JP2007228268A - Signal processing apparatus and imaging apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a signal processing apparatus capable of allowing the difference in signal level between channels to contribute to improvement correction accuracy, and to provide an imaging apparatus. <P>SOLUTION: In the signal processing apparatus for processing an imaging signal to be output from an imaging means capable of executing multi-channel reading, a plurality of groups of monitoring pixels are provided in the vicinity of a boundary of separated imaging regions, detection values corresponding to the signal level difference between channels are calculated on the basis of the imaging signal to be acquired from the pixels in the group of monitoring pixels for each group of monitoring pixels, and a first correction value is calculated from a large number of calculated detection values. In this case, the distribution of the detection values is referred to, and the distribution is used to calculate the first correction value. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a signal processing device used in an imaging device such as a digital camera, and more particularly to correct an output signal level difference between channels that occurs when an imaging device is divided into a plurality of regions. The present invention relates to a signal processing device.

  In order to improve the readout speed of a CCD (Charge Coupled Devices) during moving image shooting or high-speed continuous shooting, a two-channel CCD in which the CCD is divided into left and right parts has been developed. In a 2-channel CCD, a CDS (Correlated Double Sampling) / AGC (Auto Gain Control) circuit and an AD (Analog Digital) conversion circuit are provided for each channel.

  When a two-channel CCD is used, the left and right signal levels (output signal levels) are caused by the accuracy difference and gain difference of the CDS / AGC circuit provided for the left and right imaging areas, and the characteristic difference of the AD conversion circuit. The problem that does not match. Even if a fixed means for matching the left and right output signals is provided, it is unlikely that the left and right CDS / AGC circuits have the same temperature characteristics, so the left and right signal levels (output signal levels) are matched. It is difficult to let

  As a method of correcting the signal level difference (output signal level difference) between channels, a flat image is taken by emitting a light emitting diode (LED) inside the shutter with the shutter closed before shooting, and the captured image Based on the above, a method for correcting a signal level difference between channels has already been developed.

JP 2003-18472 A JP 2002-142158 A JP 2002-125149 A JP 2002-218186 A JP 2002-320142 A JP 2003-259224 A

  As described above, various techniques for correcting a signal level difference between channels have been studied, but a signal processing technique having higher correction accuracy is desired.

  In view of the above problems, an object of the present invention is to provide a signal processing device and an imaging device that can contribute to improvement in correction accuracy of a signal level difference between channels.

  In order to achieve the above object, a first signal processing device according to the present invention includes an imaging device in which an imaging region is divided into a plurality of divided imaging regions and an electric signal is read from each of the divided imaging regions, and the divided imaging. In the signal processing apparatus that processes the imaging signal output from the imaging unit that is separately assigned to each area and outputs the electrical signal as an imaging signal, the plurality of divided imaging areas Among the two divided imaging areas adjacent to each other, the first and second divided imaging areas are set as the first and second output channels corresponding to the first and second divided imaging areas, respectively. And the imaging signal output from at least one of the second output channels is a correction target signal, and the imaging signal is in a plurality of signal ranges having different signal levels. In a similar case, a plurality of monitoring pixel groups each including a plurality of pixels are provided in the vicinity of the boundary between the first and second divided imaging regions, and each monitoring pixel group is obtained from the pixels in the monitoring pixel group. Detection value calculation means capable of calculating a detection value according to a signal level difference between the imaging signals from the first and second output channels based on the imaging signal, and a basis for calculating the detection value The signal range to which each detection value belongs is determined based on the imaging signal, and the signal processing device further determines, for each signal range, the signal level difference based on the detection value belonging to the signal range. First correction value calculating means for calculating a first correction value for reducing the signal, and, for each signal range, the correction target signal belonging to the signal range using the first correction value corresponding to the signal range. To correct Correction means for correcting the signal level difference, wherein the first correction value calculation means refers to the distribution of the detection values belonging to the signal range for each signal range and uses the distribution to The first correction value is calculated.

  In order to achieve the above object, the second signal processing apparatus according to the present invention includes an imaging device in which an imaging region is divided into a plurality of divided imaging regions and an electric signal is read from each divided imaging region, and the divided imaging. In the signal processing apparatus that processes the imaging signal output from the imaging unit that is separately assigned to each area and outputs the electrical signal as an imaging signal, the plurality of divided imaging areas Among the two divided imaging areas adjacent to each other, the first and second divided imaging areas are set as the first and second output channels corresponding to the first and second divided imaging areas, respectively. And the imaging signal output from at least one of the second output channels is a correction target signal, and the imaging signal is in a plurality of signal ranges having different signal levels. In a similar case, a plurality of monitoring pixel groups each including a plurality of pixels are provided in the vicinity of the boundary between the first and second divided imaging regions, and each monitoring pixel group is obtained from the pixels in the monitoring pixel group. Detection value calculation means capable of calculating a detection value according to a signal level difference between the imaging signals from the first and second output channels based on the imaging signal, and a basis for calculating the detection value The signal range to which each detection value belongs is determined based on the imaging signal, and the signal processing device further determines, for each signal range, the signal level difference based on the detection value belonging to the signal range. First correction value calculating means for calculating a first correction value for reducing the signal, and, for each signal range, the correction target signal belonging to the signal range using the first correction value corresponding to the signal range. To correct Correction means for correcting the signal level difference, wherein the first correction value calculation means calculates, for each signal range, an average value of the detection values belonging to the signal range and within a predetermined limit range. The first correction value is calculated by calculating, and in the latest calculation of the first correction value, the limit range is corrected based on the average value calculated in the past.

  In order to achieve the above object, the third signal processing device according to the present invention includes an imaging device in which an imaging region is divided into a plurality of divided imaging regions and an electric signal is read from each of the divided imaging regions, and the divided imaging. In the signal processing apparatus that processes the imaging signal output from the imaging unit that is separately assigned to each area and outputs the electrical signal as an imaging signal, the plurality of divided imaging areas Among the two divided imaging areas adjacent to each other, the first and second divided imaging areas are set as the first and second output channels corresponding to the first and second divided imaging areas, respectively. And the imaging signal output from at least one of the second output channels is a correction target signal, and the imaging signal is in a plurality of signal ranges having different signal levels. In a similar case, a plurality of monitoring pixel groups each including a plurality of pixels are provided in the vicinity of the boundary between the first and second divided imaging regions, and each monitoring pixel group is obtained from the pixels in the monitoring pixel group. Detection value calculation means capable of calculating a detection value according to a signal level difference between the imaging signals from the first and second output channels based on the imaging signal, and a basis for calculating the detection value The signal range to which each detection value belongs is determined based on the imaging signal, and the signal processing device further determines, for each signal range, the signal level difference based on the detection value belonging to the signal range. First correction value calculating means for calculating a first correction value for reducing the signal, and, for each signal range, the correction target signal belonging to the signal range using the first correction value corresponding to the signal range. To correct Correction means for correcting the signal level difference, wherein the first correction value calculation means is obtained from the pixels in the plurality of monitoring pixel groups and belongs to the signal range for each signal range. The first correction value is calculated by referring to a signal level distribution of the imaging signal and using the signal level distribution.

  In order to achieve the above object, an imaging apparatus according to the present invention includes any of the imaging means described above and the signal processing apparatus described above.

  According to the present invention, the signal level difference between channels can contribute to the improvement of correction accuracy.

<< First Embodiment >>
Hereinafter, embodiments of the present invention will be specifically described with reference to the drawings. In each drawing, the same parts are denoted by the same reference numerals. FIG. 1 is a schematic block diagram of a digital camera according to the first embodiment of the present invention. The digital camera in FIG. 1 can shoot moving images and still images.

  In FIG. 1, reference numeral 1 denotes a CCD (Charge Coupled Devices) configured by arranging a plurality of pixels in a matrix. Each pixel generates an electrical signal corresponding to the amount of incident light. The electrical signal generated in each pixel is read out by a CDS (Correlated Double Sampling) / AGC (Auto Gain Control) circuit 2 or 4 that performs correlated double sampling and auto gain control.

  The imaging area of the CCD 1 is divided into two areas on the left and right. The CCD consisting of the divided left area (first divided imaging area) is called the left CCD 1a, and the CCD consisting of the divided right area (second divided imaging area) is called the right CCD 1b.

  The electrical signal read from the left CCD 1 a by the CDS / AGC circuit 2 is converted into a digital signal by an AD (Analog Digital) conversion circuit 3 and then sent to the image composition circuit 6. The electrical signal read from the right CCD 1 b by the CDS / AGC circuit 4 is converted into a digital signal by the AD conversion circuit 5 and then sent to the image composition circuit 6.

  Thus, the output channel 14 having the CDS / AGC circuit 2 and the AD conversion circuit 3 is assigned to the left CCD 1a, and the output channel 15 having the CDS / AGC circuit 4 and the AD conversion circuit 5 is assigned to the right CCD 1b. Is assigned. A signal output from the output channel 14 (that is, a digital signal output from the AD conversion circuit 3) is referred to as a signal CH1. A signal output from the output channel 15 (that is, a digital signal output from the AD conversion circuit 5) is referred to as a signal CH2.

  The image synthesis circuit 6 synthesizes the signals CH1 and CH2 of both channels and converts them into one system signal. The signal obtained by the image composition circuit 6 is sent to a correction value calculation circuit 8 and a level difference correction circuit 9 via a clamp circuit 7 for making the black level constant.

  The correction value calculation circuit 8 is a circuit for calculating a correction value for correcting a difference between signal levels output from both output channels (inter-channel signal level difference). The level difference correction circuit 9 is a circuit for correcting the inter-channel signal level difference using the correction value calculated by the correction value calculation circuit 8. Even when uniform light is incident on the entire imaging region of the CCD 1, a difference is generated between the signal level of the signal CH1 and the signal level of the signal CH2. This difference is called an inter-channel signal level difference. The inter-channel signal level difference is caused by a characteristic difference between the output channels 14 and 15.

  The output signal from the level difference correction circuit 9 is sent to the YRGB generation circuit 10. The YRGB generation circuit 10 generates Y, R, G, and B signals based on the output signal from the level difference correction circuit 9. The Y signal generated by the YRGB generation circuit 10 is sent to the encoder 13 via the Y process circuit 11. The R, G, and B signals generated by the YRGB generation circuit 10 are sent to the encoder 13 via the C process circuit 12. The encoder 13 performs signal compression using a compression method such as MPEG4 (Moving Picture Experts Group phase 4) or JPEG (Joint Photographic Experts Group), and a signal (data) obtained by the compression is not illustrated, for example. Stored in a memory (including an external recording medium).

  As the amount of light incident on each pixel increases, the analog signal level (signal intensity) of the electrical signal generated at each pixel increases. As the analog signal level increases, the digital signal level (that is, the digital value) output from the AD conversion circuit 3 or 5 also increases. “Signal level” and “signal value” are synonymous.

  The digital signals output from the AD conversion circuits 3 and 5 are defined as signals CH1 and CH2, respectively, but are sent to the correction value calculation circuit 8 and the level difference correction circuit 9 via the image synthesis circuit 6 and the clamp circuit 7. Signals CH1 and CH2 are also simply referred to as signals CH1 and CH2.

  FIG. 2 is a partial block diagram of the digital camera of FIG. 1 including an internal block diagram of the correction value calculation circuit 8. The correction value calculation circuit 8 corrects the inter-channel signal level difference based on the detection value calculation unit 21 that calculates a large number of detection values reflecting the inter-channel signal level difference, and the detection value calculated by the detection value calculation unit 21. An actual correction value calculation unit (first correction value calculation unit) 22 for calculating an actual measurement correction value (first correction value) for use, and a predetermined initial correction value (second correction value) optimized under a predetermined temperature condition ) In advance, and a final correction value to be supplied (applied) to the level difference correction circuit 9 based on the actual correction value and the initial correction value. An application correction value calculation unit 24 that calculates (application correction value), and an update correction table 25 that sequentially updates and stores the correction value (application correction value) calculated by the application correction value calculation unit 24. Yes.

  In the present embodiment (and other embodiments described later), the correction value is calculated as a correction value to be applied to the signal CH1.

  The level difference correction circuit 9 corrects the signal CH1 using the correction value stored in the update correction table 25. The corrected signal CH1 and the signal CH2 not subjected to the correction process by the level difference correction circuit 9 are supplied to the YRGB generation circuit 10. As will be described in detail later, correction processing using the initial correction value stored in the initial correction table 23 as a correction value is performed at a predetermined reference time such as when the digital camera of FIG. 1 is turned on. At the time of power-on and immediately after that, the initial correction value is strongly involved in the correction, but after a sufficient time has passed, the correction value follows the actual correction value.

  The inter-channel signal level difference varies depending on the accumulated charge amount of each pixel proportional to the signal level as indicated by a solid line 60 in FIG. 3 (the broken line 61 will be described later). For this reason, a plurality of correction values are provided as correction values used for correcting the signal CH1. FIG. 4 is a schematic diagram showing a plurality of correction values stored in the update correction table 25 with the horizontal axis representing the signal level of CH1. As an example, the signal level is divided into 50 signal ranges according to the magnitude of the signal level.

  For example, a signal having a digital value of 0 to 10 belongs to the first signal range (SR1 in FIG. 4), a signal having a value of 11 to 30 belongs to the second signal range, and so on. For example, the dynamic range of the signal is divided into 50 (see FIG. 5).

  Then, one correction value is assigned to each signal range. Similarly, one actual correction value and one initial correction value are assigned to each signal range. FIG. 5 shows how correction values, actual correction values, and initial correction values are assigned to each signal range. The correction value assigned to the m-th signal range is calculated based on the “actual correction value and initial correction value” assigned to the m-th signal range (m is an arbitrary integer from 1 to 50).

  FIG. 6 is a flowchart for explaining the operation of the detection value calculation unit 21. The operations of steps S1 to S4 shown in FIG. Further, the operation of the detection value calculation unit 21 will be described with reference to FIG.

  In the upper side of FIG. 7, pixels in one horizontal line of the CCD 1, that is, a total of eight pixels P <b> 1 to P <b> 8 of four pixels on both sides of the boundary L between the left CCD 1 a and the right CCD 1 b are shown. The pixels P1, P2, P3, P4, P5, P6, P7, and P8 are arranged in this order from left to right. Signal values (signal levels) corresponding to the pixels P1, P2, P3, P4, P5, P6, P7, and P8, which are given to the correction value calculation circuit 8 and the level difference correction circuit 9, are p1, p2, p3, respectively. Represented by p4, p5, p6, p7 and p8. For convenience of explanation, in FIG. 7, the signal values p1, p3, p5, and p7 are shifted and displayed with respect to the signal values p2, p4, p6, and p8.

  First, in step S1, three pixels of the same color that are continuous in the horizontal direction are selected from these eight pixels. The CCD 1 employs a single plate method, and each pixel of the CCD 1 is provided with, for example, one of red (R), green (G), and blue (B) color filters (not shown). . The same color filter pixel means a pixel provided with a color filter of the same color among the pixels of the CCD 1. A case where the color filters are arranged in a so-called Bayer arrangement is taken as an example. In this case, the pixels P2, P4, P6, and P8 form one set of the same color filter pixels, and the pixels P1, P3, P5, and P7 form another set of the same color filter pixels.

  Therefore, there are four combinations of (P1, P3, P5), (P2, P4, P6), (P3, P5, P7) and (P4, P6, P8) as combinations of three selected pixels. Each of the pixel groups composed of the three pixels functions as a monitoring pixel group for calculating a detection value.

  In step S2, detection values for each of the four combinations are calculated. A detection value calculation method for the group (P2, P4, P6) will be described. Among the three pixels P2, P4, and P6, from the signal values p2 and p4 of the two pixels P2 and P4 belonging to one divided imaging region (in this example, the left CCD 1a), the other divided imaging region (in this example, The signal value of one pixel P6 belonging to the right CCD 1b) is predicted. That is, the signal value of the pixel P6 is predicted from the slope of the line segment connecting p2 and p4 in the coordinate system having the pixel position on the horizontal axis and the signal value (signal level) on the vertical axis. This predicted signal value is represented by p6 '.

  The actual signal value of the pixel P6 (hereinafter, “actual signal value” may be referred to as “actual value”) p6 and the predicted signal value of the pixel P6 (hereinafter, “predicted signal value”). Is detected as “predicted value”), and a detection value d6 corresponding to the set of (P2, P4, P6) is calculated.

  The detected value is expressed by a difference value between the actual value and the predicted value or a ratio between the actual value and the predicted value. In the following, unless otherwise stated, it is assumed that the detected value is represented by a ratio. When attention is paid to only one detection value, the detection value represents a gain to be multiplied with the signal CH1 in order to eliminate the inter-channel signal level difference.

  As described above, the correction value obtained according to the detection value is handled as a correction value to be applied to the signal CH1. For this reason, when the predicted value is the predicted value of the pixel in the right CCD 1b, the actual value / the predicted value is calculated as the detected value. That is, for example, d6 = p6 ÷ p6 ′. When the predicted value is the predicted value of the pixel in the left CCD 1a, the predicted value ÷ actual value is calculated as the detected value.

  Similar processing is performed for the other combinations (P1, P3, P5), (P3, P5, P7) and (P4, P6, P8). That is, the predicted values p5 ′, p3 ′, and p4 ′ of the signal values of the pixels P5, P3, and P4 are calculated, and (P1, P3, P5), (P3, P5, P7), and (P4, P6, P8) Detection values d5 (= p5 / p5 ′), d3 (= p3 ′ / p3) and d4 (= p4 ′ / p4) corresponding to the set are calculated.

  The processes in steps S1 and S2 are repeated for each horizontal line of an image obtained by photographing with the CCD 1. For example, the processes in steps S1 and S2 are performed for each horizontal line for all horizontal lines constituting the image. That is, for example, in step S3 following step S2, it is determined whether or not the processing in steps S1 and S2 has been performed for all horizontal lines. If not performed for all horizontal lines, another horizontal line is selected. (Step S4) Return to Step S1. On the other hand, if it is determined that the process has been performed for all horizontal lines, the process of FIG. 6 ends.

  When the number of horizontal lines of the image (CCD1) is n (n is a natural number), (n × 4) detection values are calculated from one frame image. The calculated detection value is used for calculation of the actual correction value by the actual correction value calculation unit 22. However, as will be understood from the following description, it is not always necessary to calculate (n × 4) detection values in each frame, and calculation of actual correction values (or correction values in the correction table 25 for update) It is sufficient to calculate at least the detection values necessary for (update) in each frame.

  Next, the operation of the actual measurement correction value calculation unit 22 will be described. The actual measurement correction value calculation unit 22 calculates 50 actual correction values based on the plurality of detection values calculated by the detection value calculation unit 21. The calculation of 50 actual correction values is performed, for example, for each frame. However, as will be understood from the following description, it is not always necessary to calculate 50 actual correction values in each frame, and only the actual correction values necessary for updating the correction values in the update correction table 25 are stored in each frame. It is sufficient to calculate at a minimum.

  In order to calculate each actual measurement correction value, first, a signal range to which each detection value calculated by the detection value calculation unit 21 belongs is determined. That is, it is determined whether each detection value belongs to the first, second,..., 50th signal range. This determination is made based on the signal value that is the basis for calculating each detected value. In addition, this determination is performed by the actual measurement correction value calculation unit 22, for example, but may be performed by the detection value calculation unit 21.

  Each detection value is calculated based on the three signal values. Among the three signal values, each detection value is determined using a signal value corresponding to a pixel in the left CCD 1a as a determination value for signal range classification. Specify the signal range to which the value belongs. When there are two signal values corresponding to the pixels in the left CCD 1a, for example, the signal range to which each detection value belongs is specified using the signal value of the pixel closer to the boundary L as the determination value. When a determination value corresponding to a certain detection value belongs to the m-th signal range (m is an arbitrary integer from 1 to 50), it is determined that the detection value belongs to the m-th signal range.

  For example, the signal range to which the detection value d6 for the set (P2, P4, P6) belongs is specified by using the signal value p4 as the determination value. When the value of p4 is “12” and the second signal range is “11-30” as described above (see FIG. 5), since “12” belongs to the second signal range, the detection is performed. The value d6 is determined to belong to the second signal range.

  However, various methods can be applied as a method for determining which signal range the detected value belongs to. For example, a value obtained by using any one or more of the three signal values from which the detection value is calculated is set as a determination value for signal range classification, and each detection is performed using the determination value. It is also possible to determine which signal range the value belongs to.

  For example, any one of p2, p4, and p6 may be adopted as a determination value for specifying the signal range to which the detection value d6 belongs, or two or more values of p2, p4, and p6 May be adopted, or a value obtained by adding p6 ′ to them may be adopted.

  An actual measurement correction value corresponding to the m-th signal range is calculated using the detected value determined to belong to the m-th signal range. However, when calculating the actual correction value corresponding to the m-th signal range, the validity of each detected value is determined based on the signal value or the detected value that is the basis for calculating the detected value, and the detection is determined to be effective. The actual measurement correction value corresponding to the m-th signal range is calculated using only the value (the detected value determined to be invalid is ignored).

This valid / invalid determination method will be exemplified by focusing on the detection value d6 for the set (P2, P4, P6). The actual measurement correction value calculation unit 22 determines that the detection value d6 is valid only when both of the following formulas (1) and (2) are satisfied. In other cases, the reliability of the detection value d6 is affected by noise and the like. It is determined that the detected value is low and the detected value d6 is determined to be invalid.
| P4-p2 | ≦ p4 × K ₁ (1)
th1 ≦ d6 ≦ th2 (2)

Here, K ₁ is arbitrarily set within a range of 0 <K ₁ ≦ 1. Further, th1 and th2 are threshold values, and th1 <th2 is established. The thresholds th1 and th2 will be described in detail later.

  Various methods can be adopted as a method of calculating the actual correction value by the actual correction value calculation unit 22.

  First, the most basic method that can be adopted will be described as “basic measurement correction value calculation method”. In the basic actual correction value calculation method, for each signal range, an average value of detection values for one frame determined to be valid is calculated, and the average value is adopted as the actual correction value.

  For example, it is assumed that in one frame, detection values that belong to the first signal range and are determined to be valid are “1.11”, “1.09”, and “1.10”. In this case, the actual correction value corresponding to the first signal range is “1.10” from “(1.11 + 1.09 + 1.10) /3=1.10”. This actual correction value represents a gain to be multiplied with the signal CH1 in order to eliminate (ideally) a signal level difference between channels. (As described above, after a sufficient time has elapsed since the power was turned on, the correction value follows the actual correction value). Similarly, actual correction values are calculated for other signal ranges.

  Next, the internal configuration and operation of the applied correction value calculation unit 24, the update method of the correction value stored in the update correction table 25, and the correction method by the level difference correction circuit 9 will be described. FIG. 8 is a block diagram illustrating an example of an internal configuration of the applied correction value calculation unit 24. FIG. 8 shows a filter 30 including multipliers 31 and 34, an adder 32, and a register (delay circuit) 33. The filter 30 is an IIR (Infinite Impulse Response) filter.

The multiplier 31 outputs a multiplication result of the input value and the coefficient (1-K ₂ ). The adder 32 outputs the addition result of the output value of the multiplier 31 and the output value of the multiplier 34. The register 33 holds the output value of the adder 32 and outputs the output value with a delay of unit time. The multiplier 34 outputs a multiplication result between the output value and the coefficient K ₂ of the register 33. The delay time (that is, the unit time) by the register 33 is a time corresponding to, for example, 50 frames, but can be appropriately changed during the operation. The coefficient K ₂ is a positive number less than 1.

  The filter 30 functions as a low-pass filter that smoothes and outputs the input value to the filter 30 (that is, the input value to the multiplier 31) in the time direction. The output value of the filter 30 (the output value of the register 33) is sent to the update correction table 25 as the correction value (applied correction value) calculated by the applied correction value calculation unit 24.

  It is also possible to calculate a total of 50 correction values corresponding to the first to 50th signal ranges by using one filter 30. However, for simplification of explanation, it is now possible to calculate a single signal range. A case will be described in which one filter 30 is provided to calculate one corresponding correction value. In this case, the applied correction value calculation unit 24 is configured to include 50 filters 30.

  The mth filter 30 calculates a correction value (applied correction value) corresponding to the mth signal range based on the “initial correction value and actual correction value” corresponding to the mth signal range (m is , Any integer from 1 to 50). 1 is turned on and each unit shown in FIG. 1 is activated (this time point is referred to as “reference time point”), the stored value and output value of the register 33 of the mth filter 30; In addition, the correction value (applied correction value) corresponding to the m-th signal range stored in the update correction table 25 is the initial correction value corresponding to the m-th signal range stored in the initial correction table 23. It is supposed to match.

  The actual correction value corresponding to the m-th signal range calculated in a certain frame after the reference time becomes an input value to the multiplier 31 of the m-th filter 30. For this reason, the output value of the register 33 of the m-th filter 30 that coincides with the initial correction value corresponding to the m-th signal range at the reference time point gradually increases from the reference time point to the m-th signal range. It follows the actual measurement correction value corresponding to.

  The output value of the register 33 of the mth filter 30 is stored in the update correction table 25 as a correction value corresponding to the mth signal range. When the output value of the register 33 of the mth filter 30 is updated, the correction value corresponding to the mth signal range in the update correction table 25 is also updated.

  Level difference correction circuit 9 receives signal CH1. The signal CH1 includes a plurality of imaging signals corresponding to electrical signals generated at each pixel in the left CCD 1a. The level difference correction circuit 9 determines to which signal range each imaging signal constituting the signal CH1 belongs. The level difference correction circuit 9 then stores the latest value corresponding to the m-th signal range stored in the update correction table 25 for the signal value of the imaging signal determined to belong to the m-th signal range. Multiply the correction value (applied correction value). The signal value obtained by the multiplication is sent to the YRGB generation circuit 10 as a signal value in which the inter-channel signal level difference is corrected.

  For example, when the signal value of a certain imaging signal is “15” and the second signal range is “11-30” as described above (see FIG. 5), the imaging signal falls within the second signal range. Judged to belong. When the correction value (applied correction value) corresponding to the second signal range is “1.2”, a signal having a value of “18” (= 15 × 1.2) is converted to the YRGB generation circuit 10. Sent to.

  Although the correction process performed by the level difference correction circuit 9 is expressed as “multiplication”, any other method capable of obtaining the same result or approximate result as this multiplication can be employed. For example, a value obtained by subtracting the median value of the mth signal range from the multiplication result of the median value of the mth signal range and the correction value corresponding to the mth signal range is calculated, and the value (subtraction result) is calculated. May be added to each imaging signal belonging to the mth signal range to correct the inter-channel signal level difference.

  When the operating temperature of the CCD 1 and the output channels 14 and 15 and the ambient temperature of the digital camera of FIG. 1 change, the relationship between the signal level and the inter-channel signal level difference changes from a solid line 60 to a broken line 61 in FIG. Therefore, it is necessary to appropriately change the correction value according to the actual operating temperature or the like. However, if the correction value is suddenly changed, the signal level of the signal CH1 is also sharply corrected. Will occur.

  However, if configured as described above, immediately after the power is turned on, an initial correction value optimized under a predetermined temperature condition is used as a correction value to be actually applied, and the inter-channel signal level difference correction process is performed. Gradually, the correction value used for the correction process is changed to a value suitable for the actual operating temperature (that is, the actually measured correction value). For this reason, the signal level difference between channels is corrected smoothly, and contributes particularly to the improvement of the quality of moving images.

  Further, the level difference correction circuit 9 corrects only the signal CH1 to adjust the signal level of the signal CH1 to the signal level of the signal CH2. However, the level difference correction circuit 9 corrects only the signal CH1. The same effect may be obtained by performing correction.

  In this case, the level difference correction circuit 9 receives both signals CH1 and CH2 as correction target signals, and determines to which signal range each imaging signal constituting the signals CH1 and CH2 belongs. Then, the latest correction value (applied correction value) corresponding to the m-th signal range stored in the update correction table 25 is stored as the signal value of each imaging signal determined to belong to the m-th signal range. Use to correct.

  For example, the same effect as described above can be obtained by reflecting (about) 50% of the correction value in the signals CH1 and CH2. Considering a certain signal range, it is assumed that the latest correction value (applied correction value) is J (for example, 1.2). In this case, the signal value of the imaging signal of the signal CH1 is multiplied by “(1+ (J−1) / 2)”, and the signal value of the imaging signal of the signal CH2 is “(1+ (J−1) / 2”. ) ÷ J ”. Then, the signal values of the signals CH1 and CH2 obtained by the multiplication are sent to the YRGB generation circuit 10 as signal values in which the inter-channel signal level difference is corrected.

  The above-described method of correcting the inter-channel signal level difference by reflecting the correction value on the signals CH1 and CH2 by 50%, for example, also applies to various modifications (including other embodiments) described later. Applicable.

  In addition, the actually measured correction value obtained sequentially is reflected in the correction value that is actually used for the correction process via a low-pass filter in the time direction (in this embodiment, the filter 30 is exemplified). The time constant of the low-pass filter may be changed immediately after the input) and thereafter.

For example, immediately after the reference time point, the time constant of the low-pass filter is set as τ _1, and the time constant is moved toward a larger τ ₂ with the passage of time (for example, as the shooting progresses). The time constant is fixed at τ ₂ (where τ ₁ <τ ₂ ). When the filter 30 shown in FIG. 8 is used, the coefficient K ₂ may be changed with the passage of time.

  After the reference time point, the correction value (applied correction value) is gradually changed toward the value corresponding to the latest measured correction value, but it can be corrected by changing the time constant as described above. The speed at which the value is directed to the actual correction value (or the response speed of the correction value with respect to the change in the actual correction value calculated one after another) is fast immediately after the power is turned on, and gradually decreases thereafter.

  That is, immediately after the power is turned on, the update correction table 25 is quickly updated to the correction table suitable for the current temperature situation, so that the correction content of the inter-channel signal level difference is suitable for the current temperature situation. It is quickly changed to something. On the other hand, since it is considered that the temperature change is moderate after a while since the power is turned on, it is considered that the change in the actual correction value is small. In addition, a large change in the actually measured correction value after a while has passed since power-on is often due to the influence of noise or the like. Therefore, at a point in time after the power is turned on, the amount of change in the correction value in the update correction table 25 is relatively large, and a sharp change in the correction value due to noise or the like is reduced.

  Further, a specific example of the update timing of the correction value in the update correction table 25 will be shown with reference to FIG. In the first frame after the reference time point, the correction value corresponding to the first signal range is updated for the first time according to the output value of the filter 30 corresponding to the first signal range. In the second frame to be visited next, the correction value corresponding to the second signal range is updated for the first time according to the output value of the filter 30 corresponding to the second signal range.

  Thus, one correction value is updated in one frame. In this case, the output value of the filter 30 (register 33) corresponding to the mth signal range is updated only in the frame in which the correction value corresponding to the mth signal range is updated. That is, for example, in the first frame, only the output value of the filter 30 (register 33) corresponding to the first signal range is updated, and the filter 30 (register 33) corresponding to the second to 50th signal ranges is updated. ) Output value is not updated.

  After the 50 update values are updated for the first time using the 50 frame period, the 50th correction value is updated for the second time using the 51st to 100th frames. After the second update is completed, after a pause period, the third update of 50 correction values is performed. For example, the pause period is 20 frame periods, and the third update is performed using the 121st to 170th frames. After the third update, the fourth update of 50 correction values is performed after another pause period. For example, the other pause period is 40 frame periods, and the fourth update is performed using the 211st to 260th frames.

  In this way, even when the correction value update interval is initially shortened and gradually widened, the speed at which the correction value moves toward the actual correction value is high immediately after the power is turned on, and gradually thereafter. Become slow.

  Moreover, although the example which employ | adopts the cyclic | annular circuit (filter 30) like FIG. 8 as an applied correction value calculation part 24 was shown, by adding the past actual correction value to the newest actual correction value, the above-mentioned is demonstrated. You may make it acquire the effect | action similar to the filter 30. FIG. For example, the average value or the weighted average value of the latest actual measurement correction value, the previous actual measurement correction value, and the previous actual measurement correction value (...) belonging to the m-th signal range should be updated this time. You may employ | adopt as a correction value which belongs to the m-th signal range. As a result, the latest correction value follows the latest actual correction value while being influenced by the past actual correction value.

  By the way, there is a point which should be improved in the calculation method of the actual measurement correction value by the actual measurement correction value calculation part 22 mentioned above as "basic actual measurement correction value calculation method". Consider this point to be improved.

A curve 63 in FIG. 10 shows a distribution for one frame of detection values belonging to a certain signal range. The sample value corresponding to the peak 64 in this distribution, that is, the mode value, represents the ideal value G _{id of the} gain to be multiplied to the signal CH1 in order to eliminate the inter-channel signal level difference.

It is not necessary to consider the influence of noise or the like at all, and when the average value of all detection values for one frame is calculated, the average value is considered to be substantially the same as the ideal value _Gid . However, there is actually an influence of noise and the like. Therefore, when the detection value d6 corresponding to the set (P2, P4, P6) in FIG. 7 is considered, the detection value d6 is invalidated if neither of the above formulas (1) and (2) is satisfied. There is a need.

  On the other hand, since the output channels 14 and 15 are created so that the signal level difference between channels is as close to zero as possible, the thresholds th1 and th2 for determining the validity / invalidity of the detected value are centered at “1.0”. Set to That is, th1 = 1.0−Δd and th2 = 1.0 + Δd are set (where 0 <Δd <1, and Δd = 0.2, for example).

If the signal level difference between channels is zero without correction, the sample value corresponding to the peak of the distribution of detected values matches (substantially) “1.0”. However, in practice, as shown in FIG. 10, the sample value corresponding to the peak 64 is usually different from “1.0”. On the other hand, detection values less than the threshold th1 and detection values exceeding the threshold th2 need to be excluded from the calculation of the average value as invalid detection values. As a result, the average value G _r of the detection values in the range from th1 to th2 deviates from the mode value (ideal value G _id ), and there arises a problem that the value to be detected cannot be detected with high accuracy.

  As a calculation method of the actual correction value by the actual correction value calculator 22 that solves such a problem, first, second, and third actual correction value calculation methods will be exemplified below.

  Also in the first to third actual correction value calculation methods, a detection value that does not satisfy the same conditions as the above formulas (1) and (2) is invalid and is ignored in calculating the actual correction value. A range from the threshold th1 to the threshold th2 is referred to as a limited range.

[First measurement correction value calculation method]
First, the first actual correction value calculation method by the actual correction value calculator 22 will be described. In the first actual correction value calculation method, a distribution of detection values for one frame belonging to the mth signal range (hereinafter referred to as “detection value distribution α”) is referred to. FIG. 11 shows the detected value distribution α. Then, a detection value (sample value) corresponding to the peak of the detection value distribution α, that is, the mode value G ₁ is searched, and the mode correction value corresponding to the m-th signal range is searched using the mode value G _1. Is calculated. Of course, this calculation is performed for each of m = 1 to 50. Even if the mode value of the distribution exists outside the limit range (th1 to th2), the mode value is naturally ignored. Hereinafter, for simplification of description, the description of the first actually measured correction value calculation method will be continued focusing on only one signal range.

Most simply, the mode G ₁ itself of the detection value distribution α is adopted as the actual measurement correction value. Referring to FIG. 10, this is equivalent to adopting the ideal value G _id as the actual correction value. In addition, an average or a weighted average of the detection value (sample value) having the second highest frequency in the detection value distribution α and the mode value G ₁ (the detection value having the highest frequency) is adopted as the actual measurement correction value. Also good. Further, the actually measured correction value may be calculated in consideration of the detection value having the third highest frequency in the detection value distribution α (the fourth highest detection value,...). Furthermore, the actual measurement correction value may be calculated in consideration of the detection value (sample value) adjacent to the mode value G ₁ .

  If the first actual correction value calculation method is used, the value that is originally desired to be detected is detected with high accuracy, and the correction accuracy of the inter-channel signal level difference is improved.

[Second actual correction value calculation method]
Next, the second actual correction value calculation method by the actual correction value calculator 22 will be described with reference to FIGS. 12 (a) and 12 (b). In the second actual correction value calculation method, the limit range of the detection value is sequentially changed according to the detection value. That is, the threshold values th1 and th2 are sequentially changed according to the detection value. Hereinafter, for simplification of description, the description of the second actually measured correction value calculation method will be continued focusing on only one signal range, but the same processing is performed for each of the 50 signal ranges.

  Similarly to the first actual correction value calculation method, the second actual correction value calculation method refers to the distribution of detection values (detection value distribution α) for one frame belonging to a certain signal range. FIG. 12A shows the detected value distribution α.

Then, in the detection value distribution α, an average value G ₂ of detection values (sample values) within a limited range from th1 to th2 is calculated. Initially, as described above, th1 and th2 are set as th1 = 1.0−Δd and th2 = 1.0 + Δd, and the limit range at this point is indicated by reference numeral 66 in FIG. . The calculated average value G ₂ is sent to the applied correction value calculation unit 24 as an actual measurement correction value corresponding to the signal range.

  When one or two or more frame periods have elapsed and the actual correction value corresponding to the signal range is to be calculated next, the distribution of detection values for the latest one frame belonging to the same signal range (hereinafter referred to as “detection”). Value distribution β) ”. FIG. 12B shows this detected value distribution β. The detected value distribution β can be considered to be typically the same as the detected value distribution α.

Then, the detection value distribution beta, but calculates an average value G ₃ of the detection value within the limited range from th1 to th2, this time, is th1 and th2 depending on the average value G ₂ obtained in the previous calculation Be changed. Specifically, th1 = G ₂ −Δd and th2 = G ₂ + Δd. In other words, the average value G ₃ of the detected values within the restricted range from th1 (= G ₂ −Δd) to th2 (= G ₂ + Δd) is calculated, and this average value G ₃ corresponds to the signal range. And sent to the applied correction value calculation unit 24 as the latest actual measurement correction value. A limit range for calculating the average value G ₃ is indicated by reference numeral 68 in FIG.

Further, when calculating the next actual correction value, th1 and th2 are changed in the same manner as described above, and detection within the limited range from th1 (= G ₃ -Δd) to th2 (= G ₃ -Δd) is performed. The average value is the latest actual correction value.

  If the second actual correction value calculation method is used, the center of the limit range is changed so as to (substantially) coincide with the value that should be detected, so that the correction accuracy of the inter-channel signal level difference is improved.

[Third measurement correction value calculation method]
Next, a third actual correction value calculation method by the actual correction value calculation unit 22 will be described.

  First, a description will be given regarding FIG. FIG. 14 shows a detection value calculation area 75 near the boundary L between the left CCD 1a and the right CCD 1b. In FIG. 14, the detection value calculation area 75 is surrounded by a two-dot chain line. In each horizontal line of the CCD 1, a total of 8 pixels, 4 pixels on both sides of the boundary L, are specified. This specified collection of pixels forms a detection value calculation area 75. If the number of horizontal lines of the CCD 1 is n (n is a natural number), the detection value calculation area 75 is composed of (n × 8) pixels. The eight pixels in the detection value calculation area 75 on each horizontal line are the pixels P1 to P8 described with reference to FIG. A group of a total of (n × 4) pixels arranged in the detection value calculation area 75 and in the left CCD 1 a is referred to as a pixel group 77.

  In the third actual correction value calculation method, a distribution of signal values (signal levels) corresponding to a total of (n × 4) pixels in the pixel group 77 is referred to.

  First, the signal value (signal level) corresponding to each pixel in the pixel group 77 provided to the correction value calculation circuit 8 is classified into one of the above-described first to fifty signal ranges. For example, when a certain signal value (signal level) is “12” and the second signal range is “11-30” as described above (see FIG. 5), “12” is the second. Therefore, it is determined that the signal value (signal level) belongs to the second signal range.

  Then, for each signal range, a distribution of signal values (signal levels) of pixels in the pixel group 77 (hereinafter referred to as “signal level distribution”) is formed. Each signal level distribution is formed using only “signal values of pixels in the pixel group 77” belonging to each signal range.

  For example, the signal values of the pixels in the pixel group 77 are “8”, “9”, “10”, “11”, “12”, “13”, “14”, and “50”, respectively ( Let us consider a case where n = 2) and the first and second signal ranges are “0 to 10” and “11 to 30” as described above (see FIG. 5). In this case, the signal level distribution for the first signal range is formed using “8”, “9”, and “10” as sample values, and the signal level distribution for the second signal range is “11”, “12”, “13” and “14” are formed as sample values. “50” is a sample value of the signal level distribution for a signal range different from the first and second signal ranges.

  Hereinafter, for simplification of description, the description of the third actually measured correction value calculation method will be continued focusing on only one signal range, but the same processing is performed for each of the 50 signal ranges. It is assumed that the second signal range is focused, and the second signal range is “11 to 30”.

FIG. 13 shows a signal level distribution 70 for the second signal range. The actual correction value calculation unit 22 obtains a signal value (signal level) corresponding to the peak of the signal level distribution 70, that is, the mode value F ₁ of the signal level distribution 70. A pixel that gives the same signal value (signal level) as the mode value (signal level) F ₁ is selected from the pixel group 77. The selected pixel is hereinafter referred to as “selected pixel”. Subsequently, a detection value calculated using the selected pixel is specified, and an average value of the specified detection values is calculated as an actual correction value.

Specific numerical examples will be given. For simplicity of explanation, it is assumed that the CCD 1 has only one horizontal line. As described with reference to FIG. 7, the pixels P1 to P8 and the corresponding signal values (signal levels) p1 to p8 are defined for the horizontal line. Further, it is assumed that the mode F ₁ is 20, and p1, p2, p3, and p4 are “22”, “20”, “12”, and “20”, respectively. In this case, the pixels P2 and P4 are selected pixels, and the detection value calculated using the selected pixels is a set of detection values d6 and (P4, P6, P8) corresponding to the set of (P2, P4, P6). The corresponding detection value d4 is obtained. Accordingly, in this case, the actual correction value to be calculated by the actual correction value calculator 22 is (d4 + d6) / 2.

  Further, as can be seen from the above numerical examples, some detection values (d3 and d5 in the above numerical examples) may not be used in calculating the actual correction value. Therefore, when the third actual correction value calculation method is employed, the detection value calculation unit 21 and the actual correction value calculation unit 22 cooperate to omit calculation of detection values that are not used. . That is, first, after determining which detection value is necessary to calculate the actual correction value, only the necessary detection value may be calculated.

A pixel that gives a signal value (signal level) close to the mode value (signal level) F ₁ may also be included in the selected pixel. For example, when the signal value of the pixel in the pixel group 77 is represented by “p”, all the pixels that satisfy “(F ₁ −ΔF) ≦ p ≦ (F ₁ + ΔF)” may be selected pixels. Here, ΔF is an integer of 1 or more. In the above numerical example, when ΔF = 2, the pixel P1 is also included in the selected pixel, and the detection value d5 corresponding to the set of (P1, P3, P5) is also used for calculating the actual correction value. become. In this case, the actually measured correction value is (d4 + d5 + d6) / 3.

A curve 71 in FIG. 13 represents the relationship between the signal level and the inter-channel signal level difference within one signal range. Since each signal range has a signal level width, the signal level difference between channels changes in accordance with the signal level even in the same signal range. Considering this, in the third actual correction value calculation method, the actual correction value is calculated using the detection value calculated based on the signal value corresponding to the mode value F ₁ of the signal level distribution 70. As a result, an actual correction value corresponding to the actual inter-channel signal level difference (DIF in FIG. 13) to be corrected is calculated, and the correction accuracy of the inter-channel signal level difference is improved.

  I will explain this further. The image disturbance (visual influence) due to the presence of the signal level difference between channels becomes noticeable when the image corresponding to the vicinity of the boundary L of the CCD 1 is a flat image (for example, a so-called solid image). When a flat image is given in the vicinity of the boundary L, the signal level distribution becomes a distribution having a clear peak like the signal level distribution 70 of FIG. 13, and in such a case, the third actually measured correction value calculation method is particularly effective. To work. Even if the signal level distribution is a broad distribution without a clear peak, the adverse effect of adopting the third actual correction value calculation method is slight (or none). This is because, in an image giving such a distribution, image disturbance (visual effect) due to the presence of a signal level difference between channels is small.

[Detection value]
Although the case where the detection value (for example, d6) is expressed by the ratio of the actual measurement value (for example, p6) and the predicted value (for example, p6 ′) has been mainly described, as described above, It can also be expressed by a difference value between the actual value and the predicted value. For example, when attention is paid to the set of (P2, P4, P6), the detection value d6 is calculated by the equation: d6 = p6-p6 ′ (see FIG. 7). Similarly, the detection values d5 (= p5-p5 ′) and d3 (= p3′−p3) corresponding to the set of (P1, P3, P5), (P3, P5, P7) and (P4, P6, P8) And d4 (= p4′−p4) are calculated.

  Even if the detected value is expressed as a difference value between the actual value and the predicted value, the detected value includes information indicating the signal level difference between channels. Therefore, the detected value is expressed as a ratio between the actual value and the predicted value. It is possible to realize the same correction processing (correction processing for the signal level difference between channels) as expressed.

  When the detected value is expressed by a difference value between the actual value and the predicted value, for example, the actual correction value calculated by the actual correction value calculation unit 22 and the initial correction value stored in the initial correction table 23 are associated therewith. The correction value (applied correction value) calculated by the applied correction value calculation unit 24 is also expressed in the form of a difference value. For example, the level difference correction circuit 9 corrects the inter-channel signal level difference by adding a value corresponding to the correction value (applied correction value) expressed in the form of the difference value to the given signal CH1. Processing will be performed. Further, even when the correction value is expressed in the form of a difference value, the correction value is expressed in accordance with the correction value (applied correction value) expressed in the form of the difference value, as in the case where the correction value is expressed by a ratio. The correction process for the inter-channel signal level difference may be performed by reflecting 50% of the value in the signals CH1 and CH2.

  In addition, although an example is shown in which four monitoring pixel groups for calculating detection values are selected for each horizontal line and four detection values are calculated for each horizontal line, the present invention is not limited to this. For example, in each horizontal line, the same color filter pixel is selected from 8 pixels P1 to P8 located in the detection value calculation area 75 (see FIG. 14). When the color filters are arranged in a Bayer array, the same color filter pixel combinations to be selected are (P1, P3, P5, P7) and (P2, P4, P6, P8).

In each combination, the detection value is the ratio or difference between the average value of the signal values of two pixels belonging to one divided imaging area and the average value of the signal values of two pixels belonging to the other divided imaging area. . For example, when focusing on the combination (monitor pixel group) of (P2, P4, P6, P8), “signal values p2 and P2 of two pixels P2 and P4 belonging to one divided imaging region (in this example, the left CCD 1a) and Ratio R ₁ or difference value R ₂ between “average value of p4” and “average value of signal values p6 and p8 of two pixels P6 and P8 belonging to the other divided imaging region (right CCD 1b in this example)” The detection value is used as a detection value calculated by the detection value calculation unit 21. The ratio R ₁ and the difference value R ₂ are represented by the following formulas (3) and (4), respectively.
R ₁ = ((p6 + p8) / 2) / ((p2 + p4) / 2) (3)
R ₂ = ((p6 + p8) / 2) − ((p2 + p4) / 2) (4)

  Similarly, detection values corresponding to combinations of (P1, P3, P5, P7) are also calculated. The same process is repeated for each horizontal line. When the number of horizontal lines of the image (CCD1) is n (n is a natural number), (n × 2) detection values are calculated from one frame image.

In this case as well, the signal range to which the detected value belongs is specified based on the signal value that is the basis for calculating the detected value. For example, the signal range to which the detection value (R ₁ or R ₂ ) corresponding to (P2, P4, P6, P8) belongs is “the signal range to which the signal value p4 (, p2, p6 or p8) belongs” or “ It is the same as “a signal range to which a value calculated using two or more of the signal values p2, p4, p6, and p8 belongs”.

  “By providing the applied correction value calculation unit 24, the correction value used for correcting the signal level difference between channels is gradually changed from the initial correction value to a value suitable for the actual operating temperature (ie, the actual correction value). In the above example, the function of the applied correction value calculation unit 24 can be omitted. That is, the actual correction value calculated by the actual correction value calculation unit 22 can be used as it is as a correction value (applied correction value) used for correcting the signal level difference between channels. This is equivalent to short-circuiting the input and output of the filter 30 in FIG. Even in this case, for example, by using any one of the first to third calculation methods, it is possible to expect improvement in the correction accuracy of the inter-channel signal level difference.

  In FIG. 1, the correction value calculation circuit 8 and the level difference correction circuit 9 are provided at the subsequent stage of the clamp circuit 7. It is possible to change to an arbitrary position after 15. However, when providing them before the clamp circuit 7, the clamp circuit 7 needs to perform processing according to the correction value used by the level difference correction circuit 9.

<< Second Embodiment >>
As is clear from the above description, the detection value calculation unit 21 in FIG. 2 calculates the detection value based on signals obtained from the pixels in the detection value calculation region 75 in FIG. Therefore, if a noise component is mixed in a signal obtained from a pixel in the detection value calculation area 75, an accurate correction value cannot be calculated. In consideration of this, a modification may be made in which a spatial smoothing circuit is inserted before the correction value calculation circuit 8. An embodiment with this modification will be described as a second embodiment of the present invention.

  FIG. 15 shows a schematic block diagram of a digital camera according to the second embodiment of the present invention. 15, the same parts as those in FIG. 1 are denoted by the same reference numerals. The digital camera shown in FIG. 15 matches the digital camera shown in FIG. 1 except that a spatial smoothing circuit 16 is added. Therefore, only the function of the spatial smoothing circuit 16 will be described.

  In FIG. 15, the spatial smoothing circuit 16 is provided before the correction value calculation circuit 8. The spatial smoothing circuit 16 receives the signals CH1 and CH2 sent from the clamp circuit 7 and performs a smoothing process in the spatial direction on these signals. The smoothed signals CH1 and CH2 are sent to the correction value calculation circuit 8. The correction value calculation circuit 8 in FIG. 15 uses the signals CH1 and CH2 after the smoothing process sent from the spatial smoothing circuit 16 as signals CH1 and CH2 from the clamp circuit 7 in the first embodiment (FIG. 1). As in the first embodiment, the correction value is calculated and output.

  The spatial smoothing circuit 16 performs vertical LPF (low-pass filter) processing, for example, as the smoothing processing. This vertical LPF process is performed on each pixel in the detection value calculation area 75.

For example, attention is paid to nine pixels arranged in the vertical direction in the region 76 of FIG. The area 76 is included in the detection value calculation area 75. FIG. 16 shows an enlarged view of the region 76. Nine pixels arranged in the vertical direction in the region 76 are referred to as A ₀ , A ₁ , A ₂ , A ₃ , A ₄ , A ₅ , A ₆ , A ₇ and A ₈ from the lower side. The central pixel A ₄ is a target pixel for vertical LPF processing. Signal values (signal levels) corresponding to the pixels A _{0 to} A ₈ from the clamp circuit 7 are S _{0 to} S ₈ , respectively. In this case, the signal value S ₄ ′ after the vertical LPF processing of the target pixel A ₄ is expressed by, for example, the following expression (5). Here, V _kq (q = 0, 1,..., 8) is a value in the range of 0-7.

The signal value S ₄ ′ after the vertical LPF processing is sent to the correction value calculation circuit 8. The correction value calculation circuit 8 handles the signal value S ₄ ′ as the signal value of the pixel A ₄ . A similar vertical LPF process is performed on the signal value of each pixel in the detection value calculation area 75, and each signal value after the vertical LPF process is sent to the correction value calculation circuit 8. The correction value calculation circuit 8 in FIG. 15 calculates and outputs a correction value based on the signal value after vertical LPF processing of each pixel in the detection value calculation area 75, as in the first embodiment.

  As an example of the smoothing process performed by the spatial smoothing circuit 16, the vertical LPF process has been described as an example, but a horizontal LPF process or a two-dimensional LPF process may be used instead of the vertical LPF process.

  In the vertical LPF process, as described above, the signal value of the target pixel is subjected to the LPF process using the signal values of the peripheral pixels arranged in the vertical direction of the target pixel. On the other hand, in the horizontal LPF process, the signal value of the target pixel is subjected to the LPF process using the signal value of the peripheral pixels arranged in the horizontal direction of the target pixel.

  That is, in the horizontal LPF process, attention is paid to a horizontal region including a plurality of pixels arranged in the horizontal direction, and the center pixel of the plurality of pixels is set as a target pixel for the horizontal LPF process. Then, using the signal value corresponding to each pixel in the horizontal region, the signal value corresponding to the target pixel is smoothed as in the vertical LPF process. For example, each pixel in the detection value calculation area 75 is a target pixel for horizontal LPF processing.

  In the two-dimensional LPF process, the LPF process is performed on the signal value of the target pixel using the signal values of the peripheral pixels arranged in the vertical and horizontal directions of the target pixel. In the two-dimensional LPF process, it is also possible to use signal values of peripheral pixels arranged in an oblique direction of the target pixel. For example, each pixel in the detection value calculation area 75 is a target pixel for the two-dimensional LPF process.

  Note that in any of the vertical LPF processing, the horizontal LPF processing, and the two-dimensional LPF processing, if the diameter of the LPF becomes too large, the probability that an edge component is included in the signal value used for the LPF increases. Therefore, conversely, the diameter of the LPF is appropriately set so as not to cause noise on the detected value.

<< Third Embodiment >>
In the first and second embodiments, the CCD 1 in which the imaging region is divided into two left and right regions is used as an example. The technical matters described in the first and second embodiments are as follows. Of course, the present invention can be applied even when a CCD whose imaging region is divided into three or more is used. As an example, an embodiment using a CCD whose imaging region is divided into four will be described as a third embodiment according to the present invention.

  FIG. 17 is a schematic partial block diagram of a digital camera according to the third embodiment of the present invention. In FIG. 17, the same parts as those in FIG. 1 are denoted by the same reference numerals, and redundant description of the same parts is omitted.

  In FIG. 17, reference numeral 41 denotes a CCD (Charge Coupled Devices) configured by arranging a plurality of pixels in a matrix. Each pixel generates an electrical signal corresponding to the amount of incident light. The imaging area of the CCD 41 is divided into four areas in the left-right direction. The four divided areas are referred to as areas 41a, 41b, 41c and 41d from the left.

  The CDS / AGC circuits 42a, 42b, 42c and 42d have the same function as the CDS / AGC circuit 2 (and 4) in FIG. 1, and the AD conversion circuit circuits 43a, 43b, 43c and 43d are the same as those in FIG. This has the same function as the AD conversion circuit 3 (and 5).

The electrical signal generated in each pixel in the area 41a is sent to the image composition circuit 46 via the output channel 44a having the CDS / AGC circuit 42a and the AD conversion circuit 43a. The electric signal generated in each pixel in the area 41b is sent to the image composition circuit 46 via the output channel 44b having the CDS / AGC circuit 42b and the AD conversion circuit 43b.
The electrical signal generated in each pixel in the region 41c is sent to the image composition circuit 46 via the output channel 44c having the CDS / AGC circuit 42c and the AD conversion circuit 43c. The electrical signal generated in each pixel in the region 41d is sent to the image composition circuit 46 via the output channel 44d having the CDS / AGC circuit 42d and the AD conversion circuit 43d.

  The image synthesis circuit 46 synthesizes the output signal CHa of the output channel 44a, the output signal CHb of the output channel 44b, the output signal CHc of the output channel 44c, and the output signal CHd of the output channel 44d, and converts them into one system signal. . The signal obtained by the image composition circuit 46 is sent to a correction value calculation circuit 48 and a level difference correction circuit 49 via a clamp circuit 47 for making the black level constant.

  The correction value calculation circuit 48 and the level difference correction circuit 49 have functions according to the correction value calculation circuit 8 and the level difference correction circuit 9 of FIG.

  For example, the correction value calculation circuit 48 and the level difference correction circuit 49 correct the inter-channel signal level difference related to the signals CHa and CHb by treating the signals CHa and CHb as CH1 and CH2 in the first embodiment, respectively. By treating the signals CHc and CHd as CH1 and CH2 in the first embodiment, respectively, the signal level difference between channels involved in the signals CHc and CHd is corrected. Further, by treating the “signals CHa and CHb” and the “signals CHc and CHd” as the CH1 and CH2 in the first embodiment, the inter-channel signals related to the “signals CHa and CHb” and the “signals CHc and CHd”, respectively. Correct the level difference.

  The signal from which the inter-channel signal level difference is corrected from the level difference correction circuit 49 is sent to the YRGB generation circuit 10. The subsequent stage of the YRGB generation circuit 10 is the same as in the first embodiment.

  Similarly to the modification of the first embodiment to the second embodiment, a spatial smoothing circuit can be provided in the previous stage of the correction value calculation circuit 48 in the third embodiment.

<< Deformation, etc. >>
It should be noted that various methods other than the above-described methods can be employed as a method for obtaining a value corresponding to the signal level difference between channels (in the above-described embodiments, the detection value). For example, a plurality of monitoring areas MA ₁ , MA ₂ ,... Are provided near the boundary L of the CCD 1 as shown in FIG. FIG. 19 shows an enlarged view of one monitoring area (in FIG. 19, MA ₁ is taken as an example). Each monitoring area includes (3 × 3) 9 pixels located on the left side of the boundary L and (3 × 3) 9 pixels located on the right side of the boundary L.

  Then, by comparing the average value of the signal values of 9 pixels located on the left side of the boundary L with the average value of the signal values of 9 pixels located on the right side of the boundary L for each monitoring area, A detection value corresponding to the difference is calculated. Also in this case, as described with reference to FIG. 5 and the like, each detection value is classified into a plurality of signal ranges. On this basis, for example, for each signal range, an average value of detection values belonging to the signal range is employed as an actual measurement correction value corresponding to the signal range.

  Then, a predetermined initial correction value is used as a correction value for correcting a signal level difference between channels at the time of power-on, while the correction value is measured after that by using the filter 30 of FIG. Change gradually to a value according to the correction value. As a result, it is possible to smoothly correct the signal level difference between channels.

  Further, all of the above description can be applied not only when shooting moving images but also when shooting still images.

  In each of the embodiments described above, the CCD is taken as an example of the image sensor, but a C-MOS (Complementary Metal Oxide Semiconductor) image sensor can also be used as the image sensor.

  Further, the CCD 1 (or 41) and the output channels 14 and 15 (or 44a to 44d) function as imaging means (see FIGS. 1 and 17). A signal processing apparatus that processes a signal output from the imaging means includes at least a correction value calculation circuit 8 (or 48) and a level difference correction circuit 9 (or 49).

  Further, the correction value calculation circuit 8 and the level difference correction circuit 9 (FIG. 1 and the like), or the correction value calculation circuit 48 and the level difference correction circuit 49 (FIG. 17) may be implemented by hardware or by a microcomputer and the microcomputer. This can be realized by a program that determines the operation of the above or a combination thereof.

  The present invention can be widely applied to imaging apparatuses having a photographing function, such as a digital still camera and a digital video camera. In particular, it is suitable for an imaging apparatus capable of shooting a moving image.

1 is a block diagram of a schematic configuration of a digital camera according to a first embodiment of the present invention. FIG. 2 is a partial block diagram of the digital camera of FIG. 1 including an internal block diagram of the correction value calculation circuit of FIG. 1. It is a figure which shows the relationship between the charge amount of accumulation | storage of each pixel of CCD of FIG. 1, and the signal level difference between channels. FIG. 3 is a schematic diagram showing a plurality of correction values stored in the update correction table of FIG. 2 with the horizontal axis as the signal level of CH1. It is a figure which shows a mode that the correction value calculated by the correction value calculation circuit of FIG. 2 or stored, the actual measurement correction value, and the initial correction value are allocated to the several signal range. It is a flowchart for demonstrating operation | movement of the detection value calculation part of FIG. It is a schematic diagram for demonstrating operation | movement of the detection value calculation part of FIG. It is a block diagram which shows the example of an internal structure of the legal correction value calculation part of FIG. FIG. 3 is a diagram for explaining an example of update timing of correction values in the update correction table of FIG. 2. It is a figure for demonstrating the characteristic of the calculation method (basic measurement correction value calculation method) of the measurement correction value employable by the measurement correction value calculation part of FIG. It is a figure for demonstrating the characteristic of the calculation method (1st measured correction value calculation method) of the measured correction value employable in the measured correction value calculation part of FIG. It is a figure for demonstrating the characteristic of the calculation method (2nd measured correction value calculation method) of the measured correction value employable in the measured correction value calculation part of FIG. It is a figure for demonstrating the characteristic of the calculation method (3rd measured correction value calculation method) of the measured correction value employable in the measured correction value calculation part of FIG. It is a figure which shows pixel arrangement | positioning in CCD of FIG. It is a schematic block diagram of a digital camera according to a second embodiment of the present invention. FIG. 15 is an enlarged view of a certain area in the CCD of FIG. 14. It is a schematic block diagram of a digital camera according to a third embodiment of the present invention. It is a figure for demonstrating the modification with respect to each embodiment. It is a figure for demonstrating the modification with respect to each embodiment.

Explanation of symbols

1 CCD
1a Left CCD
1b Right CCD
2, 4 CDS / AGC circuit 3, 5 AD conversion circuit 6 Image composition circuit 7 Clamp circuit 8 Correction value calculation circuit 9 Level difference correction circuit 16 Spatial smoothing circuit 21 Detected value calculation unit 22 Actual correction value calculation unit 23 Initial correction table 24 applied correction value calculation unit 25 correction table for update

Claims (4)

An imaging device in which an imaging region is divided into a plurality of divided imaging regions and an electrical signal is read from each divided imaging region, and a plurality of outputs that are separately assigned to each of the divided imaging regions and output the electrical signal as an imaging signal A signal processing apparatus for processing the imaging signal output from the imaging means having a channel,
Two divided imaging areas adjacent to each other among the plurality of divided imaging areas are defined as first and second divided imaging areas, and the output channels corresponding to the first and second divided imaging areas are first and second, respectively. As an output channel, when the imaging signal output from at least one of the first and second output channels is a correction target signal, and the imaging signal is classified into a plurality of signal ranges having different signal levels,
A plurality of monitoring pixel groups each including a plurality of pixels are provided in the vicinity of the boundary between the first and second divided imaging regions, and the imaging signal obtained from the pixels in the monitoring pixel group is provided for each monitoring pixel group. Based on detection value calculation means capable of calculating a detection value according to a signal level difference between the imaging signals from the first and second output channels,
The signal range to which each detection value belongs is determined based on the imaging signal that is a basis for calculation of the detection value,
The signal processing device further includes:
First correction value calculation means for calculating a first correction value for reducing the signal level difference for each signal range based on the detection value belonging to the signal range;
Correction means for correcting the signal level difference for each signal range by correcting the correction target signal belonging to the signal range using the first correction value corresponding to the signal range;
The first correction value calculation means is provided for each signal range.
A signal processing apparatus that refers to a distribution of the detection values belonging to the signal range and calculates the first correction value using the distribution. An imaging device in which an imaging region is divided into a plurality of divided imaging regions and an electrical signal is read from each divided imaging region, and a plurality of outputs that are separately assigned to each of the divided imaging regions and output the electrical signal as an imaging signal A signal processing apparatus for processing the imaging signal output from the imaging means having a channel,
Two divided imaging areas adjacent to each other among the plurality of divided imaging areas are defined as first and second divided imaging areas, and the output channels corresponding to the first and second divided imaging areas are first and second, respectively. As an output channel, when the imaging signal output from at least one of the first and second output channels is a correction target signal, and the imaging signal is classified into a plurality of signal ranges having different signal levels,
A plurality of monitoring pixel groups each including a plurality of pixels are provided in the vicinity of the boundary between the first and second divided imaging regions, and the imaging signal obtained from the pixels in the monitoring pixel group is provided for each monitoring pixel group. Based on detection value calculation means capable of calculating a detection value according to a signal level difference between the imaging signals from the first and second output channels,
The signal range to which each detection value belongs is determined based on the imaging signal that is a basis for calculation of the detection value,
The signal processing device further includes:
First correction value calculation means for calculating a first correction value for reducing the signal level difference for each signal range based on the detection value belonging to the signal range;
Correction means for correcting the signal level difference for each signal range by correcting the correction target signal belonging to the signal range using the first correction value corresponding to the signal range;
The first correction value calculation means is provided for each signal range.
Calculating the first correction value by calculating an average value of the detection values belonging to the signal range and within a predetermined limit range;
In the latest calculation of the first correction value, the limit range is corrected based on the average value calculated in the past. An imaging device in which an imaging region is divided into a plurality of divided imaging regions and an electrical signal is read from each divided imaging region, and a plurality of outputs that are separately assigned to each of the divided imaging regions and output the electrical signal as an imaging signal A signal processing apparatus for processing the imaging signal output from the imaging means having a channel,
Two divided imaging areas adjacent to each other among the plurality of divided imaging areas are defined as first and second divided imaging areas, and the output channels corresponding to the first and second divided imaging areas are first and second, respectively. As an output channel, when the imaging signal output from at least one of the first and second output channels is a correction target signal, and the imaging signal is classified into a plurality of signal ranges having different signal levels,
A plurality of monitoring pixel groups each including a plurality of pixels are provided in the vicinity of the boundary between the first and second divided imaging regions, and the imaging signal obtained from the pixels in the monitoring pixel group is provided for each monitoring pixel group. Based on detection value calculation means capable of calculating a detection value according to a signal level difference between the imaging signals from the first and second output channels,
The signal range to which each detection value belongs is determined based on the imaging signal that is a basis for calculation of the detection value,
The signal processing device further includes:
First correction value calculation means for calculating a first correction value for reducing the signal level difference for each signal range based on the detection value belonging to the signal range;
Correction means for correcting the signal level difference for each signal range by correcting the correction target signal belonging to the signal range using the first correction value corresponding to the signal range;
The first correction value calculation means is provided for each signal range.
The first correction value is calculated using the signal level distribution with reference to the signal level distribution of the imaging signal obtained from the pixels in the plurality of monitoring pixel groups and belonging to the signal range. Signal processing device. An imaging apparatus comprising the imaging means and the signal processing apparatus according to claim 1.