclk) to a frequency suitable for the operation of the image sensor 200.
  • the register value that defines the waveform of the pixel control signal is written by the register control signal.
  • the TG generates a horizontal synchronization signal and a vertical synchronization signal from the system clock signal multiplied by the PLL and supplies the horizontal synchronization signal and the vertical synchronization signal to the row decoder 210 and the column decoder 250. Further, the TG supplies the register value to the row decoder 210 and causes the row decoder 210 to output a pixel control signal.
  • the row decoder 210 cyclically selects each row of the pixel array unit, vertically scans the pixel array unit, and outputs a signal from the pixel unit 220 of the selected row.
  • the row decoder 210 includes a vertical readout scanning circuit 211, a transfer gate scanning circuit 212, a reset scanning circuit 213, and a reset drain scanning circuit 214.
  • the row decoder 210 outputs a pixel control signal to each row of the pixel array unit.
  • the pixel control signal includes a reset drain signal (hereinafter described as ⁇ RDi), a reset signal (hereinafter described as ⁇ RSTi), a transfer gate signal (hereinafter described as ⁇ TXi), and a row selection signal (hereinafter referred to as ⁇ RSTi).
  • ⁇ RDi reset drain signal
  • ⁇ RSTi reset signal
  • ⁇ TXi transfer gate signal
  • ⁇ RSTi row selection signal
  • ⁇ VSEN, ⁇ TX, ⁇ RST, and ⁇ RD are output from the vertical readout scanning circuit 211, the transfer gate scanning circuit 212, the reset scanning circuit 213, and the reset drain scanning circuit 214, respectively.
  • the column ADC array 240 includes m AD conversion units (ADC) 241 (an example of a reading unit) provided corresponding to each column of the pixel array unit.
  • the ADC 241 performs AD conversion on the signal output from the pixel unit 220.
  • the ADC 241 includes a memory that temporarily holds a digital signal obtained by AD conversion.
  • the ADC 241 is a single slope type ADC that AD-converts a signal by comparing an analog signal output from the pixel unit 220 with a ramp waveform output from the ramp waveform generation circuit 260.
  • the ADC 241 may employ an ADC other than the single slope type (for example, a double slope type ADC).
  • the column decoder 250 sequentially selects the ADCs 241 in each column according to the horizontal synchronization signal output from the timing control unit 230, horizontally scans the column ADC array 240, and sequentially outputs digital signals from the ADC 241.
  • the ramp waveform generation circuit 260 generates a ramp waveform used when the ADC 241 performs AD conversion, and outputs the ramp waveform to the ADC 241.
  • the temperature measurement unit 270 includes, for example, a thermistor and a temperature sensor including an AD conversion circuit that AD converts a voltage value of the thermistor, measures the environmental temperature of the image sensor 200, and supplies digital temperature data to the image processing unit 300. Output.
  • the output circuit 280 shapes the digital signal waveform sequentially output from the column ADC array 240 and outputs the waveform to the image processing unit 300.
  • the output circuit 280 includes a sense amplifier and an LVDS serializer.
  • the sense amplifier shapes a digital signal waveform output from the column ADC array 240.
  • the LVDS serializer converts a parallel signal whose waveform is shaped by the sense amplifier into a serial signal and outputs the serial signal to the image processing unit 300.
  • FIG. 3 is a circuit diagram of the pixel unit 220 shown in FIG.
  • the pixel portion 220 includes a photoelectric conversion element (hereinafter referred to as PD) and four NMOS transistors (hereinafter referred to as TR1 to TR4).
  • PD has an anode connected to the ground (hereinafter referred to as GND) and a cathode connected to the source of TR1.
  • GND ground
  • the PD generates a photocurrent Ip proportional to the luminance of the subject, and accumulates charges corresponding to the photocurrent Ip as parasitic capacitance.
  • TR1 allows photocurrent to flow so that the charge accumulated in the PD has a linear logarithmic characteristic.
  • TR1 is a transfer transistor that transfers charges accumulated in the PD to the FD.
  • the source of TR1 is connected to PD, the drain of TR1 is connected to TR2, and the gate of TR1 is connected to the ⁇ TX signal line through which ⁇ TX flows, and is controlled by the transfer gate scanning circuit 212.
  • ⁇ TX can be set to three levels: High, Middle, and Low. When High, TR1 is in the ON state, when Low, when OFF, and when photocurrent Ip is large, TR1 is in the subthreshold state where linear logarithmic conversion is performed. Become.
  • TR2 is a reset transistor that resets FD (floating diffusion).
  • the drain of TR2 is connected to a ⁇ RD control line through which ⁇ RD flows, the source of TR2 is connected to TR1 through FD, and the gate of TR2 is connected to a signal line through which ⁇ RST flows.
  • TR2 is controlled by supplying ⁇ RST to the gate from the reset drain scanning circuit 214 (see FIG. 2).
  • ⁇ RST can be set to two states, High and Low. When High, TR2 is turned ON, and the voltage set by ⁇ RD is set as the source voltage of TR2. When ⁇ RST is Low, TR2 is OFF.
  • the nodes between TR1 and TR2 form an FD.
  • FD charges accumulated in the PD by TR1 are transferred, and a voltage signal corresponding to the photocurrent Ip is generated.
  • TR3 The drain of TR3 is connected to the PVDD line, and the source of TR3 is connected to TR4.
  • the gate of TR3 is connected to the FD.
  • TR3 When a voltage signal corresponding to the photocurrent Ip is generated in the FD, TR3 operates as a source follower and amplifies the signal.
  • PVDD is supplied by, for example, a power supply circuit (not shown) and set to a positive potential.
  • TR4 The drain of TR4 is connected to TR3, the source of TR4 is connected to the VIDEO line, and the gate of TR4 is connected to the ⁇ VSEN control line.
  • TR4 is controlled by supplying ⁇ VSEN to the gate from the vertical readout scanning circuit 211 (see FIG. 2).
  • ⁇ VSEN When ⁇ VSEN is High, TR4 is turned ON, and the signal amplified by TR3 flows through the VIDEO line.
  • TR4 is OFF. That is, TR4 functions as a row selection transistor.
  • FIG. 4 is a timing chart of the pixel unit 220 shown in FIG.
  • TR1 will be described as a transfer gate and TR2 as a reset gate.
  • ⁇ TX Middle
  • ⁇ RST High
  • the photocurrent Ip generated in the PD according to the luminance of the subject flows through the transfer gate and the reset gate and flows into the ⁇ RD control line, and the charge obtained by linear logarithmic conversion is accumulated in the parasitic capacitance of the PD.
  • the Ref signal includes information on KTC noise, which is random noise generated when the reset gate is closed.
  • ⁇ TX High and the transfer gate is turned on, the charge accumulated in the PD is transferred to the FD, and the transfer process is executed.
  • the FD a voltage signal corresponding to the charge obtained by adding the charge transferred from the PD to the charge of the Ref signal output at time t1 appears.
  • ⁇ TX Low
  • the transfer gate is turned off, and a signal signal output process in which the voltage signal of FD is output as a sig signal (signal signal) is executed.
  • the sig signal output at time t3 is subjected to a difference process that takes a difference from the ref signal by the ADC 241 to obtain a LinLog signal (first signal).
  • LinLog signal first signal
  • the period from time t1 to time t3 corresponds to the execution period of the first acquisition process in which the charge accumulated in the PD by the exposure process is converted into a voltage by the transfer gate being transferred to the FD and the first signal is acquired. To do.
  • ⁇ TX Middle
  • ⁇ RD High
  • the charge accumulated in the PD is leaked to the drain side of the reset gate, and leak processing is executed.
  • the transfer gate shifts from the ON state to the subthreshold state in a very short time. For this reason, after the end of this leak period, a charge of a level corresponding to the potential level of the transfer gate remains in the PD.
  • the leak period is, for example, about several us to several hundred us.
  • the reference signal output process, the transfer process, and the signal signal output process are performed, respectively, similarly to the times t1 to t3.
  • the sig signal output at time t8 is differenced from the ref signal by the ADC 241 and random noise included in the sig signal is canceled by the ref signal, thereby obtaining a WR signal (second signal).
  • the period from time t6 to time t8 corresponds to the execution period of the second acquisition process in which the charge remaining in the PD is converted into a voltage by the transfer gate being transferred to the FD after the leak process is completed and the second signal is acquired. To do.
  • FIG. 5 is a potential diagram of the pixel unit 220 at times t0 to t9 in FIG.
  • the transfer gate since the transfer gate is in the subthreshold state, the charge accumulated in the PD that exceeds the potential level LV1 of the transfer gate leaks from the PD to the FD. Accordingly, the PD accumulates charges while leaking charges exceeding the potential level LV1, and accumulates charges with a linear logarithmic characteristic.
  • the transfer gate and the reset gate are closed, and the voltage signal of FD is output as the ref signal.
  • the transfer gate is opened, and the charge accumulated in the PD is transferred to the FD.
  • the transfer gate is closed and the voltage signal of FD is output as the sig signal.
  • the transfer gate reset gate is opened and white reset is performed, and the charge of the PD is full.
  • the transfer gate is in the subthreshold state, and the charge accumulated in the PD leaks to the FD. At this time, since the FD is reset by the reset gate, the charge leaked to the FD is discharged to the outside.
  • the transfer gate is opened, and the charge remaining in the PD is transferred to the FD.
  • the transfer gate is closed, and the charge transferred to the FD is output as a sig signal.
  • the transfer gate is set to the sub-threshold state, the FD is reset by the reset gate, and exposure processing is performed.
  • FIG. 6 is a graph showing photoelectric conversion characteristics of the LinLog signal and the WR signal.
  • the vertical axis indicates digital output values of the LinLog signal and the WR signal
  • the horizontal axis indicates the photocurrent Ip having a value corresponding to the luminance of the subject on a logarithmic scale.
  • the photoelectric conversion characteristic of the LinLog signal has a linear characteristic in which the output value increases linearly as the luminance increases on the low luminance side. It can also be seen that the high luminance side has a logarithmic characteristic in which the output value increases logarithmically as the luminance increases.
  • the photoelectric conversion characteristic of the WR signal has a flat characteristic without depending on the luminance of the subject.
  • the WR signal can be obtained by providing a leak period after white reset and leaking charge from the PD.
  • the leakage current flowing in the leakage period decreases with time, but when the leakage period is short, the leakage current is very large and is not affected by the luminance of the subject. Therefore, the photoelectric conversion characteristic of the WR signal shows a flat characteristic.
  • the photoelectric conversion characteristic of the WR signal represents information on the high luminance side of the LinLog signal. This is because the leak period set when obtaining the WR signal is short. That is, when the leak period is short, the leak current is very large, and the value indicates the value of the photocurrent Ip that flows through the PD when a high-luminance subject is exposed.
  • the photoelectric conversion characteristics of the LinLog signal vary in the logarithmic region on the higher luminance side than the inflection point because the inflection point, which is a switching point between the linear characteristic and the logarithmic characteristic, is different for each pixel.
  • the photoelectric conversion characteristic of the WR signal also has a flat characteristic as a whole, but varies from pixel to pixel.
  • both the LinLog signal and the WR signal are signals obtained by setting the transfer gate to the subthreshold state
  • the LinLog signal includes information on the potential level LV1 of the transfer gate that causes the variation of the inflection point. Therefore, the WR signal variation is highly correlated with the inflection point variation. Therefore, inflection point information inherent in the pixel unit 220 can be obtained from the WR signal.
  • a plurality of correction data in which a representative value of a WR signal and a representative characteristic that is a photoelectric conversion characteristic corresponding to the representative value are associated with each other are stored in advance in the correction data storage unit 301.
  • the representative characteristic is data obtained by grouping photoelectric conversion characteristics into a plurality of groups according to information on the inflection point and averaging the photoelectric conversion characteristics belonging to each group.
  • the LinLog signal is corrected using correction data associated with a representative value whose value is close to that of the WR signal obtained at the time of imaging. Therefore, in the present embodiment, the WR signal is used as an index for selecting which correction data to use from among a plurality of correction data for correcting the LinLog signal.
  • FIG. 7 is a timing chart of the solid-state imaging device according to Embodiment 1 of the present invention.
  • the pixel units 220 are arranged in N rows.
  • the output of the LinLog signal and the output of the WR signal can be performed during one horizontal period. Therefore, in the present embodiment, the output of the LinLog signal, the white reset process and the leak process, and the output of the WR signal are performed within one horizontal period.
  • one horizontal period indicates an output cycle of the horizontal synchronization signal (HSYNC).
  • N rows of data can be obtained.
  • the row decoder 210 selects the 0th row and supplies the 0th row to the pixel unit 220. , LinLog signal output, white reset processing, leak processing, and WR signal output are sequentially performed. Thereby, the LinLog signal and the WR signal of the 0th row are obtained.
  • the row decoder 210 selects the first row, and causes the pixel unit 220 of the 0th row to sequentially perform the same processing as the pixel unit 220 of the 0th row. Thereby, the LinLog signal and the WR signal in the first row are obtained.
  • the row decoder 210 sequentially selects each row and outputs a LinLog signal, white reset processing, and leak processing to the pixel unit 220 of the selected row. And the WR signal are sequentially output. Thereby, the LinLog signal and WR signal of all the pixels are obtained.
  • the subsequent image processing unit 300 can correct the LinLog signal based on the WR signal without providing a frame memory. Therefore, the circuit scale of the image processing unit 300 can be reduced.
  • FIG. 8 is a circuit diagram of the ADC 241 and the ramp waveform generation circuit 260.
  • the ADC 241 includes a sample hold (S / H) 801, a comparator 802, an AND gate 803, and an output counter 804.
  • the ramp waveform generation circuit 260 includes a DAC (digital-analog converter) 811, a counter 812, and an initial value setting unit 813.
  • the S / H 801 holds a VIDEO signal input from the VIDEO line.
  • the VIDEO signal corresponds to the ref signal and the sig signal.
  • the comparator 802 sets the output signal to High.
  • the comparator 802 compares the VIDEO signal held in the S / H 802 with the ramp waveform output from the DAC 811, and changes the output signal from High to Low when the ramp waveform crosses the VIDEO signal.
  • the AND gate 803 outputs a clock signal (CLK) to the output counter 804 when a high output signal is input from the comparator 802, and outputs a clock signal to the output counter 804 when a low signal is input from the comparator 802. Is stopped and the clock signal is gated.
  • CLK clock signal
  • the output counter 804 starts the count operation when the input of the ramp waveform is started and the clock signal is output from the AND gate 803. Further, when the ramp waveform crosses the VIDEO signal and the output of the clock signal from the AND gate 803 is stopped, the output counter 804 stops the count operation and holds the count value. The count value at this time becomes the digital value of the VIDEO signal. That is, in the period from when the input of the ramp waveform is started until the ramp waveform and the VIDEO signal cross, the count number of the clock signal counted by the output counter 804 becomes the digital value of the VIDEO signal.
  • the counter 812 detects the output timing of the ramp waveform by counting the number of clocks of CLK, and instructs the DAC 811 to start outputting the ramp waveform. In addition, when the counter 812 determines the output timing of the ramp waveform for AD conversion of the sig signal of the WR signal, the counter 812 outputs the initial value set by the initial value setting unit 813 to the DAC 811, and responds to the initial value for the ramp waveform. Have an offset.
  • the initial value setting unit 813 holds a preset initial value.
  • FIG. 9 is a waveform diagram of the VIDEO signal and the ramp waveform.
  • a ramp waveform for AD conversion of the ref signal is first output from the DAC 811, and then the AD conversion of the sig signal is performed.
  • a ramp waveform is output from the DAC 811.
  • the ramp waveforms of the ref signal and the sig signal both have the same slope and have a waveform in which the voltage value decreases linearly.
  • the ref signal has a higher voltage than the sig signal and the amount of change in the voltage is small, so that the AD conversion period is short. Therefore, the amplitude AM1 of the ramp waveform for AD converting the ref signal is set smaller than the amplitude AM2 of the ramp waveform for AD converting the sig signal. Thereby, the AD conversion period of the ref signal can be shortened.
  • the WR signal output period When the LinLog signal output period ends, the white reset process and the leak process are performed, and then the WR signal output period is started. Since the WR signal output period also includes two output periods of the ref signal and the sig signal, a ramp waveform for AD conversion of the ref signal is first output from the DAC 811, and then a ramp for outputting the sig signal A waveform is output from the DAC 811.
  • Sig signal of WR signal contains inflection point variation information, but voltage is within a certain range. This range is smaller than the range of the sig signal of the LinLog signal, and the voltage is also low. Therefore, if the sig signal of the WR signal is AD-converted using the same ramp waveform as the sig signal of the LinLog signal, the AD conversion period becomes long.
  • an offset OS corresponding to the initial value is set for the sig signal of the WR signal.
  • the AD conversion period of the sig signal of the WR signal is shortened.
  • the offset OS for example, a voltage value in a higher voltage range that can be taken by the sig signal of the WR signal may be employed.
  • FIG. 10 is a flowchart showing a process for generating correction data stored in the correction data storage unit 301. It is desirable to generate correction data for each manufactured solid-state imaging device individually because high-precision correction data is generated. However, in order to simplify the correction data generation process, several representative solid-state imaging devices are used. The correction data may be generated by the device, and the generated correction data may be used for the remaining solid-state imaging devices. Therefore, in this embodiment, one correction data is generated for each lot, and the correction data is used in a plurality of solid-state imaging devices in the same lot.
  • the correction data will be described as being generated by a solid-state imaging device in which the correction data generation unit 303 is mounted on the image processing unit 300. However, this is only an example, and the correction data generation unit 303 may be provided in each solid-state imaging device, and the correction data generation unit 303 may generate correction data for each solid-state imaging device.
  • correction data that takes into account the variation of the inflection point that varies according to the environmental temperature and the transfer gate voltage is generated. Therefore, first, the correction data generation unit 303 sets the environmental temperature T (S1).
  • the environmental temperature to be set for example, the temperature of the image sensor 200 is adjusted by heating or cooling the image sensor 200, the adjusted temperature is measured by the temperature measurement unit 270, and the obtained temperature data A value may be adopted.
  • the correction data generation unit sets the voltage V of the transfer gate (S2).
  • the brightness of a light source that is disposed in front of the image sensor 200 and irradiates the image sensor 200 with light is set (S3).
  • the image sensor 200 is caused to measure the LinLog signal (S4).
  • S3 and S4 are repeatedly executed to change the luminance of the light source with a predetermined resolution, and each time the luminance of the light source is changed, the photocurrent Ip having a value corresponding to the luminance and the output value of the LinLog signal
  • the correction data generation unit 303 records data in which the photocurrent Ip and the LinLog signal are associated with each other. In this way, the photoelectric conversion characteristic of the LinLog signal (hereinafter referred to as the LinLog characteristic) is measured.
  • the correction data generation unit 303 groups the pixel units 220 on the basis of the measured inflection point of LinLog characteristics (an example of information on the inflection point) (S7).
  • LinLog characteristics are grouped into, for example, 256 groups. However, this number is only an example, and another number may be used for grouping.
  • the correction data generation unit 303 divides the width between the maximum value and the minimum value of the inflection points into 256 equal parts to create 256 classes, and to which class the inflection points of each pixel unit 220 belong. And assigning each pixel unit 220 to the class to which the inflection point belongs, thereby dividing each pixel unit 220 into 256 groups.
  • the correction data generation unit 303 calculates an average characteristic of the LinLog characteristic for each group (S8).
  • 256 LUTs (LUT0 (V, T) to LUT255 (V, T)) of group 0 to group 255 are generated by associating the average characteristic, the environmental temperature T, and the transfer gate voltage V (S9).
  • This LUT is an LUT having an output value of the LinLog characteristic as an LUT input value and an optical current Ip as an LUT output value. That is, the LUT is a LUT that outputs a digital value of the photocurrent Ip according to the output value of the LinLog signal.
  • the correction data generation unit 303 calculates an average value WR (V, T) of WR characteristics for each group (S10). Finally, the correction data generation unit 303 generates a correspondence table in which the average value WR (V, T) and the LUT number (0 to 256) are associated (S11). Thus, when the WR value of a certain pixel is obtained, the LUT associated with the WR value closest to the WR value can be specified, and the LUT used for correcting the LinLog signal can be determined using the WR value as an index. Can be identified.
  • inflection points are adopted as information on inflection points, and grouping is performed based on the inflection points.
  • the WR values and the inflection points are highly correlated, the WR values are inflected. It may be adopted as information about points and grouped based on the WR value. In this case, it is not necessary to obtain an inflection point from the LinLog characteristics, and the correction data generation process can be simplified.
  • FIG. 11 is a graph for explaining the principle of correction.
  • the first quadrant in the upper right indicates the LinLog characteristic and the WR characteristic for each group, the horizontal axis indicates the photocurrent Ip on a logarithmic scale, and the vertical axis indicates the output value of the image sensor 200.
  • the second quadrant in the upper left indicates the input / output relationship of the LUT for each group, the vertical axis indicates the LinLog signal input to the LUT, and the horizontal axis indicates the corrected LinLog signal (corresponding to the photocurrent) output from the LUT. Therefore, it is expressed as Ip ′).
  • the lower left third quadrant shows a graph G3 representing the photoelectric conversion characteristics of the corrected LinLog signal
  • the horizontal axis shows Ip ′ output from the LUT
  • the vertical axis shows the photocurrent Ip on a logarithmic scale.
  • the lower right fourth quadrant shows a graph G4 obtained by turning the graph G3 around the vertical axis.
  • the LinLog characteristic and the WR characteristic are associated with the group number of the group to which the self belongs. At the time of imaging, an output value at one point on the LinLog characteristic and an output value at one point on the WR characteristic are output from the image sensor 200.
  • the correction unit 302 acquires a WR signal of a target pixel from the image sensor 200.
  • the correction unit 302 acquires the environmental temperature T measured by the temperature measurement unit 270.
  • the correction unit 302 acquires the transfer gate voltage V from the register of the timing control unit 230.
  • the correction unit 302 refers to the correspondence table and specifies the group number of the group to which the target pixel belongs from the acquired WR signal, the environmental temperature T, and the voltage V.
  • the correction unit 302 firstly includes a group in which the average value of the WR signal closest to the WR signal of the target pixel is associated with each of the environmental temperature T and the transfer gate voltage V among the groups described in the correspondence table. Is identified. Next, the correction unit 302 selects a group associated with the environmental temperature T and the voltage V closest to the acquired environmental temperature T and the voltage V among the groups specified for the environmental temperature T and the voltage V. Identify as a group to which it belongs.
  • the correction unit 302 inputs the LinLog signal of the target pixel to the LUT associated with the group number, and obtains a corrected LinLog signal.
  • the correction unit 302 inputs LinLog_x0 to LUT0, and obtains Ip_x ′ that is a corrected LinLog signal corresponding to LinLog_x0. Since Ip_x ′ corresponds to Ip_x, when it is plotted in the third quadrant, it rides on the point P_x on the graph G3.
  • the correction unit 302 inputs LinLog_x1 to the LUT1, and obtains Ip_x ′ that is a corrected LinLog signal corresponding to LinLog_x1. Since Ip_x ′ corresponds to Ip_x, when it is plotted in the third quadrant, it rides on P_x on the graph G3.
  • the LinLog signal of all the pixel units 220 is corrected by the correction unit 302 so as to be on the graph G3, and the entire range is converted into data showing logarithmic characteristics with respect to the photocurrent Ip, and at the same time, the inflection point. Variation is also eliminated.
  • Embodiment 2 In Embodiment 1, the example which correct
  • the coefficient of the function defining the LinLog characteristic is stored in advance as correction data instead of the LUT indicating the input / output relationship of the photoelectric conversion characteristic. It is characterized by.
  • the correction data generation process of the correction data generation unit 303 is shown below.
  • the LinLog characteristic is divided into a linear characteristic area and a logarithmic characteristic area, and both areas are represented by the following equations.
  • the correction data generation unit 303 can obtain the coefficients a, b, c, and d of each group from the LinLog characteristics obtained in the flowchart shown in FIG. Specifically, the coefficients a, b, c, and d can be obtained by appropriately selecting four points of average characteristics in each group obtained in S8 and substituting them into the equations (1) and (2).
  • the correction data generation unit 303 generates a correspondence table that associates the average value WR (V, T) of the WR signal obtained for each group in S10 with the group numbers of the coefficients a to d calculated for each group.
  • the correction unit 302 refers to the correspondence table and identifies the group number associated with the average value WR (V, T) with the closest WR value.
  • the coefficients a to d can be specified from the specified group number.
  • FIG. 12 is a graph showing the LinLog characteristics, where y is the output value of the image sensor 200, and the horizontal axis is x, which is the luminance of the incident surface of the pixel unit 220, on a logarithmic scale.
  • the correction data generation unit 303 may obtain the value of (xth, yth) approximately using a numerical solution method such as a least square method. Then, the correction data generation unit 303 calculates (xth, yth) for each group using the values of the coefficients a to d of each group, and associates the calculated yth with the group number in the correction data storage unit 301. Store in advance. As a result, correction data in which the coefficients a to d, yth and the average value WR (V, T) of the WR signal are associated is generated. The correction data generation unit 303 stores the generated correction data in the correction data storage unit 301 in advance.
  • the output value of the LinLog signal measured at the time of imaging is LL, and the corrected value of the LinLog signal is yA.
  • the corrected output value LLA that is, the correction equation for the linear characteristic region is expressed by the following equation.
  • Linear characteristic region correction formula: LLA ln ((LL ⁇ b) / a) (4)
  • the corrected output value LLA that is, the logarithmic characteristic region correction equation is expressed by the following equation.
  • Logarithmic characteristic region correction formula: LLA (LL ⁇ d) / c (5)
  • the processing of the correction unit 302 is summarized as follows. First, the correction unit 302 obtains a WR signal and a LinLog signal from the target pixel. Next, the correction unit 302 refers to the correspondence table from the WR signal, identifies the group number of the target pixel, and identifies the coefficients a to d and yth associated with the identified group number.
  • the correction unit 302 determines whether the output value LL of the LinLog signal is greater than or less than yth.
  • the output value LL is LL ⁇ yth
  • the output value LL and the coefficients a and b are substituted into the equation (4) to obtain the corrected output value LLA.
  • the corrected output value LLA indicating the logarithmic characteristic can be obtained in the entire area of the LinLog characteristic.
  • the coefficients a to d and yth of each group are associated with the WR signal and stored in advance in the correction data storage unit 301, the LinLog signal can be corrected. As a result, the consumption of memory resources can be further reduced.
  • FIG. 13 is a graph of the LinLog characteristic showing the variation in the slope of the logarithmic characteristic region, the vertical axis shows the output value of the LinLog characteristic, and the horizontal axis shows the logarithm of the photocurrent Ip according to the luminance of the imaging surface of the pixel unit 220. Shown in scale.
  • Embodiment 3 is characterized in that at the time of imaging, two types of WR signals having different leak periods are acquired, and the variation in the slope of the logarithmic characteristic region is corrected using these WR signals.
  • the WR signal has a characteristic that the output value changes depending on the length of the leak period.
  • the leak period is short, the output value is large, and when the leak period is long, the output value is small.
  • the WR1 signal with a short leak period and the WR2 signal with a long leak period are acquired, the WR1 signal reproduces the case where a large photocurrent Ip flows with a high luminance, and the WR2 signal has a low photocurrent Ip with a low luminance. The case has been reproduced.
  • the relationship between the leak period and the photocurrent Ip can be measured in advance.
  • FIG. 14 is a graph showing photoelectric conversion characteristics of the WR1 signal (hereinafter referred to as WR1 characteristics), photoelectric conversion characteristics of the WR2 signal (hereinafter referred to as WR2 characteristics), and LinLog characteristics. Indicates the output value from the ADC 271, and the horizontal axis indicates the photocurrent Ip flowing according to the luminance of the imaging surface of the pixel unit 220 on a logarithmic scale.
  • the dotted LinLog characteristic is the LinLog characteristic indicated by one LUTa (a indicates the LUT number and has a value of 0 to 255) selected by the correction unit 302 from the WR1 values during imaging.
  • the solid line LinLog characteristic is the LinLog characteristic of the target pixel.
  • the WR1 signal has an output value in which the leak period is set so that the leak current becomes Ip1.
  • Ip1 has a predetermined value and indicates a photocurrent at the intersection P_max between the LinLog characteristic and the standard characteristic.
  • the WR2 signal has an output value in which the leak period is set so that the leak current becomes Ip2 ( ⁇ Ip1).
  • Ip2 is a photocurrent Ip at a certain point in the logarithmic characteristic region, and has a predetermined value.
  • Ip1 and Ip2 are stored in the correction data storage unit 301 in advance.
  • the correction unit 302 acquires the WR1 signal (an example of the second signal) from the target pixel in the same manner as in the first embodiment, and uses the acquired WR1 signal as LUTa (Ip) (here, the LUT is converted into the photocurrent Ip). Select as a function).
  • the correction unit 302 acquires a WR2 signal (an example of a third signal) from the target pixel.
  • the correction unit 302 calculates a target pixel inclination SLx that is an inclination of the LinLog characteristic of the target pixel in the logarithmic characteristic region from the acquired WR1 and WR2 signals using the following formula.
  • the correction unit 302 calculates a standard gradient SLa_n that is the gradient of the LinLog characteristic of the LUTa in the logarithmic characteristic region using the following equation.
  • SLa_n (LUTa (Ip1) ⁇ LUTa (Ip2)) / (Ip1-Ip2)
  • SLa_n may be calculated in advance by the correction data generation unit 303 at the time of generation of correction data.
  • the correction unit 302 may read the pre-calculated slope SLa_n from the correction data storage unit 301. .
  • LUTa (Ip1) and LUTa (Ip2) indicate values corresponding to Ip1 and Ip2 in the photoelectric conversion characteristics of the selected LUTa.
  • the target pixel inclination SLx is larger than the standard inclination SLa_n, it can be seen that the target pixel inclination is steep. Conversely, when the target pixel inclination SLx is smaller than the standard inclination SLa_n, it can be seen that the target pixel inclination is gentle.
  • FIG. 14 illustrates an example in which the inclination SLx of the target pixel is gentler than the inclination SLa_n.
  • the correction unit 302 obtains a correction amount of the inclination deviation from the standard inclination SLa_n and the target pixel inclination SL_x.
  • the correction unit 302 first determines whether the LinLog signal of the target pixel belongs to the linear characteristic region or the logarithmic characteristic region.
  • the photocurrent Ipb_a at the inflection point is specified.
  • the photocurrent Ipb_a is calculated in advance for each group by the correction data generation unit 303 and stored in the correction data storage unit 301. That is, the correction unit 302 corrects the LinLog signal by inputting the LinLog signal to the LUTa selected from the WR1 signal without correcting the correction data if the LinLog signal of the target pixel is smaller than the photocurrent Ipb_a.
  • the correction unit 302 applies a process for correcting correction data, which will be described later, and corrects the LinLog signal using the corrected correction data.
  • Ipb_a When the brightness of the inflection point of LUTa is low, Ipb_a is set to the dark side, and when the brightness of the inflection point is high, Ipb_a is set to the bright side.
  • the correction unit 302 calculates DLa (Ip) represented by the following equation so that the correction amount of the inclination shift has a value only in the logarithmic characteristic region.
  • DLa (Ip) is represented by the graph in FIG. In FIG. 15A, the vertical axis represents DLa (Ip), and the horizontal axis represents the photocurrent Ip on a logarithmic scale.
  • the correction unit 302 calculates a coefficient DIp (Ip) using the following equation in order to calculate the correction amount of the inclination deviation of the inclination according to the photocurrent Ip.
  • DIp (Ip) is represented by the graph in FIG. 15B, the vertical axis indicates DIp (Ip), and the horizontal axis indicates the photocurrent Ip on a logarithmic scale.
  • DIp (Ip) M ⁇ (log (Ip1) ⁇ log (Ip))
  • the coefficient M is a constant, and an arbitrary value is set according to the strength of correction.
  • the intersection between the LinLog characteristic of the target pixel and the LinLog characteristic of the LUTa may vary from P_max.
  • a value capable of minimizing the tilt deviation may be obtained in advance based on this variation and set as the value of the coefficient M.
  • LSCa (Ip) is represented by the graph in FIG. In FIG. 15C, the vertical axis indicates LSCa (Ip), and the horizontal axis indicates the photocurrent Ip on a logarithmic scale. As shown in FIG. 15C, a correction amount LSCa (Ip) in which a region having a lower luminance than Ipb_a is removed from DIp by DLa ⁇ DIp and the slope of DIp is adjusted by coefficient M is obtained. .
  • the correction unit 302 adds the correction amount LSCa (Ip) of the inclination deviation to the LUTa (Ip). Thereby, LUTa (Ip) is corrected so that the standard inclination SLa_n matches the target pixel inclination SLx.
  • the correction unit 302 inputs the output value LL of the LinLog signal of the target pixel to the corrected LUTa, and obtains the output value LLA.
  • the method of the third embodiment works well when the condition that the LinLog characteristic to be corrected is substantially the same as the LinLog characteristic of LUTa is satisfied.
  • This condition corresponds to a case where the variation in inclination between the target pixel inclination SLx and the standard inclination SL_a is very small.
  • the variation of the inflection point was large, and the inclination of the logarithmic characteristic region was slightly different depending on the variation of the inflection point. Specifically, a pixel with a high luminance value at the inflection point has a steep slope in the logarithmic characteristic region, and a pixel with a low luminance value at the inflection point tends to have a gentle inclination in the logarithmic characteristic region.
  • the LUTa (Ip) is corrected from the correction amount LSCa (Ip) of the inclination deviation between the target pixel inclination SLx and the standard inclination SLa_n.
  • the LinLog signal is corrected using the corrected LUTa (Ip). Therefore, in addition to the variation of the inflection point of the LinLog characteristic, the variation of the slope of the linear logarithmic characteristic region can also be corrected.
  • the fourth embodiment is characterized in that the LUT is corrected so that a non-linear region near the inflection point is reproduced without clearly distinguishing the linear characteristic region and the logarithmic characteristic region.
  • the correction unit 302 calculates the target pixel inclination SLx and the standard inclination SLa_n. Next, the correction unit 302 identifies the LUTa (Ip) selected from the WR1 signal and the LUTa + 1 (Ip) adjacent thereto.
  • FIG. 16 is a graph showing the LinLog characteristics of LUTa and LUTa + 1.
  • the vertical axis indicates the output value of the LinLog signal
  • the horizontal axis indicates the photocurrent Ip corresponding to the luminance on a logarithmic scale.
  • LUTa + 1 adjacent to LUTa for example, LUTa + 1 having the next highest inflection point luminance value with respect to LUTa is adopted.
  • the correction unit 302 obtains DLa (Ip) by subtracting LUTa (Ip) from LUTa + 1 (Ip) as shown in the following equation.
  • FIG. 17A is a graph showing the difference between DLTa of LUTa + 1 and DLTa of LUTa, where the vertical axis shows DLa obtained by subtracting LUTa (Ip) from LUTa + 1 (Ip), and the horizontal axis is photocurrent. Ip is shown on a logarithmic scale.
  • DLa (Ip) is almost zero in the linear characteristic region and has a substantially constant value in the logarithmic characteristic region.
  • DLa (Ip) changes with a value corresponding to the difference in shape between LUTa + 1 and LUTa in the region near the inflection point.
  • the correction unit 302 calculates DIp (Ip) using the following equation in order to calculate the correction amount of the tilt shift according to the photocurrent Ip.
  • DIp (Ip) M ⁇ (log (Ip1) ⁇ log (Ip))
  • DIp (Ip) is represented by the graph in FIG. 17B, the vertical axis indicates DIp (Ip), and the horizontal axis indicates the photocurrent Ip on a logarithmic scale.
  • the correction unit 302 calculates an inclination shift correction amount LSCa (Ip) according to the photocurrent Ip using the following equation.
  • LSCa (Ip) is represented by the graph in FIG. 17C, the vertical axis represents LSCa (Ip), and the horizontal axis represents photocurrent Ip on a logarithmic scale. As shown in FIG. 17C, a region having a lower luminance than Ipb_a is removed from DIp by DLa ⁇ DIp, linearly increases from Ipb_a to Ipb_a + 1, and the slope of DIp is adjusted by Ipb_a + 1 to M. (Ip) is obtained.
  • the correction unit 302 adds the correction amount LSCa (Ip) of the inclination deviation to the LUTa (Ip).
  • LUTa (Ip) is corrected so that the standard inclination SLa_n matches the target pixel inclination SLx.
  • the inclination deviation correction amount LSCa (Ip) gradually increases from Ipb_a to Ipb_a + 1, the inclination deviation correction amount taking into account the characteristics of the nonlinear region near the inflection point. LSCa (Ip) is calculated.
  • the correction unit 302 inputs the output value LL of the LinLog signal of the target pixel to the corrected LUTa, and obtains the output value LLA.
  • the correction amount LSCa (Ip) of inclination deviation is calculated using the LUTa and the LUTa + 1 adjacent thereto.
  • the present invention is not limited to this, and the LUT separated from the LUTa by an arbitrary number (for example, LUTa + 3). ) May be used to calculate the tilt shift correction amount LSCa (Ip).
  • the adjustment of the absolute value of the inclination shift correction amount LSCa (Ip) may be performed by the coefficient M.
  • the LUT can be corrected in consideration of the nonlinear region, and the slope shift in the logarithmic characteristic region of the LinLog characteristic can be achieved. Can be corrected with high accuracy.
  • One possible cause of the output value shift in the high luminance range is the influence of unnecessary charges accumulated in the PD during periods other than the exposure period.
  • the PD accumulates charges.
  • unnecessary charges accumulated in the PD during these periods are accumulated in the FD and read out as a sig signal, so that information on unnecessary charges is included in the sig signal.
  • FIG. 18 is a sequence diagram of the pixel unit 220 according to Embodiment 5 of the present invention.
  • the output order is the sig signal and the ref signal, even if the difference between the sig signal and the ref signal is taken, the random noise generated when the FD is reset cannot be canceled accurately, and the S / N ratio is lowered. Is concerned. Therefore, in the present embodiment, in the LinLog signal acquisition process, the output order of the ref signal and the sig signal is performed in this order as in FIG.
  • the WR signal acquisition process is performed a plurality of times in one horizontal period, and these WR signals Are held in a frame memory provided in the subsequent stage. Then, an average of the WR signals held in the frame memory using an averaging circuit may be acquired as the final WR signal. Thereby, the random noise contained in the WR signal can be reduced.
  • the solid-state imaging device includes a plurality of pixel units including a photoelectric conversion element, a transfer transistor that flows a photocurrent so that charges accumulated in the photoelectric conversion element have a linear logarithmic characteristic, and a floating diffusion.
  • a pixel control unit that controls the pixel unit; and an image processing unit that processes a signal output from the pixel unit.
  • the pixel control unit applies an intermediate level voltage to a control terminal of the transfer transistor.
  • An exposure process for exposing the subject to the photoelectric conversion element, and the charge accumulated in the photoelectric conversion element by the exposure process is converted into a voltage by the transfer transistor being transferred to the floating diffusion to obtain a first signal.
  • a white reset is performed to reset the photoelectric conversion element to a zero bias state.
  • a leak process for applying the intermediate level voltage to the control terminal of the transfer transistor to leak the charge accumulated in the photoelectric conversion element for a predetermined first leak period; After the leakage process, the charge remaining in the photoelectric conversion element is converted into a voltage by the transfer transistor being transferred to the floating diffusion, and a second acquisition process for acquiring a second signal is performed on the pixel unit.
  • the image processing unit based on the second signal, from among a plurality of correction data in which a representative value of the second signal and a representative characteristic that is a photoelectric conversion characteristic corresponding to the representative value are associated with each other.
  • a correction unit is provided that selects one correction data and corrects the first signal using the selected correction data.
  • the correction data in which the representative value of the second signal is associated with the representative characteristic that is the photoelectric conversion characteristic corresponding to the representative value is stored in advance.
  • the second signal is a signal obtained by applying an intermediate level voltage to the control terminal of the transfer transistor after the white reset is completed and leaking the charge accumulated in the photoelectric conversion element during the first leak period. is there. Therefore, the second signal has a value corresponding to the variation of the inflection point of the photoelectric conversion characteristic inherent to each pixel portion.
  • correction data associated with the representative value of the second signal closest to the acquired second signal is specified, and the first signal is used using this correction data. Is corrected. Therefore, the representative characteristic indicated by the specified correction data indicates the characteristic closest to the photoelectric conversion characteristic of the target pixel among the representative characteristics stored in advance. As a result, even if correction data corresponding to each pixel portion is not stored in advance, the variation of the inflection point of the first signal can be corrected with high accuracy. Therefore, the amount of correction data can be reduced, and the variation of the inflection point of the first signal can be accurately corrected without increasing the scale of the solid-state imaging device.
  • the representative characteristic is preferably data obtained by grouping photoelectric conversion characteristics into a plurality of groups according to information on inflection points and averaging the photoelectric conversion characteristics belonging to each group. .
  • the representative characteristic is not the photoelectric conversion characteristic inherent to each pixel unit, but the average characteristic of the photoelectric conversion characteristics grouped according to the information on the inflection point. Therefore, the amount of correction data can be reduced as compared with the case where a configuration in which correction data is stored for each pixel unit is employed as in the prior art.
  • the pixel control unit executes the first acquisition process, the white reset process, the leak process, and the second acquisition process within one horizontal period.
  • the first and second signals are acquired and the first signal is acquired every horizontal period.
  • One signal can be corrected.
  • the storage capacity of the frame buffer can be reduced.
  • the correction data is data in which an environmental temperature and the intermediate level voltage applied to the control terminal of the transfer transistor are associated with each other, and the correction unit further includes the exposure in addition to the second signal. It is preferable to select one correction data based on the environmental temperature and the intermediate level voltage during processing.
  • the first signal can be corrected in consideration of the exposure time and the environmental temperature.
  • the correction data is configured by a look-up table indicating the representative characteristics.
  • the first signal can be quickly corrected using the lookup table.
  • the correction data has a coefficient of a function that defines a linear characteristic region in the representative characteristic and a coefficient of a function that defines a logarithmic characteristic region as the representative characteristic, and the correction unit uses the coefficient It is preferable to correct the first signal.
  • the pixel units are arranged in a matrix to form a pixel array unit, and are provided corresponding to a ramp waveform generation circuit that generates a ramp waveform and each column of the pixel array unit, A single slope AD converter that compares two signals with the ramp waveform and AD-converts the first and second signals, and the ramp waveform generation circuit performs AD conversion of the second signal
  • the ramp waveform has a predetermined offset.
  • the second signal is acquired with the same leak period in all the pixel portions, and the range that the second signal can take is not as wide as the first signal that is an image signal. Therefore, the second signal can be AD converted at high speed by AD conversion of the second signal by giving the ramp waveform the start level of the assumed range that the second signal can take as a predetermined offset.
  • the pixel control unit obtains a third signal by causing the pixel unit to perform the same process as the leak process in a second leak period different from the first leak period, and the correction unit includes the first leak period. It is preferable that the correction data is further corrected based on a difference between the two signals and the third signal, and the first signal is corrected using the corrected correction data.
  • the third signal is acquired by leaking in the second leak period.
  • the slope of the representative characteristics in the logarithmic characteristic region can be obtained by appropriately taking two values of the representative characteristics. Then, using both the inclinations, the inclination of the representative characteristic of the correction data is corrected to become the inclination of the photoelectric conversion characteristic of the target pixel, and the first signal is corrected using the corrected representative characteristic. Thereby, it is possible to remove the variation in the slope in the logarithmic characteristic region from the first signal.
  • the second leak period is longer than the first leak period.
  • the third signal indicates information of lower luminance of the target pixel than the second signal.
  • the correction unit calculates a standard gradient that is a gradient of the logarithmic characteristic region of the representative characteristic of the correction data selected based on the second signal, and based on the difference between the second and third signals, the target pixel
  • the target pixel inclination which is the slope of the logarithmic characteristic region of the photoelectric conversion characteristic of, is calculated, the correction amount of the standard inclination deviation with respect to the target pixel inclination is calculated based on the inclination of the standard inclination and the target pixel inclination, and calculated. It is preferable to correct the selected correction data based on the correction amount.
  • the target pixel inclination that is the inclination of the logarithmic characteristic region of the photoelectric conversion characteristic of the target pixel is calculated based on the difference between the second and third signals. Further, a standard inclination that is the inclination of the logarithmic characteristic region of the representative characteristic of the correction data selected from the second signal is calculated. Then, the correction amount of the standard pixel inclination deviation is calculated from both inclinations, and the representative characteristic of the correction data selected from the second signal is corrected using this correction amount. Therefore, by correcting the first signal using the corrected correction data, it is possible to remove variation in inclination in the logarithmic characteristic region from the first signal.
  • the correction unit calculates the correction amount based on a difference between the representative characteristic of the selected correction data and the representative characteristic of the other one correction data.
  • the correction amount is calculated based on the difference between the representative characteristics of the correction data and the other one correction data. Therefore, the first signal can be corrected in consideration of the non-linear characteristic near the inflection point.
  • the correction data includes data of a start point of a photoelectric conversion characteristic region to be corrected by the correction unit, and the correction unit selects the selection when the first signal is larger than the level of the start point. It is preferable to correct the corrected data.
  • the logarithmic characteristic region can be set as the correction target region by setting the start point as the inflection point. Also, by setting the starting point at an arbitrary point on the logarithmic characteristic area, an arbitrary area on the logarithmic characteristic area can be set as a correction target area.
  • a reading unit that reads a signal output from the pixel unit is further provided, and the pixel unit includes a floating diffusion (FD) to which charges accumulated in the photoelectric conversion element are transferred from the transfer transistor and the FD.
  • a reset transistor for resetting, and the first and second acquisition processes include a reference signal output process for turning off the reset transistor and outputting a reference signal from the FD, and turning on the transfer transistor and the photoelectric conversion element.
  • a difference process that causes the reading unit to calculate as a second signal and the first acquisition process includes the basic process Signal output processing is performed prior to the signal signal acquisition processing, the second acquisition process, the reference signal output processing is preferably performed after the signal signal acquisition process.
  • the first signal is acquired by performing the reference signal output process before the signal signal output process, the first signal having a high S / N ratio can be acquired.
  • the second signal is acquired by performing the reference signal output process after the signal signal output process, it is possible to suppress unnecessary charges from being included in the second signal.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Nonlinear Science (AREA)
  • Transforming Light Signals Into Electric Signals (AREA)

Abstract

L'invention concerne une unité de stockage de données de correction qui stocke à l'avance une pluralité de données de correction dans lesquelles des valeurs représentatives d'un signal WR et des caractéristiques représentatives qui sont des caractéristiques de conversion photoélectrique correspondant aux valeurs représentatives sont associées les unes aux autres. Une unité de correction corrige un signal LinLog à l'aide de données de correction dans lesquelles une valeur représentative qui est proche d'un signal WR obtenu lors de la capture d'une image est associée.
PCT/JP2013/000786 2012-02-28 2013-02-13 Dispositif de capture d'image à semi-conducteurs WO2013128817A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2013515614A JP5299597B1 (ja) 2012-02-28 2013-02-13 固体撮像装置

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2012-041740 2012-02-28
JP2012041740 2012-02-28

Publications (1)

Publication Number Publication Date
WO2013128817A1 true WO2013128817A1 (fr) 2013-09-06

Family

ID=49082033

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2013/000786 WO2013128817A1 (fr) 2012-02-28 2013-02-13 Dispositif de capture d'image à semi-conducteurs

Country Status (2)

Country Link
JP (1) JP5299597B1 (fr)
WO (1) WO2013128817A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20150116699A (ko) * 2014-04-08 2015-10-16 삼성전자주식회사 선형-로그형 이미지 센서 및 이를 포함하는 전자 기기
JP2019022198A (ja) * 2017-07-21 2019-02-07 キヤノン株式会社 撮像素子および撮像装置

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005509319A (ja) * 2001-04-03 2005-04-07 インスティチュート フュア ミクロエレクトロニク シュトゥットガルト イメージ信号のｆｐｎ補正方法及びそのための装置
JP2006074663A (ja) * 2004-09-06 2006-03-16 Renesas Technology Corp 固体撮像装置
JP2006140666A (ja) * 2004-11-11 2006-06-01 Renesas Technology Corp 撮像装置の駆動方法
JP2006303768A (ja) * 2005-04-19 2006-11-02 Sharp Corp 固体撮像装置
JP2007074488A (ja) * 2005-09-08 2007-03-22 Konica Minolta Holdings Inc 撮像装置
JP2007288479A (ja) * 2006-04-17 2007-11-01 Konica Minolta Holdings Inc 撮像装置
JP2008263546A (ja) * 2007-04-13 2008-10-30 Konica Minolta Holdings Inc 固体撮像装置、固体撮像装置の駆動方法、及びこれを用いた撮像システム
JP2009038772A (ja) * 2007-08-03 2009-02-19 Sony Corp 参照電圧回路および撮像回路
WO2009131190A1 (fr) * 2008-04-24 2009-10-29 コニカミノルタオプト株式会社 Dispositif d'imagerie à semiconducteur, procédé de commande de ce dispositif et système d'imagerie

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005509319A (ja) * 2001-04-03 2005-04-07 インスティチュート フュア ミクロエレクトロニク シュトゥットガルト イメージ信号のｆｐｎ補正方法及びそのための装置
JP2006074663A (ja) * 2004-09-06 2006-03-16 Renesas Technology Corp 固体撮像装置
JP2006140666A (ja) * 2004-11-11 2006-06-01 Renesas Technology Corp 撮像装置の駆動方法
JP2006303768A (ja) * 2005-04-19 2006-11-02 Sharp Corp 固体撮像装置
JP2007074488A (ja) * 2005-09-08 2007-03-22 Konica Minolta Holdings Inc 撮像装置
JP2007288479A (ja) * 2006-04-17 2007-11-01 Konica Minolta Holdings Inc 撮像装置
JP2008263546A (ja) * 2007-04-13 2008-10-30 Konica Minolta Holdings Inc 固体撮像装置、固体撮像装置の駆動方法、及びこれを用いた撮像システム
JP2009038772A (ja) * 2007-08-03 2009-02-19 Sony Corp 参照電圧回路および撮像回路
WO2009131190A1 (fr) * 2008-04-24 2009-10-29 コニカミノルタオプト株式会社 Dispositif d'imagerie à semiconducteur, procédé de commande de ce dispositif et système d'imagerie

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20150116699A (ko) * 2014-04-08 2015-10-16 삼성전자주식회사 선형-로그형 이미지 센서 및 이를 포함하는 전자 기기
KR102182870B1 (ko) 2014-04-08 2020-11-25 삼성전자주식회사 선형-로그형 이미지 센서 및 이를 포함하는 전자 기기
JP2019022198A (ja) * 2017-07-21 2019-02-07 キヤノン株式会社 撮像素子および撮像装置

Also Published As

Publication number Publication date
JPWO2013128817A1 (ja) 2015-07-30
JP5299597B1 (ja) 2013-09-25

Similar Documents

Publication Publication Date Title
US9680494B2 (en) Photoelectric conversion apparatus and image pickup system
JP4952301B2 (ja) 撮像装置およびカメラ
CN104471860B (zh) 信号处理装置和方法、成像元件以及成像设备
JP5034610B2 (ja) 固体撮像装置、固体撮像装置の信号処理方法および撮像装置
CN102647566B (zh) 用于cmos图像传感器的模拟行黑色电平校准
JP5218309B2 (ja) 固体撮像素子およびその制御方法、並びにカメラシステム
JP4929090B2 (ja) 固体撮像装置およびその駆動方法
CN102348074B (zh) 固态成像器件和照相机系统
JP5591586B2 (ja) 固体撮像装置、画像処理装置、カメラシステム
CN109155827B (zh) 高动态范围成像传感器阵列
TWI615031B (zh) 信號處理裝置及方法、攝像元件與攝像裝置
US10630912B2 (en) Image sensor, image capturing apparatus, and method for controlling image sensor
JP6643919B2 (ja) 信号処理装置及び方法、及び撮像装置
JP6355459B2 (ja) 撮像装置及びその制御方法
JP5299597B1 (ja) 固体撮像装置
JP2018014602A (ja) 撮像装置及びその制御方法、プログラム、記憶媒体
JP2016029759A (ja) 撮像装置の欠陥画素検出方法
JP2013183347A (ja) 固体撮像装置
JP6365040B2 (ja) 撮像装置、画像読取装置及び画像形成装置
JP2012244379A (ja) 固体撮像装置
JP5454741B1 (ja) 固体撮像装置
JP6433276B2 (ja) 撮像装置及びその制御方法、プログラム、記憶媒体
JP6432185B2 (ja) 撮像装置、画像読取装置及び画像形成装置
US11284017B2 (en) Image pickup device and storage medium
JP2013118595A (ja) 固体撮像装置

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2013515614

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13755049

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13755049

Country of ref document: EP

Kind code of ref document: A1