JPH03224372A - Electronic camera - Google Patents

Abstract

PURPOSE:To effectively input an object with a wide dynamic range as a latent image input by providing a means repeating the pickup of an object by an image pickup element and the readout of a picture signal for plural number of times, a means reducing a fixed pattern noise signal from the picture signal and a means accumulating sequentially the picture signal with the fixed pattern noise signal reduced therefrom. CONSTITUTION:A fixed pattern noise(FPN) component including a background noise component generated in an image pickup element 1, an amplifier 2 and a signal processing circuit 3 in advance is stored in an FPN data ROM 42 as an A/D-converted data. Then in the case of pickup, after an input signal is A/D-converted, an FPN data of a relevant picture element is read from the FPN data ROM 42 and subtracted sequentially from the input signal at a subtractor 41 to accumulate the signal subtracted by FPN component after the stage of an adder 5. Since a random noise component is decreased by the succeeding addition, the S/N is improved and a picture signal with excellent wide dynamic range is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an electronic camera that can effectively image and input a subject with a wide dynamic range and display the image on an image monitor.

[Prior Art] Recently, various solid-state image sensors such as CCD image sensors and MO3 type image sensors have been developed, and electronic cameras that electronically image a subject using these solid-state image sensors have been developed.
For example, various types of cameras have been developed, such as video cameras and electronic still cameras. This type of electronic camera records an image signal (video signal) of a subject image electronically captured and input using a solid-state image sensor on a so-called magnetic recording medium such as a videotape or floppy disk, or an IC memory card. TV this
It is used for an image element by an image monitor such as a television receiver.

By the way, the dynamic range for a video signal that can be displayed on an image monitor such as the TV receiver is at most 40 dB.
That's about it. For this reason, the imaging capability (dynamic range) of solid-state imaging devices for boat fishing is often set at about 50 dB.

In comparison, the dynamic range of a subject to be imaged by the solid-state image sensor (electronic camera) often reaches about 80 dB. That is, the dynamic range (signal level width) of the subject often greatly exceeds the dynamic range that can be displayed by an image monitor and inputted by a solid-state image sensor.

[Problem to be solved by the invention] However, when an image of such a wide dynamic range subject is electronically input as it is using the solid-state image sensor, the solid-state image sensor is affected by high signal level areas (high brightness areas). Saturation occurs and a so-called blown-out highlight condition occurs. On the other hand, if an attempt is made to suppress this saturation, it is undeniable that portions with low signal levels (low brightness portions) will be underexposed, resulting in a so-called blackish state. In other words, there is a problem in that signal components that do not fall within the dynamic range of the solid-state image sensor or image monitor cannot be imaged and input, nor can they be displayed.

In this way, in the conventional technology, there are many objects that have a wider dynamic range than the dynamic range of the solid-state image sensor, and if the image is electronically input as is with the solid-state image sensor, high-level parts or There was a problem that subject information in low-level areas was lost.

In other words, even though the subject has a wide dynamic range, if it is imaged and input electronically using a solid-state image sensor,
The dynamic range of the subject is limited by the dynamic range of the solid-state image sensor and image monitor, resulting in an image signal (video signal) with a limited dynamic range.
There was a problem that I could only get it. Further, in particular, solid-state image sensing devices inherently have the problem of generating vertical stripes on the display screen due to their fixed pattern noise.

The present invention was made in consideration of these circumstances, and its purpose is to effectively capture and input a subject having a dynamic range that exceeds the dynamic range of a solid-state image sensor, and to eliminate so-called blown-out highlights and blown-out shadows. Another object of the present invention is to provide a highly practical electronic camera that can be used for image display on an image monitor without causing problems such as vertical stripes due to fixed pattern noise.

"Means for Solving the Problems" The present invention relates to an electronic camera that electronically captures and inputs images of a subject using a solid-state image sensor and displays the image on an image monitor, and the present invention relates to an electronic camera that uses a solid-state image sensor to electronically capture and input an image of a subject and display the image on an image monitor. A solid-state image sensor that can be read outside the device at a relatively high speed is used, and the solid-state image sensor captures an image of a subject and reads out the image signal at high speed multiple times within the image display cycle. The fixed pattern noise can be reduced when the image signals repeatedly read out from the solid-state image sensor are cumulatively added over the image display period and outputted in this way. It is characterized by:

Further, the present invention is characterized in that the dynamic range of the cumulatively added image signal is compressed and then output to the image monitor.

In other words, by repeatedly capturing an image of an object using a solid-state image sensor and reading out the image signal to the outside of the element at high speed, the exposure time of the object using the solid-state image sensor can be shortened. Prevent saturation in the element. The image signals obtained by this short-time exposure are repeated at high speed and cumulatively added to obtain an image signal with an expanded dynamic range. At the same time, fixed pattern noise can be reduced, and The present invention is characterized in that it obtains an image signal having a wide dynamic range without impairing the dynamic range of the subject and not causing vertical stripes on the display screen.

[Operations] According to the present invention, exposure of a subject by a solid-state image sensor and reading of the image signal to the outside of the element are repeated at high speed, and fixed pattern noise is reduced and cumulatively added to the resulting image signal. , it becomes possible to expand the dynamic range of cumulatively added image signals by the number of cumulative additions while keeping the dynamic range of each individual image signal itself narrow.

As a result, even for subjects with a wider dynamic range compared to solid-state image sensors, image signals with a wide dynamic range that are effectively imaged and input without being subject to the dynamic range limitations of the solid-state image sensors can be obtained. It becomes possible to ask for it.

On top of that, this wide dynamic range image signal is
The dynamic range is compressed in accordance with the dynamic range of the image monitor, and this is used for image display, so that image display can be performed without causing problems such as so-called self-discipline, blackout, or vertical stripes caused by fixed pattern noise. becomes possible.

[Embodiments] First, electronic cameras according to some basic examples of the present invention will be described with reference to FIGS. 1 to 12.

FIG. 1 is a schematic diagram of the main parts of an electronic camera according to a first basic example of the present invention, in which numeral 1 denotes a solid-state image sensor that electronically captures and inputs an image of a subject. This solid-state image sensor 1 basically includes a plurality of photoelectric conversion sections arranged in a matrix, and is driven by an image pickup control section (not shown) to send signal charges to each of the photoelectric conversion sections according to the amount of light of the subject. The exposure is done by the occurrence. By sequentially reading out pixel signals corresponding to the signal charges generated in each of these photoelectric conversion parts,
An image signal electronically captured by the solid-state image sensor 1 is obtained.

Here, however, as the solid-state image sensor 1, an AMI (amplification type MOS imager) is used, which has a function of being able to read out at high speed an image signal of a subject that has been imaged electronically. The imaging of the object by the solid-state imaging device 1 and the reading of the image signal thereof are repeated at high speed a plurality of times within the period of image display.

The image signal read out at high speed from the solid-state image sensor 1 in this way is amplified to a predetermined signal level via the preamplifier 2, and then subjected to clipping processing etc. via the signal processing circuit 3. In the A/D converter 4, for example, 1
It is sequentially digitally encoded into a 2-bit signal.

That is, in conventional boat fishing, the image input of the subject by the solid-state image sensor 1 corresponds to the image cycle on the image monitor.
For example, only the imaging and image signal reading are performed every 1 field period (17 msec), or here the solid-state image sensor 1 is driven at high speed, and the image signal is read out, for example, 10 times within the 1 field period (17 msec). Imaging and image signal reading are repeatedly performed at high speed.

In this way, the image signals repeatedly read out at high speed from the solid-state image sensor 1 are sequentially digitally encoded as described above.

The digitally encoded image signal is then sequentially stored in a frame memory 6 frame by frame via an adder 5. The adder 5 adds the image signal stored in the frame memory 6 and the image signal of the next frame that is sequentially and repeatedly read out from the solid-state image sensor 1, and rewrites the result in the frame memory 6. By doing so, the image signals that are repeatedly read out as described above are cumulatively added to the frame memory 6. By repeating such cumulative addition of image signals over the aforementioned image display period, for example over one field period, the dynamic range of the image signal can be expanded.

That is, in this electronic camera, the electronic image capture input of the subject by the solid-state image sensor 1 is repeatedly performed multiple times (for example, 10 times) within the image display period (one field), as described above. The subject exposure time is set to be sufficiently short. Therefore, the amount of signal charges generated in the solid-state image sensor 1 according to the amount of light from the object increases in proportion to the exposure time, and when the exposure time is set short as described above, the dynamic range of the object becomes larger than that of the solid-state image sensor. Even if the dynamic range is sufficiently wider than the dynamic range of the solid-state image sensor 1, it is possible to perform exposure without causing saturation of the solid-state image sensor 1. In other words, exposure is performed without saturating the high-brightness portions of the subject. However, since the exposure time is short, the amount of exposure for low-luminance portions of the subject is insufficient.

By repeatedly and cumulatively adding such image signals over a plurality of times within the image display period (one field), the level of the image signal is increased by the number of repetitions, and the dynamic range is substantially increased. Expansion is possible. For example, by repeatedly reading image signals n times within one field period of image display and performing cumulative addition, the dynamic range of the cumulatively added image signal can be compared with the dynamic range of each individual image signal. The image is enlarged by n times.

Incidentally, random noise components such as dark current mixed into the image signal when the solid-state image sensor 1 exposes the subject increases in a square root function with respect to the length of the exposure time. For example, if the exposure time is (1/n ), the noise level is reduced to (1/(M)) as illustrated in FIG.

Specifically, when the exposure time is set to (1/10), the noise level becomes (A/JT completed-), and a dynamic range improvement effect of 10 dB can be expected. Furthermore, since the noise component itself has randomness, the noise component increases by only J'T times by cumulative addition of the image signals described above. As a result, the dynamic range of the image signal cumulatively added n times is f
x, "i'i""', that is, it is improved by n times.

In this way, solid-state imaging is achieved by the above-described high-speed repeated readout of image signals from the solid-state imaging device 1, which electronically captures and inputs the image of the subject, and by the cumulative addition of the image signals using the frame memory 6 and the adder 5. The level of the image signal captured within the dynamic range of the element 1 is increased, and the dynamic range is expanded.

In other words, the dynamic range is expanded by n times by cumulatively adding the image signals input within the dynamic range of the solid-state image sensor 1 n times. This is possible by compressing the dynamic range through the short-time exposure mentioned above and expanding the dynamic range through cumulative addition through high-speed readout when photographing a subject with a wide dynamic range of brightness levels compared to the dynamic range of the solid-state image sensor 1. This means effective image input. As a result, e.g. 50 dB
A subject having a dynamic range of 80 dB is imaged and input by effectively using the solid-state image sensor 1 having a dynamic range of
dB.

Moreover, for imaging input of such a wide dynamic range object, it is only necessary to change the driving conditions of the solid-state image sensor 1 as described above and provide a means for accumulatively adding image signals that are repeatedly read out at high speed. In terms of hardware, it is very simple. Therefore, make it very simple,
It becomes possible to effectively expand the dynamic range of the image signal to be imaged and input.

By the way, as mentioned above, the video processor 7 inputs the image signal whose dynamic range has been expanded by cumulative addition.
separates and extracts the luminance signal component Y from the image signal, and also decomposes the image signal into three primary color components of R, G, and B, respectively. The dynamic range control circuit 8 inputting this luminance signal component Y controls the dynamic range of the image signal (80 dB in the above example) on the image monitor (
For example, a compression coefficient for compression conversion to a dynamic range (for example, 40 dB) possessed by a TV receiver) is determined.

That is, when displaying an image signal with a wide dynamic range as described above using an image monitor with a dynamic range of less than 40 dBL, if the offset is set to [0], the image signal will be reduced by 40 dBL or less as shown in A in Figure 3. It cannot be displayed, and the high-brightness portion becomes blown out due to saturation.

On the other hand, if the offset is set to a negative value and the high-brightness portion is adjusted to the maximum display level of the image monitor, it becomes impossible to display an image at a high range of 40 dBL of the image signal, as shown in FIG. 3B. In this case, the low luminance level portion of the image signal is in a state of crushed black.

Therefore, the dynamic range control circuit 8 compresses the dynamic range of the image signal as shown in C in FIG. The compression coefficient is determined according to the dynamic range of the image signal.

However, in this basic example, the dynamic range of the image signal (separated three primary color components R, G, B) can be easily compressed.
In order to do this effectively, the image signal separated into the three primary color components R, G, and B and the luminance signal component Y are each subjected to logarithmic transformation, and then the dynamic range controlled signal is subjected to antilogarithmic transformation. and is configured to restore the original signal form.

Regarding the method of controlling this dynamic range,
For example, the method proposed in Japanese Patent Application Laid-Open No. 63-232591 can be used as is. Therefore, here:
This dynamic range control will be briefly explained.

That is, the dynamic range control circuit 8 logarithmically transforms the luminance signal component Y into log Y via the log amplifier 8a, and inputs the luminance signal component Y, for example, as shown in FIG.
The uneven illuminance component is removed via the dimensional filter 8b. This two-dimensional filter 8b, for example, as shown in FIG.
By correcting the luminance signal obtained through F, clipping it, and subtracting it from the luminance signal component,
This is to remove the component of uneven illuminance.

Then, the signal from which the uneven illuminance component has been removed is passed through the two-dimensional filter 8b to a dynamic range gain controller (DGC) configured as shown in FIG.
) 8c, and an output 10gαYβ (where α is a gain adjustment coefficient and β is a dynamic range adjustment coefficient) is obtained by matching the dynamic range of the luminance signal component Y to the dynamic range of the image monitor. In Figure 5, a D/A converter is installed at the input and an A/D converter is installed at the output.
/A, A/D converter may be omitted.

Incidentally, the same output can be obtained even if the circuit portion shown surrounded by the dotted line in FIG. 5 is simplified and configured as shown in FIG. 6, for example. The ratio log (αYβ/Y) between such log αYβ and the luminance signal component log Y is obtained by the subtracter 8e, and this is used as a compression coefficient indicating the degree of dynamic range compression to be performed.

That is, this DGC 8c performs feedback control so that the standard deviation value of one screen (or one part thereof) of the luminance signal is constant. In other words, by feeding back the amplified error value between the average value of the luminance signal and its gain reference voltage to the adder, the gain is adjusted so that the average value of the luminance signal is equal to the gain reference voltage. configured to adjust the range. Therefore, the output lo of DGC8c whose dynamic range is made constant in this way
The ratio between gαYβ and the luminance signal component log Y is 10g (
If αYβ/Y) is calculated, the value becomes a compression coefficient that indicates how much the input image signal should be compressed to match its dynamic range to the dynamic range of the image monitor, as described above. become.

The three primary color components R, G, and B of the image signal converted and outputted by the video processor 7 are each converted to a log amplifier 9.
logR via r, 9g, 9b.

It is logarithmically transformed as log G and log B. Then their respective signals log R, log G, l
og B is provided to adders 1.1r, l1g, and llb via delay circuits 10r, 10g, and 10b, respectively, and the dynamic range compression coefficient log (αYβ/Y) determined by the DGC 8c is added thereto.

As a result, each output of the adders 11r, l1g, and ll'b is log R+log (αYβ/Y)−1ogR(α
Yβ/Y) ogG+log(αYβ/Y) −logG(αYβ/Y) log B+log(αYβ/Y)−IogB(α
Yβ/Y), which is converted into reverse log amplifiers 12r and 12g.

12b, and the D/A converters 13r, 1
The signals output through 3g and 13b are (αYβ/Y) R9 (αYβ/Y) G (αYβ/Y)
B is a signal component whose dynamic range has been compressed.

Thus, by adjusting the dynamic range of the output image signal using such a dynamic range adjustment means, it is possible to effectively match the required wide dynamic range image signal to the dynamic range of the image monitor as described above. It becomes possible. As a result, as shown in FIG. 3, it becomes possible to effectively display the image signal without causing blown-out highlights or blown-out shadows.

In addition, in the embodiment described above, the dynamic range control circuit 8
In the above, unevenness in illumination was removed using two-dimensional filtering processing, and then a luminance signal with dynamic range adjustment was obtained using DGB8c, from which a compression coefficient for dynamic range control was obtained. It is possible to suppress the dynamic range of the luminance signal to a certain range simply by removing the uneven illuminance caused by the above.

Therefore, in such a case, for example, the output of the clipping circuit for adjusting the luminance signal level in the two-dimensional filter 8b shown in FIG.
Instead of l1g and llb, a subtracter 3 as shown in FIG.
1r, 31g.

31b, respectively. In other words, if the unevenness of illumination of the subject is removed, the dynamic range of the image signal can be suppressed to 40 to 50 dB, so even if the compression coefficient obtained from the two-dimensional filter 8b is used as is, the dynamic range will be effective to some extent. The range can be adjusted. In addition, in this case, the above-mentioned DGC 80 and the like are not required, so that the configuration can be simplified.

By the way, the above-mentioned basic example is an example of a video movie camera that continuously captures and inputs an image of a subject at each image display cycle and displays the image, but recently, still images are captured and input using a solid-state image sensor 1. Various electronic still cameras have also been developed. For example, while monitoring images that are continuously captured and input using a solid-state image sensor built into an endoscope, it is possible to capture the resulting screen as a still image and store it in mass storage such as a digital VTR or digital video file. It will be done.

When the present invention is applied to such a system, it may be configured as shown in the second basic example shown in FIG. 8, for example.

In FIG. 8, the same parts as those in FIG. 1 described above are designated by the same reference numerals.

That is, in this example, the processor 7 also obtains the color difference signals (RY) and (B-Y) at the same time. Then, the color difference signals (RY) (B-Y) and the luminance signal Y are respectively taken into the memories 21a, 21b, and 21C over one field according to the release timing. And these memories 21a, 21
b.

21c, each signal component is sequentially read out via the parallel/serial converter 22, and the data compressor 23,
A compression process and a recording modulation process are performed through the recording modulation unit 24 in series, and the data is recorded in the mass storage unit 25.

The image signal read out and reproduced from the mass storage section 25 is decoded and reproduced from the demodulation section 26 via the pig decoding section 27, and the above-mentioned color difference signal (R-Y), (B -Y) and a luminance signal Y. Then, the matrix circuit 29 calculates the three primary color components R, G, and B from the color difference signals (RY), (B-Y) and the luminance signal Y, respectively.

In addition to configuring the image signal recording and reproducing section as described above, the dynamic range adjustment section similar to that described above processes the image signal required from the solid-state image sensor 1, or the mass storage section Switch circuit 3 for switching whether to process the image signal read out from 25
It is sufficient to incorporate 0.

According to such a second basic example, the wide dynamic range image signal input as described above is converted into its luminance signal Y and color difference signal (R-Y).

(B-Y) can be accumulated and stored in the mass storage unit 25 with good quality. Then, it becomes possible to effectively adjust the dynamic range of the recorded and saved wide dynamic range image signal and provide it for image display on an image monitor.

Further, on each side mentioned above, it is necessary to quantize the noise component of the image signal in order to ensure the dynamic range of the image signal, and the number of quantization bits is about 12 bits. However, the 12-bit high-speed A/D converters currently in use, especially those capable of operating at television signal rates, are very expensive. Moreover, as the resolution of images progresses and the frequency band becomes wider, 12
There is a possibility that an A/D converter with a quantization precision of about a bit will not be able to handle the above-mentioned image processing.

On the other hand, conventional digital image processing is often performed in units of 8 bits, and therefore it is practically preferable in the present invention to perform image signal processing in 8 bits.

Therefore, as in the third basic example shown in FIG. 9, the image signal repeatedly read out at high speed from the solid-state image sensor 1 is input to the video processor 7, and at this point, the luminance signal Y and the three primary color components R, G, B and logarithmically transform each of these signals. After each of the logarithmically transformed signal components is digitally encoded into 8 bits, each of these signal components is written into the frame memories 6a, 6b, 6g, and 6r and cumulatively added. This cumulative addition expands the dynamic range of each signal component. Since each signal component whose dynamic range has been expanded becomes, for example, a 12-bit digital signal,
These bit shifters 14a, 14b, 14g, 14r
The data is compressed to 8 bits via the 8 bits, and then subjected to the dynamic range compression process described above.

By configuring the signal processing circuit in this way, it becomes possible to handle 8-bit signals in most of the signal systems, making it possible to construct an electronic camera using existing general-purpose image processing circuit components (semiconductor ICs, etc.) as they are. It becomes possible. Also A
Although the noise component is greatly amplified by the log amplifier installed before the /D converter, the S/N improvement effect of the cumulative addition works as described above, and the cumulatively added image signal is converted into bits. Since the signal is shifted, it is possible to suppress the noise component to a very small level. As a result, S/H
This makes it possible to display a good image.

Next, another example for reducing the number of quantization bits of the A/D converter will be shown below. In this example, for example, prior value predictive coding is employed, and desired digital conversion is performed on the image signal using an A/D converter with a smaller number of quantization bits than the required number of quantization bits.

FIG. 10 is a schematic diagram of a digital conversion circuit that employs predictive coding to realize 12-bit digital coding using an 8-bit A/D converter. This digital conversion circuit uses an analog subtracter 40 to find the difference between the current input analog signal and the signal generated by the decoding circuit 32 one data before, and uses this as a predicted signal. This predicted signal is subjected to non-linear conversion in a non-linear conversion section 33, and its output is converted into a high-speed digital signal using an 8-bit high-speed operation type A/D converter 34. The nonlinear conversion unit 33 includes
Since the predicted signal obtained by the analog subtracter 31 concentrates around zero, a nonlinear conversion characteristic as shown in FIG. 11, for example, is set so as to finely sample the predicted signal values around zero.

By performing such nonlinear conversion, the analog input signal (predicted signal) is converted into an 8-bit digital signal in which the quantization accuracy is fine near zero, and the quantization accuracy becomes coarse as the value increases.

The decoding circuit 32 linearly converts the 8-bit digital signal using a linear converter 35 having input/output characteristics opposite to those of the non-linear converter 33. This linear conversion equalizes the quantization precision of the 8-bit digital signal. The digital adder 36 inputting this linearly converted digitized prediction signal receives the value outputted one data before (previous value).
are sequentially added to and output as 12-bit data.

The previous value is obtained by delaying the output of the digital adder 36 by one data interval via the delay circuit 37, and multiplying the output value by a predetermined coefficient γ in the coefficient multiplier 38. Note that the coefficient γ is often given as a value close to [1]. Further, when [γ-1] is used, the coefficient multiplier 38 is unnecessary.

By feeding back the previous value (output value of one data previous) output from this coefficient unit 38 to the digital adder 36, a predicted value at the current data timing is added to the output value of one data previous. is output. At the same time, the previous value (output value one data previous) output from the coefficient unit 38 is D
The signal is converted into an analog signal via the /A converter 39 and provided to the analog subtracter 40.

In this way, in the digital conversion circuit shown in FIG. 10, the difference between the result of analog conversion of the digital conversion result before data sampling and the analog input signal at the current data timing is obtained as a predicted signal, and this predicted signal is converted into a nonlinear After the conversion, an 8-bit A/D converter 24 is used to perform high-speed digital conversion. By adding this digitally converted predicted value to the digitally converted result of the previous data sample, the desired number of quantization bits (1
It is designed to obtain a 2-bit) digital signal.

As a result, a general-purpose 8-bit A that can operate at high speed and is inexpensive.
By effectively using the /D converter 24, a digital conversion result with a quantization bit number can be obtained, and image processing in each of the embodiments described above can be easily realized.

In addition, here, even if predictive coding is used as previous value prediction within a frame, interframe predictive coding, three-dimensional predictive coding, or even various planar predictive coding other than previous value prediction is adopted as appropriate, As described above, by using the inexpensive high-speed 8-bit A/D converter 24, it is possible to realize digital conversion processing with a desired number of quantization bits, such as obtaining a 12-bit digital signal, for example.

Note that the application of the present invention is not limited to the basic examples described above. For example, on each side mentioned above, the dynamic range is adjusted by converting the image signal into the three primary color components R, G, and B, but the dynamic range is adjusted by separating it into three complementary color components (yellow, magenta, cyan). You can also do it like this.

Further, the dynamic range may be adjusted for the color difference signals (R-Y) and (B-Y).

However, in this case, the color difference signal (RY) is used.

Since (B-Y) has positive and negative polarities, a log amplifier for logarithmically converting this must have characteristics as shown in FIG. 12, for example. Other solid-state image sensor 1
The number of times the image signal is repeatedly read out, the means for adjusting the dynamic range, etc. can be modified in various ways depending on the specifications.In short, the present invention can be modified and applied in various ways without departing from the gist of the invention.

By the way, when cumulative addition is used to obtain an image signal with a wide dynamic range as in each of the basic examples above, the random noise component in the image signal is relatively reduced compared to the signal component. The fixed pattern noise (hereinafter referred to as FPN) component generated in the signal processing circuit including the fixed pattern noise and clamp circuit is multiplied by the addition number and becomes as large as the signal component, reducing the random noise component. As a result, the FPN component becomes even more noticeable in the image signal.

Next, the present invention for reducing this FPN component will be explained in detail.

FIG. 13 shows a first embodiment of the present invention in which means for reducing the FPN component is added to the first basic example.

In the following embodiments, the parts configured in the same manner as in each of the basic examples described above will be designated by the same reference numerals and the explanation thereof will be omitted.
FPN components including background noise components generated in the amplifier 2 and the signal processing circuit 3 are recorded in the FPN data ROM 42 as A/D converted data.

It is assumed that this FPN data ROM 42 can store data of the same number of pixels as the frame memory 6. When capturing an image, after the input signal is A/D converted, the FP of the corresponding pixel is read from the FPN data ROM 42.
The FPN component is subtracted by reading the N data and sequentially subtracting it from the input signal in the subtracter 41, and then cumulatively adding the signal from the adder 5 onwards.

By subtracting the FPN component from the input signal in this way, only the random noise component remains in the input signal.
Since the random noise component is also reduced by the subsequent cumulative addition, a high-quality image signal with an improved S/N ratio and a wide dynamic range can be obtained. Thereby, problems such as vertical stripes caused by the FPN component on the display screen can be eliminated.

In the first embodiment, the data of the FPN component including the background noise component is stored in the ROM 42 in advance, but it is possible to store such FPN data at any time. Example 2 is the first example.
Shown in Figure 4.

In this embodiment, switches 45 and 46 for switching between the photographing mode and the FPN storage mode and an FPN shutter 47 are provided. In the shooting mode, a switch 45 provided after the frame memory 6 is connected to the video processor 7, and a switch 46 provided on one side of the subtracter 41 is turned on.
In this state, the FPN shutter 47 provided at the front stage of the image sensor 1 is opened. And when shooting, use the FP
The previously stored FPN data is read from the N frame memory 43 and subtracted from the input signal by the subtracter 41, and the signal with the FPN component subtracted is cumulatively added by the adder 5 and the frame memory 6.

On the other hand, in the FPN storage mode, the switch 45 is 1/n
connected to the multiplier 44 side, the switch 46 is 0FFSFP
The N shutter 47 is in a closed state.

It is assumed that the switch 45, the switch 46, and the FPN shutter 47 are interlocked. When storing FPN data, by setting the FPN storage mode, external light is blocked, so noise generated in the image sensor 1, amplifier 2, and signal processing circuit 3
The A/D converter 4 converts the FPN component into a digital signal, which is cumulatively added by the adder 5 and frame memory 6, reducing random noise and emphasizing the FPN component. be done. The random noise component is cumulatively added the number of times n that the random noise component is sufficiently smaller than the FPN component, and the signal is made to the same level as the signal before the cumulative addition by the 1/n multiplier 44 and then stored in the FPN frame memory 43.

According to this embodiment, even if the FPN component changes due to a change in temperature, the FPN component is re-stored and then imaged, so an image signal with a wide dynamic range and good S/H can be obtained. This embodiment is particularly effective since the FPN component of an image sensor such as a CCD changes by as much as 6 dB at 10°C. It is also effective when the FPN component changes due to changes over time in the characteristics of elements in the signal path from the image sensor 1 to the A/D converter 4.

If the above cumulative addition number n is set to an integer power of 2, the above 1
/n multiplier 44 can be replaced with a bit shifter, so
The configuration becomes easier. In addition, this embodiment
The N component is multiplied by 1/n and recorded, and the FPN component is subtracted before being cumulatively added during imaging, so the FPN
It is possible to reduce the capacity of the frame memory for use, and it is also possible to change the set value of the addition number n when storing the FPN and when capturing an image.

Note that when this embodiment is applied to a video camera, the FP
The N shutter 47 can be substituted by closing the aperture, and when applied to a still video camera, it can also be used as a normal shutter, so there is no need to provide it specially. In either application, if the FPN storage mode is automatically set when the power is turned on, and the imaging mode is set after storage, the camera can be operated in exactly the same way as a conventional video camera or still video camera. In addition, under conditions where the environment such as temperature changes rapidly, a video camera needs to re-store new FPN data every few frames or a still video camera every time the shutter is released. By doing so, the FPN component can be reliably reduced, so that an image signal with a good S/N ratio can always be obtained.

The embodiments shown in FIGS. 13 and 14 are for the case where the FPN components are added additively, but when the FPN components are added multiplicatively, the subtracter 41 in FIGS. 13 and 14 is used. A similar effect can be obtained by replacing it with a divider.

FIG. 15 shows a modification for noise reduction.

FIG. 15 shows a part of FIG. 1, showing the image sensor 1.
, the amplifier 2, the signal processing circuit 3.7, and the A/D converter 4. A cooling device 48 is added, which is thermally coupled to each of the amplifier 2, the signal processing circuit 3.7, and the A/D converter 4, and is capable of cooling each of them. This cooling device 48
As a result, the dark current in the image sensor 1 is significantly reduced,
Furthermore, thermal noise generated in each of the amplifier 2, signal processing circuit 3, and A/D converter 4 is also reduced. therefore,
Since the fixed pattern noise component due to dark current unevenness of the image sensor 1 is also reduced, both the FPN component and the random noise component are effectively reduced.

This makes it possible to reduce the number of times of addition in the cumulative addition circuit at the subsequent stage, making it possible to increase the speed and be advantageous in terms of time. Furthermore, when this modification is combined with the first embodiment shown in FIG. 14, the FPN component is significantly reduced, so the number of bits per pixel of the FPN frame memory 43 shown in FIG. 14 is reduced. There is an advantage that it can be done.

[Effects of the Invention] As explained above, according to the present invention, it is possible to very simply and effectively increase the dynamic range of an image signal inputted by a solid-state image sensor, and also to reduce fixed pattern noise. This makes it possible to effectively image and input a subject with a wide dynamic range. In addition, it is possible to obtain the effect of improving the S/H, which provides great practical effects.

[Brief explanation of drawings]

The figures show an electronic camera according to the present invention. Fig. 1 is a schematic configuration diagram showing a first basic example, Fig. 2 is a diagram for explaining the S/N improvement effect by cumulative addition of image signals, Figure 3 is a diagram showing the relationship between the dynamic range of the image signal and the dynamic range of the image monitor, and Figure 4 is a diagram showing the relationship between the dynamic range of the image signal and the dynamic range of the image monitor.
Figure 5 showing an example of the configuration of the two-dimensional filter used in the figure.
The figure is a diagram showing an example of the configuration of the dynamic range/gain controller used in Figure 1, and Figure 6 is a diagram showing the main part configuration of a modified example of the dynamic range/gain controller.
FIG. 7 is a diagram for explaining a modification example for dynamic range adjustment. FIG. 8 is a schematic diagram showing a second basic example of the present invention, FIG. 9 is a schematic diagram showing a third basic example of the invention, FIG. 10 is a diagram showing an example of the configuration of a digital conversion circuit, and FIG. Figure 11 is the 10th
FIG. 12 is a diagram showing the characteristics of nonlinear conversion in the digital conversion circuit shown in the figure, and FIG. 12 is a diagram showing the characteristics of a log amplifier used for logarithmically converting a color difference signal having positive and negative polarities. FIG. 13 is a schematic configuration diagram showing the first embodiment of the present invention, and FIG. 14 is a schematic configuration diagram showing the second embodiment of the invention.
FIG. 5 is an explanatory diagram of the configuration of main parts showing a modification for noise reduction. DESCRIPTION OF SYMBOLS 1... Solid-state image sensor (AMI), 2... Preamplifier, 3... Signal processing circuit, 4... A/D exchanger, 5
... Adder, 6... Frame memory, 7... Video processor, 8... Dynamic range control circuit,
9b, 9g. 9 r --- Log amplifier, llb, l1g, 11 r
--Adder, 12b, 12g, 12r--Reverse log amplifier, 13b, 13g, 13r-D/A converter, 41-
...Subtractor, 42...FPN data ROM, 43...
・Frame memory for FPN, 44...1/n multiplier, 4
5゜46... Switch, 47... FPN shutter.

Claims (3)

[Claims] (1) In an electronic camera that uses a solid-state image sensor to electronically image a subject and display the image on an image monitor, there is a solid-state image sensor that can read out the input image signal outside the element, and a subject that is captured by this image sensor. means for repeatedly performing imaging and reading out the image signal multiple times; means for reducing a fixed pattern noise signal from this image signal; and sequential cumulative addition of the image signal from which the fixed pattern noise signal has been reduced. An electronic camera characterized by comprising means for: (2) The fixed pattern noise reduction means includes a memory for storing a fixed pattern noise signal, and an arithmetic means for subtracting or dividing the fixed pattern noise signal from the image signal read from the solid-state image sensor. The electronic camera according to claim 1, characterized in that: (3) The electronic camera according to claim 1, wherein the fixed pattern noise reduction means includes a cooling means for cooling at least the solid-state image sensor.