JP6482216B2 - Radiation imaging apparatus and driving method thereof - Google Patents

Description

  The present invention relates to a radiation imaging apparatus and a driving method thereof.

  The radiation imaging apparatus includes, for example, a plurality of sensors arranged on a substrate and a drive unit that drives each sensor. The sensor detects radiation, and a signal having a value corresponding to the radiation dose is read from the sensor by driving the sensor with the driving unit. For example, in a shooting mode such as moving image shooting or continuous shooting, irradiation of radiation to the sensor and signal reading from the sensor are repeated.

JP 2003-190126 A Japanese Patent Laid-Open No. 7-250283

  Since the signal read from the sensor has a signal component corresponding to the radiation dose and a noise component, correction for reducing the noise component may be performed on the signal from the sensor. For example, in a shooting mode such as moving image shooting or continuous shooting, the noise component is a noise component that depends on the time from one signal reading to the next signal reading, such as a noise component caused by charges accumulated in the sensor due to dark current. Including.

  Further, in the above imaging mode, for example, irradiation of radiation to the sensor and signal reading from the sensor are repeatedly performed. Therefore, the noise component further includes a noise component caused by charges remaining in the sensor at the time of signal reading. May be included. The charge remaining in the sensor may not be sufficiently reduced even if the initialization process is simply performed on the sensor. Therefore, this can lead to a reduction in image quality, such as an afterimage in a radiographic image obtained by subsequent signal readout.

  For example, Patent Literature 1 and Patent Literature 2 exemplify a technique for performing correction for reducing a noise component caused by dark current or the like, but this is caused by electric charge remaining in the sensor at the time of signal reading. The noise component is not considered.

  An object of the present invention is to provide a technique advantageous for reducing a noise component from a signal from a sensor.

  One aspect of the present invention relates to a radiation imaging apparatus, and the radiation imaging apparatus drives a plurality of sensors arranged in rows so as to form a plurality of rows and a plurality of columns on a substrate, and the plurality of sensors in units of rows. A radiation imaging apparatus comprising: a drive unit configured to perform a selection of the plurality of rows in a first order before the irradiation of radiation to the plurality of sensors is started. A first operation for initializing a plurality of sensors, and after the radiation irradiation is completed, the plurality of rows are selected in a second order, and the first operation is performed between the sensor in each row and the sensor in the adjacent row. Performing a second operation of driving the plurality of sensors and reading signals from the plurality of sensors so as to cause a difference in time from selection in the first order to selection in the second order, The radiation imaging apparatus has the second operation. Of the signals read from the plurality of sensors, a signal from a sensor in a first row that is one row in the plurality of rows, and a signal from a sensor in a second row adjacent to the first row A correction unit for correcting a signal from at least one sensor of the first row and the second row based on the time difference between the first row and the second row; And a light source for irradiating the plurality of sensors with light before the first operation.

  According to the present invention, it is advantageous to reduce a noise component from a signal from a sensor.

It is a figure explaining the example of the system configuration example of a radiation imaging device. It is a figure explaining the example of a structure of an imaging part. It is a figure explaining the example of the operation | movement flowchart of a radiation imaging device. It is a figure explaining the example of the operation | movement timing chart of a radiation imaging device. It is a figure explaining the example of the operation | movement timing chart of a radiation imaging device. It is a figure explaining the relationship between the difference of a signal value, and a noise component. It is a figure explaining the example of the operation | movement timing chart of a radiation imaging device. It is a figure explaining the example of the operation | movement timing chart of a radiation imaging device. It is a figure explaining the example of the operation | movement timing chart of a radiation imaging device. It is a figure explaining the example of the operation | movement timing chart of a radiation imaging device. It is a figure explaining the example of the plot figure of a signal value.

(1. First embodiment)
(1-1. Example of overall configuration of radiation imaging apparatus)
FIG. 1 is a block diagram showing an example of the overall configuration of a radiation imaging apparatus or radiation inspection apparatus. Here, an example of the overall configuration of the radiation imaging apparatus IA (hereinafter simply referred to as “apparatus IA”) will be described. The apparatus IA includes, for example, an imaging unit 10, a driving unit 20, a radiation generation source 30, a control unit 40, a processing unit 50, and a display unit 60.

  The imaging unit 10 includes, for example, a sensor array 110 in which a plurality of sensors for detecting radiation are arranged. The imaging unit 10 further includes, for example, a scintillator (not shown) that converts radiation into light. In this case, a photoelectric conversion element that detects the converted light can be used for each sensor. The scintillator can be disposed on the sensor array 110. With such a configuration, the imaging unit 10 detects radiation that has passed through the body of the subject (patient or the like), and acquires image data indicating information in the body of the subject.

  The imaging unit 10 further includes a light source 11 for irradiating the sensor array 110 with light. The light source 11 only needs to irradiate each sensor of the sensor array 110 with a sufficient amount of light. For example, the light source 11 may be disposed on the back surface of the sensor array 110 and may be disposed so as to overlap the sensor array 110 in plan view (plan view with respect to the upper surface of the sensor array 110), for example. Alternatively, the light source 11 may be disposed on the outer side of the sensor array 110 along the outer periphery in plan view, or may be disposed in the vicinity of each corner of the sensor array 110. In addition to the light source 11, a member for propagating light from the light source 11 to the entire area of the sensor array 110 may be used.

  The drive unit 20 drives the imaging unit 10 based on a predetermined drive signal or control signal, and performs drive control for performing radiation imaging. The radiation generation source 30 generates radiation based on a predetermined control signal and irradiates the imaging unit 10 with radiation. Radiation includes X-rays, α rays, β rays, γ rays, and the like. The control unit 40 outputs a control signal to the drive unit 20 and the radiation generation source 30 to control operations of the drive unit 20 and the radiation generation source 30, and can perform synchronous control of each unit. To control the whole.

  The apparatus IA can be configured, for example, so that the imaging unit 10 detects that radiation irradiation has started. For example, the imaging unit 10 may be configured to detect that radiation irradiation has started using at least some of the plurality of sensors or other sensors. More specifically, for example, each sensor may be driven before radiation irradiation to detect that radiation irradiation has started based on a signal from the sensor, or a dedicated sensor may be provided individually. The start of radiation irradiation may be detected based on the signal from the sensor. According to such a configuration, synchronization control between the imaging unit 10 and another unit can be automatically performed by the control unit 40 or the like, for example.

  The processing unit 50 receives image data from the imaging unit 10 and performs predetermined data processing, and also includes, for example, a correction unit 51 and a calculation unit 52, and performs correction processing on the image data. The calculation unit 52 calculates correction information based on the image data, and may be referred to as a calculation unit. The correction unit 51 corrects the image data using the calculated correction information. The display unit 60 receives image data from the processing unit 50 and displays an image (radiation image) indicating a state inside the subject's body.

  The apparatus IA is not limited to the above-described configuration. For example, another unit may have a part of the functions of a unit, or two or more units may be integrated. Also good. For example, the imaging unit 10 may be configured to include the control unit 40 and the processing unit 50, or the control unit 40 and the processing unit 50 may be configured integrally. In addition, transmission / reception of signals between the units may be performed in a wired manner or wirelessly.

(1-2. Configuration Example of Imaging Unit)
FIG. 2 shows a configuration example of the imaging unit 10. The imaging unit 10 includes, for example, a sensor driving unit 120, a signal reading unit 130, and a signal output unit 140 in addition to the sensor array 110 in which a plurality of sensors s are arranged. Here, in order to make the drawing easier to see, a sensor array 110 of 3 rows × 3 columns is illustrated.

  The sensor driving unit 120 drives each sensor s in units of rows based on a signal from the driving unit 20, for example, and outputs a signal corresponding to the amount of electric charge generated when each sensor s detects radiation. The sensor drive unit 120 includes, for example, a shift register, and sequentially selects rows to be driven based on, for example, a clock signal.

  The sensor s includes, for example, a PIN photodiode, a MIS sensor, and the like, and is formed using, for example, amorphous silicon on a glass substrate. Each sensor s is connected to a switch element w for outputting a signal corresponding to the amount of electric charge generated in the sensor s. It is output to the signal line 150. As the switch element w, for example, a thin film transistor (TFT) can be used.

  The signal from the sensor s may simply be referred to as a “sensor signal”. Here, one sensor s and the corresponding switch element w form a unit pixel, and a sensor signal or a signal based thereon can also be referred to as a “pixel signal”. The value can also be referred to as a “pixel value”.

  The signal readout unit 130 includes, for example, a signal amplification unit 131 that amplifies signals from the sensors s in each column and a holding unit 132 that samples and holds the amplified signals. The held signals are sequentially read based on the control signal. The shift register SR supplies a control signal for reading the signal held in the holding unit 132 to the signal reading unit 130 based on, for example, a signal from the driving unit 20. With such a configuration, the signal reading unit 130 sequentially reads the signals from the sensors s of each column via the column signal line 150 and horizontally transfers them to the signal output unit 140.

  The signal output unit 140 includes, for example, an output buffer amplifier 141 for buffering the signal read by the signal reading unit 130, and an AD conversion unit 142 for analog-digital conversion (AD conversion) of the buffered signal. Have With such a configuration, the signal output unit 140 sequentially outputs the signal read by the signal reading unit 130 to the above-described processing unit 50 as image data (digital data).

(1-3. Flowchart of radiation imaging)
FIG. 3 illustrates a flowchart for performing radiation imaging. For example, radiography such as the moving image shooting mode and the continuous shooting mode is performed by irradiating a plurality of times of radiation. In this flowchart, steps S101 to S105 described below are mainly performed. Hereinafter, in this specification, step S101 and the like are simply referred to as “S101” and the like. S101 to S104 are a series of steps for acquiring radiation data corresponding to one irradiation of radiation.

  In S101, light irradiation LI for irradiating the sensor array 110 with light is performed. The light irradiation LI is performed using the light source 11 described with reference to FIG. In the light irradiation LI, as long as charges are generated and accumulated in each sensor s by the irradiated light, a sufficient amount of light may be irradiated over a predetermined period so that the signal value of the sensor signal is saturated.

  In S102, an initialization operation RS is performed to initialize (reset) each sensor s of the sensor array 110 in response to the end of the light irradiation LI. In the initialization operation RS, the charge (or at least part of the charge) accumulated in each sensor s by the light irradiation LI is removed, and the sensor s is initialized.

  In S103, in response to the completion of the initialization operation RS, for example, a control signal (exposure permission signal) that requests the start of radiation irradiation is output to the radiation source or the radiation control unit that controls the radiation source, and radiation irradiation is performed. The accumulation operation AO for accumulating charges in each sensor s is started. In the accumulation operation AO, the switch element w corresponding to each sensor s is maintained in a non-conductive state, and an amount of electric charge corresponding to the amount of irradiated radiation is accumulated in each sensor s. The accumulation operation AO can be performed, for example, from the end of radiation irradiation for one time until a predetermined period elapses.

  In S104, a read operation RO for reading a signal from each sensor s is performed in accordance with the end of the accumulation operation AO. The read operation RO is performed by bringing the corresponding switch element w into a conductive state in units of rows. As a result, a signal having a value according to the amount of charge accumulated in the accumulation operation AO is read from each sensor s.

  Referring to FIG. 2 again, the initialization operation RS in S102 and the readout operation RO are the same in that the sensor driving unit 120 makes the switch elements w conductive in units of rows. It is a driving method. In the initialization operation RS, for example, the switch element w is turned on while the column signal line 150 is connected to the reference potential by a switch (not shown). Alternatively, in the initialization operation RS, for example, the switch element w is turned on while initializing the signal amplification unit 130 (feedback capacitor thereof). As a result, the charge of each sensor s is released to the reference potential via the column signal line 150.

  In S105, it is determined whether or not the imaging is finished, specifically, whether or not there is next irradiation. If the next radiation is irradiated, the process returns to S101, and if there is no next radiation, the imaging is terminated. The determination of this step can be made based on imaging information preset by the user, the number of times of radiation irradiation, and the like. In addition, when the irradiation of the next radiation is not started even after a predetermined time has elapsed after the irradiation of a certain radiation is completed, the imaging may be terminated. In addition, here, a mode in which a plurality of times of radiation irradiation in the moving image shooting mode and the continuous shooting mode is illustrated, but for example, in the case of the still image shooting mode, radiation data corresponding to one irradiation of radiation is acquired. After that, the actual photographing may be finished without returning to S101.

  According to this flowchart, a plurality of times of radiation is irradiated when moving image shooting or continuous shooting is performed, and one time of radiation is irradiated when still image shooting is performed. Then, a series of operations of light irradiation LI (S101), initialization operation RS (S102), accumulation operation AO (S103), and readout operation RO (S104) corresponding to one irradiation of radiation is performed. By this series of operations, signals from radiation irradiation for one time are obtained from the plurality of sensors s. In this specification, a signal from the sensor s obtained by this series of operations is a signal for one frame (or unit frame), and one image data is formed based on the signal for the one frame.

  Here, by performing the initialization operation RS in S102 after the light irradiation LI in S101 is performed so that the signal value of the sensor signal is saturated, the charge remaining in each sensor s (residual charge) is uniformly reduced. . For example, in the case of moving image shooting or continuous shooting, when starting a series of operations of S101 to S104 corresponding to a signal for one frame, each sensor s has a charge when reading the signal of the immediately preceding frame. Can remain. This residual charge can offset the characteristics of the sensor s, for example, by shifting the threshold voltage of the switch element w, which is a TFT, and the charge transfer efficiency by the switch element w. Sensor s may not be initialized uniformly. This residual charge causes an afterimage in the radiographic image based on the signal for one frame read out thereafter.

  Therefore, in the present embodiment, the plurality of sensors s when performing the initialization operation RS are saturated by the light irradiation LI. Then, the residual charge of each sensor s is uniformly reduced by the initialization operation RS. Thereby, the afterimage that may occur in the radiographic image can be suppressed. Similarly, also in the case of still image shooting, by performing the light irradiation LI before performing the initialization operation RS, the residual charge of each sensor s before shooting starts can be uniformly reduced, and the afterimage can be suppressed.

(1-4. Example of Driving Method of Imaging Unit)
Hereinafter, the driving method according to the present embodiment (mainly, the initialization operation RS in S102 and the read operation RO in S104) described above will be described with reference to FIG. 4 and the like.

  FIG. 4 shows a drive timing chart of the imaging unit 10 having the sensor array 110 of X rows × Y columns. The horizontal axis is the time axis. The vertical axis represents signals Vg (1) to Vg (X) for driving each sensor s. For example, assuming that i = 1 to X, Vg (i) is a signal for driving each sensor s in the i-th row. In this configuration, the conduction state or non-conduction state of the corresponding switch element w is determined. It is a signal to control. Each switch element w in the i-th row becomes conductive when Vg (i) is at a high level (H), and becomes non-conductive when it is at a low level (L).

  Here, a period T1 for acquiring a signal for one frame is mainly described, but the same applies to other periods (for example, a period T2 for acquiring a signal for the next one frame).

  In the period T1, a series of operations of S101 to S104 described above corresponding to one irradiation of radiation, specifically, the light irradiation LI, the initialization operation RS, the accumulation operation AO, and the read operation RO are performed. Here, in the present embodiment, the initialization operation RS and the read operation RO are performed in an interlaced manner. Specifically, first, odd rows (first row, third row, fifth row,..., (X-1) row) and even rows (second row, fourth row, sixth row, .., The Xth row) is selected, and then the other sensors s of the odd and even rows are selected.

In the figure, among the initialization operations RS, an operation for initializing the sensor s while sequentially selecting odd rows is indicated as “KI _O ”, and an operation for initializing the sensor s while sequentially selecting even rows is denoted by “KI”. _E ". In the readout operation RO, an operation for driving the sensor s while sequentially selecting odd rows is denoted as “HI _O ”, and an operation for driving the sensor s while sequentially selecting even rows is denoted as “HI _E ”. Yes.

In other words, the initialization operation RS includes an operation KI _O to initialize sensor s in an odd row, and operation KI _E to initialize sensor s in an even row, the. Further, read operation RO includes an operation HI _O for reading a signal from the sensor s in an odd row, and operation HI _E for reading a signal from the sensor s in an even row, the.

In the period T1 in the figure, the initialization operation RS is performed by performing the operation KI _E after performing the operation KI _O. Thereafter, after performing the accumulation operation AO, the read operation RO is performed by performing the operation HI _O after performing the operation HI _E.

Here, in the sensor s in a certain row, the time from the initialization operation RS to the read operation RO in the period T1 is _defined as “time t _eff ”. That is, the time t _eff for a certain row is the time from the initialization for the one row in the period T1 to the driving for the one row.

Here, the time _{t eff} of the k-th row and _{"t eff} (k)", the first (k + 1) the time _{t eff} of the row as a _"t eff (k + 1)", the time required for the initialization operation RS " t _R ”and the time required for the accumulation operation AO is“ t _AO ”. Further, here, the time required for the read operation RO is assumed to be equal to the time t _R required for the initialization operation RS in order to facilitate explanation. At this time,
t _eff (k) = _3/2 × t _R + t _AO ,
t _eff (k + 1) = 1/2 × t _R + t _AO
... (Formula 1)
It can be expressed. That is, according to the driving method of the present embodiment, there is a difference in time t _eff between the k-th row and the (k + 1) -th row. Here, the case where the k-th row is an odd-numbered row is illustrated, but the same applies to the case where the k-th row is an even-numbered row.

  Here, for ease of explanation, the case where the time required for the initialization operation RS and the time required for the read operation RO are equal to each other is illustrated, but these may be different from each other. As one example, the pulse width of Vg (i) in the read operation RO may be larger than the pulse width of Vg (i) in the initialization operation RS.

(1-5. Example of image data correction method)
As described above, according to the light irradiation LI and the initialization operation RS, it is possible to suppress the afterimage due to the residual charge of the sensor s when reading the signal for the immediately preceding one frame. However, when a signal for the next one frame is read out, there is a possibility that an artifact may occur in the radiation image based on the signal. This artifact is a relatively low-frequency noise component, which can be said to be another afterimage accompanying the saturation of the plurality of sensors s by the light irradiation LI. Therefore, in the present embodiment, correction for reducing the noise component is performed using correction information calculated based on the above-described difference in time t _eff .

In principle, it is also possible to correct the afterimage resulting from the residual charge of the sensor s when reading the signal for the immediately preceding one frame using correction information calculated based on the difference in time t _eff. It is. However, such an afterimage includes information on the subject image (subject afterimage). Therefore, when a high-frequency noise component (for example, a sharp edge of a contour portion) exists in the afterimage of the subject, it is difficult to perform correction using correction information calculated based on the difference in time t _eff. there is a possibility. According to this embodiment, an afterimage can be suitably suppressed even when such a high-frequency noise component is present.

  A signal component based on the amount of charge generated in the sensor s due to radiation is defined as S0. Further, a noise component having time dependency such as a noise component caused by dark current or the like is defined as N1. Further, N2 is a noise component that does not have time dependency, such as fixed pattern noise (FPN) due to sensor configuration, element variation, and the like.

At this time, the signal SS from the sensor s is
SS = S0 + N1 + N2
... (Formula 2)
It can be expressed.

  Here, the noise component N1 can be reduced by elapse of a sufficiently long time. However, in the moving image shooting and the continuous shooting exemplified in the present embodiment, the series of operations of S101 to S104 are repeatedly performed in a relatively short time. Therefore, the noise component N1 causes an afterimage in the image corresponding to each frame. Cause.

The noise component N1 is given by a predetermined noise model, and as one typical example,
α (t) = a (constant)
... (Formula 3)
Given in. In this case, the noise component N1 uses the time t,
N1 = ∫α (t) dt
... (Formula 4)
It can be expressed. here,
ts: time when the read operation RO is performed in the period T1,
te: Assuming that the read operation RO is performed in the period T2,
N1 = a × (te−ts)
... (Formula 5)
It can be expressed.

  Here, the constant a can take a different value for each sensor s (for each pixel), but the inventor of the present application has found that the constant a is substantially equal between adjacent sensors. Further, the noise components N2 are substantially equal between adjacent sensors. Further, in a region where the change of the signal component is small (for example, a region other than the portion forming the contour in the image), the signal components S0 are substantially equal between adjacent sensors.

For example, the signal from the sensor s (m, n) in the m-th row and the n-th column is the signal SS (m, n), and the signal from the sensor s (m + 1, n) in the (m + 1) -th row and the n-th column. Is a signal SS (m + 1, n). Further, components S0, N1 and the like corresponding to each sensor s are set as S0 (m, n), N1 (m, n) and the like. At this time, from the equations (2) to (5),
SS (m, n)
= S0 (m, n) + N1 (m, n) + N2 (m, n)
= A (m, n) x {te (m)-ts (m)} + S0 (m, n) + N2 (m, n),
SS (m + 1, n)
= S0 (m + 1, n) + N1 (m + 1, n) + N2 (m + 1, n)
= A (m + 1, n) * {te (m + 1) -ts (m + 1)} + S0 (m + 1, n) + N2 (m + 1, n),
a (m, n) ≈a (m + 1, n),
S0 (m, n) ≈S0 (m + 1, n),
N2 (m, n) ≈N2 (m + 1, n),
... (Formula 6)
It can be expressed.

According to (Equation 6) above, the difference between the signal SS (m, n) and the signal SS (m + 1, n) is
SS (m, n) -SS (m + 1, n)
= A (m, n) * [{te (m) -ts (m)}-{te (m + 1) -ts (m + 1)}]
... (Formula 7)
It can be expressed. Therefore,
a (m, n)
= {SS (m, n) -SS (m + 1, n)} / [{te (m) -ts (m)}-{te (m + 1) -ts (m + 1)}]
... (Formula 8)
It can be expressed.

Here, using the above-described (Equation 1), k = m,
te (m) −ts (m) = t _eff (m) = 3/2 × t _R + t _AO ,
te (m + 1) −ts (m + 1) = t _eff (m + 1) = ½ × t _R + t _AO
It can be expressed. Therefore, from (Equation 8) above,
a (m, n) = {SS (m, n) -SS (m + 1, n)} / t _R
... (Formula 9)
Is calculated.

Therefore, referring to (Equation 5) again, the noise component N1 (m, n) for the sensor s (m, n) can be calculated. In particular,
About the m-th row (odd row)
N1 (m, n) = {SS (m, n) −SS (m + 1, n)} × {3/2 × t _R + t _AO } / t _R
... (Formula 10a)
It becomes. Also,
About the m + 1st line (even number line)
N1 (m + 1, n) = {SS (m, n) -SS (m + 1, n)} × {1/2 × t R + t AO} / t R
... (Formula 10b)
It becomes.

As described above, the signal SS (m, n) can be corrected to reduce the noise component N1 (m, n), and a corrected signal SS ′ (m, n) can be obtained. The corrected signal SS ′ (m, n) is
SS ′ (m, n) = SS (m, n) −N1 (m, n)
... (Formula 11)
It is.

  As described above, according to the present embodiment, the noise component N1 can be calculated, and correction for reducing the noise component N1 can be performed on the signal SS from the sensor s. Here, the case where the m-th row is an odd-numbered row and the (m + 1) -th row is an even-numbered row is illustrated, but in the reverse case, the constant a is calculated and the noise component N1 is calculated in the same procedure. Can do.

  Here, a reference example of the driving method of the imaging unit 10 will be described with reference to FIG. 5A and 5B show a drive timing chart in the reference example, similar to FIG. 4 described above.

FIG. 5A shows a first reference example. In the first reference example, each of the initialization operation RS and the read operation RO is performed in a progressive manner. That is, in the first reference example, the initialization operation RS is performed by initializing the sensor s while sequentially selecting the first row, the second row, the third row,..., The Xth row. In the first reference example, the reading operation RO is performed by driving the sensor s while sequentially selecting the first row, the second row, the third row,..., The Xth row. In the drawing, the operation for initializing the sensor s by the progressive method is indicated as “KP”, and the operation for driving the sensor s by the progressive method is indicated by “HP”. Here, in the first reference example,
t _eff (k) = t _eff (k + 1) = t _R + t _AO
It becomes. Therefore, according to the first reference example, there is substantially no difference in the time t _eff between the k-th row and the (k + 1) -th row.

FIG. 5B shows a second reference example. In the second reference example, although each of the initialization operation RS and read operations RO is performed in interlace method, it performs an initialization operation RS by performing the operation KI _E after performing operation KI _O, operation HI performing _O is performing a read operation RO by performing the operation HI _E from. for that reason,
t _eff (k) = t _eff (k + 1) = t _R + t _AO
It becomes. Therefore, according to the second reference example, there is substantially no difference in time t _eff between the k-th row and the (k + 1) -th row.

According to these reference examples, there is substantially no difference in time t _eff between adjacent rows. Therefore, according to these reference examples, it can be said that it is difficult to calculate the constant a (and the noise component N1).

On the other hand, according to the present embodiment, a series of operations for reading a signal for one frame (that is, a series of operations from the initialization operation RS to the read operation RO) has a difference in time t _eff between adjacent rows. Do as follows. In the example of this embodiment, illustrate embodiments for performing the operation KI _O performs the initialization operation RS by performing the operation KI _E after performing read operation RO by performing the operation HI _O after performing operation HI _E did. Based on the difference in time t _eff, a correction coefficient (here, constant a of the noise component N1) is calculated as correction information. As a result, the noise component N1 can be calculated, and correction for reducing the noise component N1 can be performed on the signal SS from the sensor s.

In the present embodiment, the apparatus IA may further include a measurement unit for measuring the time t _eff of each row in order to perform the above-described correction. The measurement result by the measurement unit is supplied to the processing unit 50 together with the image data obtained by the imaging unit 10. The measurement unit may be provided inside the imaging unit 10 or may be provided inside the control unit 40. Note that, when the order of the operation HI _O and the operation HI _E of the read operation RO is determined, the time t _eff of each row can be specified, and in this case, the measurement unit is not used. Good.

(1-6. Summary of the present embodiment)
In the present embodiment, the state of the plurality of sensors s initialized while initializing the plurality of sensors s by performing the initialization operation RS after the signal value of the signal of each sensor s is saturated by the light irradiation LI. Homogenize. Thereby, it can suppress that the afterimage resulting from the residual charge of the sensor s arises in the radiographic image based on the flame | frame acquired after that. Then, the read operation RO is performed so that a difference occurs in the time t _eff between adjacent rows. In the present embodiment performs an initialization operation RS by performing the operation KI _E after performing operation KI _O, and illustrate embodiments of performing reading operation RO by performing the operation HI _O after performing the operation HI _E. The sensor signal obtained by the read operation RO is corrected to reduce the noise component N1 having time dependency based on the difference in time t _eff between the adjacent rows. Thereby, the artifact accompanying light irradiation LI can be prevented from occurring in the radiographic image, and the quality of the radiographic image can be improved.

(1-7. Modification)
In this embodiment, although the aspect which calculates correction information thru | or a correction coefficient (here constant a) was illustrated based on the difference of the signal value between adjacent rows, this invention is not limited to this aspect.

  For example, when the sensor array 110 is divided into several regions R (not shown) and the constant a (R) in each region R is equal, the difference in signal value between two regions R adjacent to each other From this, the constant a (R) for each region R may be calculated. At this time, the division of the sensor array 110 into the region R may be performed every two or more rows, may be performed every two or more columns, or may be performed every two or more rows and two or more rows. This may be done for each unit area of the column. Further, it may be assumed that the constant a is equal for all the sensors s of the sensor array 110. That is, the correction information or the correction coefficient may be calculated for each of one or more unit areas R, and the setting of the area R can be changed according to the shooting conditions, the shooting target, and the like.

  Further, one correction information for the region R may be determined using each of the signals from the sensor s in the region R. For example, an average value of a plurality of calculation results obtained using each signal from the sensor s in a certain region R may be used as a correction coefficient, or a median value or mode value may be used as a correction coefficient instead of the average value. Good. Furthermore, the correction coefficient may be calculated based on statistics of a plurality of calculation results, such as using a standard deviation.

  Further, the target for setting the correction information or the like is not limited to the region R of the sensor array 110. For example, one correction information is determined every predetermined period in radiography, every predetermined number of signal readout ROs (a predetermined number of frames), and two or more image data are corrected using the same correction information. Also good.

  In this embodiment, the correction information is calculated based on the difference from the signal of the sensor s in the adjacent row. However, the calculation is performed using the signals of the sensors s in the adjacent row. And may be done using their average.

Further, in the present embodiment, for ease of explanation, the noise component N1 is considered with the simple noise model of (Equation 2), but other noise models may be used. For example, a noise model due to dark current can be given by β (t) = b × t ⁻¹ using a constant b. That is, a noise model constant (constant a or the like) may be calculated based on the difference in time t _eff between adjacent rows.

  In addition, in the present embodiment, an example in which the constant a is calculated on the assumption of a predetermined noise model is exemplified, but the correction coefficient can also be calculated without using the noise model. FIG. 6A shows signal values (pixel values) of the sensors s in each row in image data acquired in advance separately from the series of operations from the initialization operation RS to the read operation RO, and the horizontal axis represents the horizontal axis. This is the row number of the sensor array 110, and the vertical axis represents the signal value. According to FIG. 6A, an average value a0 of signal values between a certain row and its adjacent row, a difference d0 of signal values between the certain row and its adjacent row, another row and its adjacent row. And the signal value difference d1 between the other row and its adjacent row is obtained. FIG. 6B is a plot based on the above a0, a1, d0, and d1, the horizontal axis is the difference between signal values (d0 to d1) between adjacent rows, and the vertical axis is the average of the signals in each row. It is a value (a0 to a1). According to FIG. 6B, a noise component having time dependency can be predicted based on the average value and the difference, and the relational expression can be calculated. The correction information or the correction coefficient may be calculated based on this relational expression.

(2. Second Embodiment)
From the first embodiment described above, so that the difference in time t _eff of adjacent rows occurs, performs an initialization operation RS by performing the operation KI _E after performing operation KI _O, performs an operation HI _E and illustrate embodiments for performing the read operation RO by performing the operation HI _O. However, the present invention is not limited to these modes, and the initialization operation RS and the read operation RO may be performed by selecting rows in another order.

FIG. 7 shows a drive timing chart according to the second embodiment, as in the first embodiment (FIG. 4). In the present embodiment, each of the initialization operation RS and the read operation RO is performed by an interlace method in which a row is selected every three rows. More specifically, in this embodiment, the initialization operation RS is performed by three types of operations KI _{1 to} KI ₃ , and the read operation RO is performed by three types of operations HI _{1 to} HI ₃ . .

In operation KI _1, first row, fourth row, the seventh row, ..., initializing the sensor s while selecting the (X-2) rows in sequence. In operation KI _2, the second row, the fifth row, the eighth row, ..., initializing the sensor s while selecting the (X-1) row in order. In operation KI _3, the third row, the sixth row, row 9, ..., initializing the sensor s while selecting the X-row sequentially. Further, the operation HI _1, the first row, fourth row, the seventh row, ..., drives the sensor s while selecting the (X-2) rows in sequence. In operation HI _2, the second row, the fifth row, the eighth row, ..., drives the sensor s while selecting the (X-1) row in order. In operation HI _3, the third row, the sixth row, the ninth row, ..., drives the sensor s while selecting the X-row sequentially.

Here, according to FIG. 7, the initialization operation RS is performed in the order of operations KI ₁ , KI ₂ , and KI ₃ , while the read operation RO is performed in the order of operations HI ₁ , HI ₃ , and HI ₂ . Even with such a driving method, there is a difference in time t _eff between adjacent rows.

  From another viewpoint, in this embodiment, the sensor array 110 (arranged sensors s) is divided into three groups in units of rows, and one row of one group is divided into rows of other groups. Adjacent. That is, the sensor array 110 is divided so that two adjacent rows are in different groups. Then, the initialization operation RS and the read operation RO are sequentially performed in units of groups. How to divide the sensor array 110 into groups may be determined by the control unit 40. In this case, the control unit 40 functions as a division unit. The control unit 40 may include a determination unit (not shown) that determines how to divide the sensor array 110. The sensor array 110 may be driven in groups by the drive unit 20 (more specifically, the sensor drive unit 102) based on a control signal from the control unit 40 or the like.

In the figure, for k-th row, k = 3j−2 (k is an integer of 1 to X, j is an integer of 1 or more), that is, the case where the remainder when k is divided by 3 is 1. ing. in this case,
t _eff (k) = t _R + t _AO ,
t _eff (k + 1) = _4/3 × t _R + t _AO ,
t _eff (k + 2) = 2/3 × t _R + t _AO
... (Formula 12)
Thus, there is a difference in time t _eff between adjacent rows.

Thus, the driving method in which a difference in time t _eff between adjacent rows is not limited to the interlace method exemplified in the first embodiment, and the same effect can be obtained by another interlace method.

As described above, also in this embodiment, the initialization operation RS and the read operation RO can be performed so that a difference in time t _eff occurs between adjacent rows. Then, correction information or a correction coefficient can be calculated based on the difference in the time t _{eff in} the same procedure as in the first embodiment. Therefore, the present embodiment can provide the same effects as those of the first embodiment.

  Here, for ease of explanation, the interlace method for every three rows has been exemplified, but the interlace method for every four rows may be used, or more. X may not be a multiple of 3. Further, the other parameters are not limited to the quantities exemplified here.

(3. Third embodiment)
In the first and second embodiments described above, the mode in which the initialization operation RS and the read operation RO are performed in an interlaced manner is illustrated so that a difference occurs in the time t _eff between adjacent rows. However, the present invention is not limited to these embodiments.

FIG. 8 shows a drive timing chart according to the third embodiment in the same manner as in the first embodiment (FIG. 4). In relation to the first embodiment, in the present embodiment, the initialization operation RS is mainly performed by the progressive method (operation KP), and the read operation RO is performed by the interlace method (operations HI _O and HI _E ). It is different in that. In relation to the second embodiment described above, this embodiment mainly performs the read operation RO without dividing the sensor array 110 into groups in the period T1, and performs the division in the period T2. The difference is that read drive RO is performed.

In the figure, the case where the k-th row is an odd-numbered row is illustrated. If the kth row is an odd row,
t _eff (k) = {1− (k−1) / 2X} × t _R + t _AO ,
t _eff (k + 1) = {3 / 2−k / 2X} × t _R + t _AO
... (Formula 13)
Thus, there is a difference in time t _eff between adjacent rows. Although the case where the k-th row is an odd row is illustrated here, the noise component N1 can be calculated by calculating the constant a in the same procedure when the k-th row is an even-numbered row.

As described above, also in the present embodiment, the correction information or the correction coefficient can be calculated based on the difference in the time t _eff . Therefore, the present embodiment can provide the same effects as those of the first embodiment.

In the example of FIG. 8, the progressive method (operation KP) is applied to the initialization operation RS, and the interlace method (operations HI _O and HI _E ) is applied to the read operation RO. The method may be applied. That is, as illustrated in Figure 9, interlaced in the initialization operation RS (operation KI _O and KI _E) is applied, progressive method (operation HP) may be applied to a read operation RO. Thus, the driving method in which the difference in time t _eff between adjacent rows is not limited to the interlace method, but the interlace method is applied to one of the initialization operation RS and the read operation RO, and the progressive method is applied to the other. The same effect can be obtained by applying.

Further, in the present embodiment, the interlace method in the case where the number of rows of the sensor s to be selected (at one time) is 1 is illustrated, but as a modified example, as illustrated in FIG. The number of may be two or more. That is, the driving KI _O, first and third rows, the fifth row and the seventh row, ..., the (X-3) and the (X-1) row, and choose in units of two rows The sensor s is initialized. Further, in the drive KI _E , the sensor s is selected while selecting the second row and the fourth row, the sixth row and the eighth row,..., The (X-2) and the Xth row in order of two rows. initialize. The opposite method may be similarly applied to each of the initialization operation RS and the read operation RO. Further, the number of selected rows may be three or more.

(4. Fourth embodiment)
In the fourth embodiment, in addition to the correction for reducing the noise component N1, the correction for reducing the noise component N2 is further performed. As described above, the noise component N2 is a noise component that does not have time dependency, such as a noise component caused by FPN. This correction is performed, for example, from an initialization operation (referred to as “RS2”) to a reading operation (referred to as “RO2”) in a state in which no radiation is applied to the apparatus IA before or after the start of radiation imaging. A series of operations may be performed separately and performed based on image data obtained by the series of operations. The series of operations can be performed by the same driving method as described above, but in a state in which no radiation is applied to the apparatus IA.

Hereinafter, a signal from the sensor s (m, n) in the m-th row and the n-th column obtained in the read operation RO is referred to as a signal SS ₁ (m, n). In addition, the components S0, N1 and the like corresponding thereto are defined as S0 ₁ (m, n), N1 ₁ (m, n) and the like. At this time, the signal SS ₁ (m, n) is
SS ₁ (m, n)
= S0 ₁ (m, n) + N1 ₁ (m, n) + N2 ₁ (m, n)
= A (m, n) × {t 1 e (m) -t 1 s (m)} + S0 1 (m, n) + N2 1 (m, n)
... (Formula 14)
It can be expressed.

A signal from the sensor s (m, n) in the m-th row and the n-th column obtained in the read operation RO2 is defined as a signal SS ₂ (m, n). In addition, the components S0, N1 and the like corresponding thereto are defined as S0 ₂ (m, n), N1 ₂ (m, n) and the like. At this time, the signal SS ₂ (m, n) is
SS ₂ (m, n)
= N1 ₂ (m, n) + N2 ₂ (m, n)
= A (m, n) × {t 2 e (m) -t 2 s (m)} + N2 2 (m, n)
... (Formula 15)
It can be expressed.

A signal SS _C (m, n) obtained by correcting the signal obtained in the read operation RO based on the signal obtained in the read operation RO2 is:
SS _C (m, n)
≡SS ₁ (m, n) −SS ₂ (m, n)
= {S0 ₁ (m, n) + N1 ₁ (m, n) + N2 ₁ (m, n)}-{N1 ₂ (m, n) + N2 ₂ (m, n)}
= S0 ₁ (m, n) + a (m, n) × [{t ₁ e (m) −t ₁ s (m)} − {t ₂ e (m) −t ₂ s (m)}]
... (Formula 16)
It can be expressed.

here,
N2 ₁ (m, n) ≈N2 ₂ (m, n)
... (Formula 17)
It is.

Thereafter, the correction for reducing the noise component N1 described above may be performed on the corrected signal SS _C (m, n) in the same manner as in each of the above-described embodiments.

In other words, in the present embodiment, correction for reducing the noise component N2 is performed, and correction for reducing the noise component N1 is performed on the image data obtained by the correction using the difference in time t _eff between adjacent rows. . According to this embodiment, the same effects as those of the first embodiment described above can be obtained, and the noise component N2 caused by FPN or the like can be further reduced.

(5. Fifth embodiment)
In the first embodiment described above, it has been described that the signal component S0 between the adjacent rows is substantially equal to each other in a region where the change in the signal component is small (for example, a region other than a portion forming an outline in the image). In this case, the signal component S0 (m, n) at the sensor s (m, n) in the mth row and the nth column and the signal at the sensor s (m + 1, n) in the (m + 1) th row and the nth column. With the component S0 (m + 1, n), S0 (m, n) ≈S0 (m + 1, n) holds. In the first embodiment, based on this, the calculation unit 52 of the processing unit 50 calculates a correction coefficient for reducing the noise component N1 from the difference in signal values between adjacent rows.

  However, in the radiographic image obtained by radiography of the subject, the change in signal component is large in the portion where the contour is formed, and therefore S0 (m, n) ≠ S0 (m + 1, n). Therefore, the correction coefficient cannot be obtained properly by the calculation method by the calculation unit 52. In this case, when the calculation result from the calculation unit 52 does not satisfy the predetermined condition, the correction unit 51 does not adopt the calculation result as a correction coefficient, or omits correction processing for a portion that does not satisfy the predetermined condition. Can do.

  Fig.11 (a) is an enlarged plot figure in the said outline part of the signal value of the sensor s. In the figure, the horizontal axis represents the column number of the sensor array 110, and the vertical axis represents the signal value. In the plot in the figure, a black circle plot indicates a signal value from each sensor s in the k-th row, and a white circle plot indicates a signal value from each sensor s in the k + 1-th row. FIG. 11A shows that in the column after the y1th column, the radiation amount detected by the sensor s is larger and the signal value is larger than in the column before the y1th column.

  FIG. 11B is a plot diagram showing the difference between adjacent columns of the signal value of the sensor s as a rhombus plot. According to FIG. 11B, since the signal value is larger in the column after the y1 column than in the column before the y1 column, the plot value is specifically increased in the y1 column.

  Here, the processing unit 50 determines, for example, whether or not the difference plot value is smaller than a predetermined value. Alternatively, the processing unit 50 may further include a determination unit (not shown) for performing the determination. When the plot value is larger than the predetermined value, the plot value may not be used for calculating the correction coefficient, or the correction process may be omitted for that portion.

  Here, the mode of determining whether or not the difference between adjacent columns of the signal value of the sensor s is smaller than a predetermined value is exemplified, but the determination may be made based on statistics of calculation results. For example, the determination unit determines whether or not the variation amount of the difference between adjacent columns of the signal value is smaller than a predetermined value (for example, the calculated difference is ± 3σ with the standard deviation of the variation from the median as σ. Or less) may be determined.

  As described above, according to the present embodiment, when the calculation result from the calculation unit 52 does not satisfy the predetermined condition, the correction unit 51 of the processing unit 50 does not adopt the calculation result as a correction coefficient, or about that portion. The correction process can be omitted. Therefore, according to the present embodiment, correction for reducing the noise component N1 can be selectively performed on an appropriate portion of the image data. Although FIG. 9 illustrates the case where the signal value greatly changes in the column direction of the sensor array 110, the same applies to the case of the row direction.

(6. Other)
As mentioned above, although several suitable embodiment was described, this invention is not limited to these, The part may be changed according to the objective etc., and each embodiment may be combined. .

  Further, the present invention can also be achieved by executing the above-described embodiments by a program or software by a computer. Specifically, for example, a program that realizes the functions of the above-described embodiments is supplied to the system or apparatus via a network or various storage media. The computer of the system or apparatus (or CPU, MPU, etc.) then reads and executes the program.

  IA: radiation imaging apparatus, 10: imaging unit, 11: light source, 20: drive unit, 50: processing unit, 51: correction unit, 52: calculation unit, 110: sensor array, 120: sensor drive unit, 130: signal readout Part, AO: accumulation operation, RO: reading operation.

Claims (13)

A radiation imaging apparatus comprising: a plurality of sensors arranged so as to form a plurality of rows and a plurality of columns on a substrate; and a drive unit for driving the plurality of sensors in units of rows,
The drive unit is
A first operation of initializing the plurality of sensors while selecting the plurality of rows in a first order before radiation irradiation to the plurality of sensors is started;
After the irradiation of the radiation is finished, the plurality of rows are selected in the second order, and the sensor is selected in the first order between the sensor in each row and the sensor in the adjacent row. Performing a second operation of driving the plurality of sensors and reading signals from the plurality of sensors so as to cause a difference in time until they are selected in order;
The radiation imaging apparatus includes:
Of the signals read from the plurality of sensors in the second operation, a signal from the sensor in the first row, which is one row in the plurality of rows, and a second row adjacent to the first row. To correct a signal from at least one sensor of the first row and the second row based on a signal from the sensor and the time difference between the first row and the second row. The correction part of
A radiation imaging apparatus, further comprising: a light source for irradiating the plurality of sensors with light before the first operation. The light source is a plan view with respect to the upper surface of the substrate.
Being arranged outside the area where the plurality of sensors are arranged,
Being arranged on the outer periphery of the region; and
The radiation imaging apparatus according to claim 1, wherein the radiation imaging apparatus satisfies at least one of being arranged so as to overlap the region. A division unit for dividing the plurality of sensors into two or more groups in units of rows so that two adjacent rows are in different groups;
The drive unit performs both selection of the plurality of rows in the first order and selection of the plurality of rows in the second order, and the division unit causes the plurality of sensors to be grouped into two or more groups. The radiation imaging apparatus according to claim 2, wherein the radiation imaging apparatus is divided for each line. The number of groups divided by the dividing unit is 2,
The radiation according to claim 3, wherein the driving unit performs the selection and the initialization in the first order, and the selection and the driving in the second order, respectively, in an interlaced manner. Imaging device. A division unit for dividing the plurality of sensors into two or more groups in units of rows so that two adjacent rows are in different groups;
The drive unit selects one of the plurality of rows in the first order and the selection of the plurality of rows in the second order, and the plurality of sensors are grouped into two or more groups by the dividing unit. The radiation imaging apparatus according to claim 2, wherein the radiation imaging apparatus is divided for each line. The number of groups divided by the dividing unit is 2,
The drive unit performs the selection in the first order and the initialization in one of an interlace method and a progressive method, and performs the selection in the second order and the drive in the other. The radiation imaging apparatus according to claim 5. Correction information for the correction unit to correct a signal from the at least one sensor includes a signal from the sensor in the first row, a signal from the sensor in the second row, the first row, and the first row. An arithmetic unit that calculates based on the time difference between the two rows;
The radiographic imaging according to claim 1, wherein the correction unit corrects a signal from the at least one sensor based on the correction information calculated by the calculation unit. apparatus. The calculation unit is configured such that the signal value difference between the signal from the sensor in the first row and the signal from the sensor in the second row, and the time difference between the first row and the second row. The radiation imaging apparatus according to claim 7, wherein the correction information is calculated based on: The plurality of sensors are divided into two or more regions,
The radiation imaging apparatus according to claim 7, wherein the calculation unit calculates the correction information for each region. The plurality of sensors are divided into two or more regions in any one of at least every two rows, at least every two columns, and every unit region formed by at least two rows and at least two columns. The radiation imaging apparatus according to claim 9. A determination unit that determines whether the correction information satisfies a predetermined condition;
11. The radiation according to claim 7, wherein the correction unit does not perform the correction when the determination unit determines that the correction information does not satisfy a predetermined condition. Imaging device. A control unit;
The controller is
A first control for controlling the light source and the drive unit so that the plurality of sensors are initialized by the first operation;
A second control for controlling a radiation source for irradiating radiation so that radiation is irradiated to the plurality of sensors for a predetermined time after the first control;
The third control for controlling the driving unit so that signals are read from the plurality of sensors by the second operation after the second control. 12. The radiation imaging apparatus according to Item. A radiation imaging apparatus driving method comprising: a plurality of sensors arranged to form a plurality of rows and a plurality of columns; and a driving unit for driving the plurality of sensors in units of rows,
Irradiating the plurality of sensors with light, and then initializing the plurality of sensors irradiated with the light while selecting the plurality of rows in a first order by the driving unit;
After the irradiation of radiation to the plurality of sensors is completed after the first step, the plurality of rows are selected in the second order by the driving unit, and a sensor in each row and a sensor in the next row are selected. A second step of reading the signals from the plurality of sensors by driving the plurality of sensors so that there is a difference in time between the selection in the first order and the selection in the second order. ,
Of the signals read from the plurality of sensors in the second step, a signal from a sensor in a first row that is one row in the plurality of rows, and a second row adjacent to the first row. Based on a signal from a sensor and the difference in time between the first row and the second row, a signal for correcting a signal from at least one sensor of the first row and the second row is corrected. And a step of driving the radiation imaging apparatus.

Images