JP5861924B2 - Imaging device - Google Patents

Description

  The present invention relates to image processing in an imaging apparatus suitable for a so-called digital camera or the like that captures a still image or the like using an imaging element, and more particularly to an imaging apparatus having an image composition function using a plurality of captured images. .

In recent years, photographing using a synthetic image generation method called “comparative bright composition” or the like is being used as a photographing technique peculiar to a digital camera.
This is to take continuous images in time series under the same exposure conditions and the same composition, and compare the output values of the corresponding pixels of the same coordinates for the reference image (composition result image) and the comparison image, This is a technique in which if the output value of the comparative image is higher, it is replaced with that and image synthesis is performed. With this method, for example, in night scene photography, it is possible to generate a characteristic photographed image that leaves a light trail of stars while keeping the background at an appropriate exposure.
Conversely, continuous images are taken in time series under the same exposure conditions and in the same composition, and the output values of corresponding pixels at the same coordinates are compared for the reference image (composition result image) and the comparison image. However, if the output value of the comparative image is lower, image synthesis called “comparative dark synthesis” or the like, which replaces the output value and performs image synthesis, is also being used.
Some digital cameras already have built-in image synthesis functions such as so-called comparative bright synthesis and comparative dark synthesis.

Here, the main merits by executing the above-described image synthesis functions such as so-called comparative bright synthesis and comparative dark synthesis in the digital camera will be described.
First, in the past, when performing image synthesis such as so-called comparative bright synthesis and comparative dark synthesis, a continuous image as a material is taken and comparatively bright using image processing software on a PC (personal computer) or the like. A synthesis process was performed. In such a case, it was not possible to obtain a combined image on the spot at the time of shooting, but by installing this function in the camera, it becomes possible to immediately check the combined image at the time of shooting. .
Conventional processing using a PC or the like is mainly performed using image data such as a so-called JPEG image (generally referred to as a JPG image because jpg is generally used as an extension). In other words, in conventional processing using a PC or the like, the image data that is the material is image data that has undergone basic image processing such as noise reduction and is compressed as necessary. In some cases, it was not enough. On the other hand, if processing is performed in the camera, it is possible to perform synthesis processing with a RAW image that has not been subjected to image processing. There is an advantage that the image quality can be left as it is, and the effect of improving the image quality, which is difficult to obtain by the normal image processing, is obtained by averaging the random noise by the synthesis processing.

A technique relating to the composition processing of a plurality of images is disclosed in, for example, Patent Document 1 (Japanese Patent Laid-Open No. 2011-199786).
Patent Document 1 discloses an image synthesis technique based on electronic camera shake correction. When performing image synthesis, in order to reduce noise as much as possible while suppressing the occurrence of blurring and double images, multiple images are used. It is shown that the synthesis ratio is determined using the noise estimation value and the correlation value.
The technique disclosed in Patent Document 1 determines a combination ratio using a threshold value based on a difference value between a plurality of images and a noise estimation value, and is not related to comparatively bright combination or comparative dark combination. In addition, in shooting with high sensitivity such as high ISO sensitivity, noise characteristics such as SN ratio (signal-to-noise ratio) deteriorate, but noise and subject state change in such cases No consideration is given to image synthesis with a RAW image that is not subjected to basic image processing such as noise reduction or noise reduction.

As described above, it is considered desirable to incorporate an image synthesis function such as so-called comparative bright synthesis and comparative dark synthesis in an imaging apparatus such as a so-called digital camera. However, in such a case, noise characteristics deteriorate. It is desirable to take appropriate measures for separation of noise and subject state change in sensitivity shooting, or for image synthesis with a RAW image.
The present invention has been made in view of the above-described circumstances, and incorporates an image synthesis function such as comparative bright synthesis and comparative dark synthesis, and effectively suppresses deterioration in the quality of the synthesized image due to noise distribution. An object of the present invention is to provide an imaging device that can be used.

In order to achieve the above-described object, an imaging apparatus according to the present invention provides
An imaging means for photoelectrically converting a subject optical image by an imaging device to obtain a captured image composed of electronic image data;
Noise estimation means for estimating a noise component included in the captured image obtained by the imaging means;
Image processing means for obtaining a signed differential value between a plurality of captured images obtained under the same conditions and the same composition by the imaging means;
Combining ratio determining means for determining a different combining ratio between positive and negative based on a noise component by the noise estimating means for the signed difference value obtained by the image processing means;
And image synthesizing means for synthesizing the plurality of photographed images by weighted averaging based on the synthesis ratio obtained by the synthesis ratio determining means.

  According to the present invention, there is provided an imaging apparatus that incorporates an image composition function such as comparative light composition and comparative dark composition and that can effectively suppress deterioration in image quality of a composite image due to noise distribution. Can do.

That is, according to the imaging device according to the invention of claim 1,
An imaging means for photoelectrically converting a subject optical image by an imaging device to obtain a captured image composed of electronic image data;
Noise estimation means for estimating a noise component included in the captured image obtained by the imaging means;
Image processing means for obtaining a signed differential value between a plurality of captured images obtained under the same conditions and the same composition by the imaging means;
Combining ratio determining means for determining a different combining ratio between positive and negative based on a noise component by the noise estimating means for the signed difference value obtained by the image processing means;
An image compositing function such as comparative bright composition and comparative dark composition by comprising image composition means for combining the plurality of photographed images by weighted averaging based on the composition ratio obtained by the composition ratio determining means. It is possible to effectively suppress deterioration in the image quality of the composite image due to the built-in noise distribution.

It is a block diagram which shows typically the structure of the principal part of the digital camera which concerns on one embodiment of the imaging device by this invention. 3 is a flowchart illustrating an example of an image composition process using a weighted average of a reference (composite) image and a comparison image in the digital camera of FIG. 1. FIG. 2 is a characteristic diagram illustrating a relationship between output values for each sensitivity and dispersion of output values in the digital camera of FIG. FIG. 2 is a characteristic diagram schematically illustrating an example of a noise estimation distribution characteristic and a composition ratio characteristic with respect to a signed difference value between a reference (composite) image and a comparison image in the digital camera of FIG. 1. FIG. 3 is a characteristic diagram illustrating an example of a relationship between output values before and after noise reduction processing and dispersion of output values in the digital camera of FIG. 1. It is a characteristic view for demonstrating an example of the noise distribution characteristic with respect to the signed difference value of the reference | standard (synthesis | combination) image and a comparison image in the conventional digital camera, and the problem by the simple comparative bright combination process in that case. FIG. 7 is a characteristic diagram illustrating an example of the relationship between the number of combined images by the simple comparative light combining of FIG. 6 and the variance of output values of processing results. It is a characteristic view showing an example of setting of a composition ratio with respect to a signed difference value between a reference (composition) image and a comparative image for explaining the effect of the present invention. FIG. 9 is a characteristic diagram illustrating an example of the relationship between the number of combined images by the combining processing of FIG. 8 and the dispersion of output values of processing results. FIG. 10 is a characteristic diagram illustrating an example of setting a composition ratio for a signed difference value between a reference (composite) image and a comparative image for comparatively bright composition. FIG. 10 is a characteristic diagram illustrating an example of setting a synthesis ratio for a signed difference value between a reference (composite) image and a comparative image for comparative dark synthesis.

Hereinafter, based on an embodiment of the present invention, an imaging device according to the present invention will be described in detail with reference to the drawings.
Before describing an imaging apparatus according to an embodiment of the present invention, the principle of image processing used in the embodiment of the present invention will be described.
First, comparative bright combination and comparative dark combination in the image combining technique will be described.
The comparatively bright combination is, for example, a pixel having the same coordinates or a plurality of pixels in the same region in a plurality of digital images, which are typically photographed continuously in time series under the same exposure condition and in the same composition. The images are combined by a method in which the outputs are compared with each other and replaced with one having a high output value (luminance). By doing so, it is possible to generate an image that leaves a light trail of a star that was not easy to shoot with a conventional film camera using a so-called silver salt film. Characteristic shooting is possible. Also, comparative dark synthesis is, for example, image synthesis based on a method that compares the outputs of pixels with the same coordinates or multiple pixels in the same region in a plurality of digital images and replaces them with one having a low output value (luminance). It is.

Conventionally, the comparative light composition and the comparative dark composition as described above are generally performed with basic image processing and compressed JPEG images by using PC software. Products with a function that performs consistent processing within the company have begun to be announced.
As described above, the advantage of having a digital camera with built-in image composition functions such as comparative light composition and comparative dark composition to enable image composition such as comparative light composition and comparative dark composition within the digital camera is There is a merit in terms of operability in that it is possible to immediately see the image after image synthesis on the spot. In addition to this, a RAW image output from the image sensor (RAW data image: raw data, that is, an image of raw data output from the image sensor and not subjected to processing such as noise reduction or data compression). Since it is possible to perform composition processing using the image as it is, a merit in terms of image quality that dark stars, light-colored stars, etc. can be properly left is also important.
On the other hand, there is also a technical problem in performing image composition such as comparatively bright composition and comparative dark composition using a RAW image. Unlike a JPEG image, a RAW image undergoes only the minimum necessary image processing, and often includes a noise component. Therefore, for example, when image synthesis processing is performed by comparatively bright synthesis, in which output values are simply compared and replaced as in the past, the output level of the image rises overall due to the influence of noise components. If the output is offset or the output is offset, the color tone may become magenta.

Therefore, estimation data related to the noise component distribution is stored in advance, the difference value between the reference image and the comparison image is calculated, and the comparatively bright synthesis and the comparative dark synthesis are performed using the threshold curve calculated from the noise distribution estimation data. If it is determined whether or not this is possible, the above-described problems can be suppressed. In this case, if it is determined that there are too many noise components and appropriate combination determination cannot be performed, it is possible to switch to combination determination including information on peripheral pixels without determining only the corresponding pixel.
In short, synthesis processing such as comparatively bright synthesis and comparative dark synthesis of RAW images requires synthesis determination in consideration of noise components, unlike the case of JPEG images, and is more preferable by adding such processing. A composite image can be obtained.
A digital camera according to an embodiment of the imaging apparatus of the present invention based on the above-described principle will be described in detail below with reference to the drawings.
FIG. 1 shows the configuration of the main part of a digital camera according to an embodiment of the imaging apparatus of the present invention. FIG. 1 is a block diagram showing a system configuration mainly related to electronic control of a digital camera as an imaging apparatus.

A digital camera shown in FIG. 1 includes an optical system 1, an analog front end unit 2, a signal processing unit 3, an SDRAM (Synchronous Dynamic Random Access Memory) 4, a liquid crystal display unit (LCD) 5, a memory card 6, a ROM (Read Only). A memory) 7, an operation unit 8, and a driver 9.
The optical system 1 includes a lens group 11, a diaphragm 12, a shutter 13, and an image sensor 14. The analog front end unit 2 includes a timing generator 21, a CDS circuit 22, an AGC circuit 23, and an A / D conversion circuit 24. The signal processing unit 3 includes a control unit (CPU) 31, an image sensor interface (image sensor I / F) 32, a memory controller 33, a YUV conversion unit 34, a resize processing unit 35, a display output control unit 36, and a data compression unit. 37 and a media interface (media I / F) 38.
The optical system 1 includes a lens group 11, a diaphragm 12, a shutter 13, and an image sensor 14. An ND (intermediate density) filter for dimming or the like may be incorporated in the aperture 12, shutter 13, and image sensor 14 portions. The image sensor 14 may be a CCD (Charge Coupled Device) image sensor or a CMOS (Complementary Metal Oxide Semiconductor) image sensor. However, when the image sensor 14 is a CMOS image sensor, the latter stage is used. In many cases, the analog front end unit 2 is substantially built in the CMOS image sensor, and is unnecessary as an external circuit.

The analog front end unit 2 includes a timing generator 21, a CDS (correlated double sampling) circuit 22, an AGC (automatic gain control) circuit 23, and an A / D (analog-digital) conversion circuit 24. The CDS circuit 22 removes noise from the output signal of an image sensor such as a CCD image sensor by correlated double sampling, the AGC circuit 23 amplifies the signal as necessary, and the A / D conversion circuit 24 literally performs analog imaging. The signal is converted into digital image data, and the timing generator 21 generates a drive timing signal for each part in the image sensor 14 and the analog front end unit 2.
The signal processing unit 3 includes a CPU 31, an image sensor I / F 32, a memory controller 33, a YUV conversion unit 34, a resize processing unit 35, a display output control unit 36, a data compression unit 37, a media I / F 38, and the like. The CPU 31 is connected to the image sensor I / F 32, the memory controller 33, the YUV conversion unit 34, the resize processing unit 35, the display output control unit 36, the data compression unit 37, the media I / F 38, and the like, and exchanges information with each other. And control them. The memory controller 33 is connected to the SDRAM 4, the display output control unit 36 is connected to the LCD 5, and the media I / F 38 is connected to the memory card 6. The CPU 31 is connected to a ROM 7 that stores an operation program and the like, and is connected to an operation unit 8 that includes a push button switch and the like for inputting operation information. Further, the CPU 31 controls the driver 9 for driving the lens group 11, the diaphragm 12, the shutter 13 and the like of the optical system 1.

Next, an operation that characterizes the present invention in the digital camera according to one embodiment of the present invention configured as described above will be specifically described. Embodiment described here is not limited to it, It can implement with various corrections or deformation | transformation in the range which can be easily conceived by those skilled in the art.
The flowchart shown in FIG. 2 shows the flow of image composition processing such as comparative light composition or comparative dark composition in the digital camera described above. The flowchart in FIG. 2 shows an example of the image composition process, and this image composition process is executed by the CPU 31 of the signal processing unit 3 in FIG.
The feature of the image composition processing according to the present invention is that, for example, when performing image composition such as comparatively bright composition and comparative dark composition using a plurality of RAW images continuously taken in time series, the output values of corresponding pixels are compared. In consideration of the distribution of the noise component, the replacement determination is performed after performing the noise processing as necessary. In this embodiment, specifically, image synthesis is performed using a weighted average weighted with a synthesis ratio according to the noise distribution.

FIG. 2 shows the flow of image composition processing in this embodiment, and will be described mainly using comparatively bright composition as an example.
In this case, as a premise, images continuously shot have the same exposure settings such as aperture, exposure time, and ISO sensitivity, that is, the same exposure settings, and the digital camera is fixed to a tripod or the like with the same composition. It is assumed that a plurality of images are taken in series. However, it does not matter whether there is a moving subject in the composition of the captured image.
First, when processing is started, a first RAW image as a reference image is acquired (step S101).
Next, noise distribution data in the digital camera corresponding to the exposure setting stored in advance in the ROM 7 or the like is read, and a synthesis ratio curve for determining a synthesis ratio is generated (first half of step S102).
Here, the composition ratio curve in the first half of step S102 will be described in detail. The amount of noise included in the image data is determined by various factors, and the most influential factor is the sensitivity in the imaging system, such as ISO sensitivity. For example, FIG. 3 shows an example of the relationship between the luminance output value and the output dispersion in the range of ISO 100 to ISO 3200 in the digital camera.

As shown in FIG. 3, at the same ISO sensitivity, there is a correlation between the luminance output value and the output variance, and at the same luminance output value, there is a correlation between the ISO sensitivity and the output variance. Such data is stored in advance in, for example, the ROM 7 in the digital camera, and is calculated as noise distribution estimation data according to the ISO sensitivity setting value at the time of shooting.
Next, a synthesis ratio curve for determining a synthesis ratio is generated from the read noise distribution data (second half of step S102). For example, the variance for ISO 100 luminance output value = 256 is σ2 = 13. The range of the luminance output value that is regarded as the same subject may be determined based on the idea of 2σ, 3σ, or 4σ with respect to the variance.
From the values determined in this way, for example, as shown in FIG. 4, a synthesis determination threshold value for determining a weighted synthesis ratio for Y = 256 of ISO 100 is generated.
In FIG. 4, the horizontal axis is a signed difference value,
(Signed difference value) = (Comparison image) − (Reference image)
Sought by. In addition, the vertical axis in FIG.

From FIG. 4, it is determined that the region where the signed difference value is 3σ≈ ± 10.8 is the same subject, and is weighted equally at a ratio of 1: 1 to obtain a weighted average. In addition, it is determined that the area of + 4σ = + 14.4 or more is clearly changed to a high luminance, and the weighted average is taken with the ratio of the comparison image as 100%. That is, the reference image is replaced with a comparative image, which corresponds to comparatively bright processing. In the region where −4σ = −14.4 or less, it is determined that the subject has changed to low luminance. However, since it does not make sense in the comparatively bright process, the ratio of the comparative image is 0%, that is, the reference image. Output as is.
In the example of FIG. 4, the composite ratio curve for determining the composite ratio is approximated by a broken line at four points, but may be a curve that changes more smoothly using a lookup table or the like. However,
[ISO sensitivity] x [Luminance output value]
On the other hand, since a composite ratio curve is set for each, it is desirable to use a curve that can be calculated by simple calculation rather than being stored as a table in the memory.

  A composite ratio curve is prepared as described above, and a comparative image is acquired (step S103). Further, noise processing is performed on the reference image / comparison image as necessary (step S104). The noise processing in step S104 will be described in detail later.

Next, a difference image between the reference image and the comparison image is generated (step S105). In this case, the difference value is always signed, and is obtained for each pixel, and is assumed to be a difference image including the signed difference value for each pixel.
The obtained difference value is applied to the composition ratio curve indicating the composition threshold generated earlier, and the composition ratio is obtained for each pixel (step S106).
Through the processing in step S106, all pixels are classified into those that are to be replaced, those that take a weighted average, and those that do nothing.
Finally, a weighted average image obtained by performing a weighted average calculation for each pixel is generated according to the obtained composition ratio (step S107).
Then, it is determined whether there is a next photographed image (step S108).
If it is determined in step S108 that there is a next captured image, the process returns to step S103, and a comparative image is acquired again. If it is determined in step S108 that there is no next captured image, the weighted average image is used as the final output result. The process ends.
Here, a further explanation will be supplemented regarding the noise processing in step S104 described above.

For example, in the case of high ISO sensitivity, the dispersion of output becomes too large, so that the accuracy of synthesis determination is lowered, and it is expected that appropriate comparative bright synthesis cannot be performed as it is. Therefore, in a high ISO sensitivity region of a certain level or more, noise processing by noise reduction is performed on the reference image / comparison image before the difference image generation processing in step S105 to reduce noise components and reduce output dispersion. It is effective to keep. The algorithm used for noise reduction at this time may be arbitrarily selected, but if possible, it is preferable to use an edge preserving filter rather than a simple smoothing filter.
FIG. 5 shows an example of comparison of output variances before and after noise processing by a filter. In FIG. 5, for example, the dispersion σ2 is improved from 42 to 28 in the dark part where the output is the lowest and the dispersion σ2 is improved from 144 to 115 in the bright part where the output is the highest. Therefore, by performing noise processing as necessary in step S104 before generating the difference image, the spread of the distribution due to noise can be substantially reduced, so that the accuracy of the composite determination can be improved. .

In addition, when noise processing is performed as described above, an image used for image synthesis may be different from an image used for noise processing and difference image generation. In other words, for example, a difference image C is generated by comparing each pixel for images A ′ and B ′ obtained by performing noise processing with a filter or the like on the reference image A and the comparison image B, and for each pixel obtained thereby. When the reference image A and the comparative image B (not the images A ′ and B) are synthesized based on the synthesis ratio, a high-quality image synthesis result that does not affect the image quality due to noise processing is obtained.
In step S102 for generating the composite ratio curve, ideally, not only the set imaging sensitivity (sensor gain) such as ISO sensitivity, but also white balance and peripheral light amount correction related to the relative ratio of the gain of each color component. It is desirable to consider the gain of the imaging system such as the gain given for the reason.
Here, a problem that occurs when simple comparative bright combination using a RAW image is performed without applying the present invention will be described.
Comparative bright synthesis is a process of comparing the output of pixels with the same coordinates in a plurality of images and replacing them with one having a higher luminance. In contrast, comparative dark synthesis is the process of comparing pixels with the same coordinates in a plurality of images. This is a process of comparing the outputs and replacing them with ones with lower brightness. In general, a RAW image has undergone only the minimum necessary image processing. Therefore, the output is not necessarily the same even if the same subject is shot under the same shooting condition setting under the influence of shot noise and random noise. It will not be. For this reason, when a simple comparative bright combination is performed, a phenomenon occurs that the noise gradually approaches the maximum value of the noise distribution.

This will be described with reference to FIG. For example, a normal distribution of σ ² = 5 is assumed as a noise distribution model.
If the output of a pixel is affected by noise and fluctuates every time it is photographed in this distribution, a simple comparative light synthesis process will gradually approach the maximum value of the noise distribution. I understand. On the other hand, for example, a process of performing comparatively bright synthesis at a value higher than the maximum value of the noise distribution and nothing is performed at a lower value, thereby asymptotically approaching the maximum value of the noise distribution. It can be improved (in the case of comparative dark synthesis processing, the opposite is true).
Next, FIG. 7 shows a processing result when a simple comparative bright combination is performed on an image obtained by photographing a Macbeth Chart under a D65 light source that is substantially equivalent to a standard light source. .
FIG. 7 shows an output value (♦) of a certain pixel in the black patch portion which is the darkest portion of the Macbeth chart, a result obtained by performing a simple comparison light synthesis (■), and 200 × 200 [ px (pixel)] average (平均) is compared for each of the 20 consecutive shot images. According to FIG. 7, it can be seen that a simple comparative bright combination is asymptotic toward the maximum value of the noise distribution. As a result, an offset of approximately 40 [LSB] is added to the output that should be originally (average of 200 × 200 [px]). This problem occurs regardless of the brightness of the subject, so that it feels as a whole as an image, and white balance (WB) processing is performed with an offset, so that the image becomes magenta. End up.

In order to avoid this, instead of performing a simple comparative bright combination, a combination determination process is performed using estimation data indicating how much noise components can be included in the target pixel of the reference image (the combined image). Is required. Further, when it is determined that the subject is the same instead of just determining whether or not the replacement is to be performed, the effect of noise reduction can be expected at the same time by performing synthesis by weighted average.
For example, in this case, the dispersion of the black patch 200 × 200 [px] is σ = 42. Therefore, for example, as shown in FIG. 8, if the difference is within 2σ = 84, the weighted average is weighted equally, and if the difference is more than that, the high luminance is replaced (that is, comparatively bright combination). By applying a process of doing nothing in the case of low luminance, as shown in FIG. 9, it is possible to make a behavior that fluctuates in the vicinity of the average value without asymptotically approaching the maximum value of the distribution.
In this case, for example, a constant multiple of the noise distribution (σ) may be used. However, if it is too small, the tendency to asymptotically approach the maximum value becomes strong. Since the effect of (combination) may be reduced, it is necessary to adjust appropriately.

As a method of determining the threshold curve, in the case of Patent Document 1 (Japanese Patent Laid-Open No. 2011-199786), the horizontal axis indicates the difference absolute value and the vertical axis indicates the composition ratio in relation to FIG. However, in such a form, the horizontal axis needs to be a signed difference value because it cannot cope with comparatively bright combination processing (comparative dark combination processing). For example, if it is made into FIG. 10, it will become comparatively bright composition, and if it is made to FIG. 11, it will become a threshold curve which functions as comparative dark composition.
The above-described imaging apparatus according to the present invention has been made based on such considerations, and has a built-in image synthesis function such as comparative bright synthesis and comparative dark synthesis, and further reduces the quality of the synthesized image due to noise distribution. Can be effectively suppressed.

DESCRIPTION OF SYMBOLS 1 Optical system 2 Analog front end part 3 Signal processing part 4 SDRAM (Synchronous dynamic random access memory)
5 Liquid crystal display (LCD)
6 Memory card 7 ROM (Read only memory)
DESCRIPTION OF SYMBOLS 8 Operation part 9 Driver 11 Lens group 12 Aperture 13 Shutter 14 Image pick-up element 21 Timing generator 22 CDS circuit 23 AGC circuit 24 A / D conversion circuit 31 Control part (CPU)
32 Image sensor interface (Imaging device I / F)
33 Memory Controller 34 YUV Conversion Unit 35 Resize Processing Unit 36 Display Output Control Unit 37 Data Compression Unit 38 Media Interface (Media I / F)

JP 2011-199786 A

Claims (10)

An imaging means for photoelectrically converting a subject optical image by an imaging device to obtain a captured image composed of electronic image data;
Noise estimation means for estimating a noise component included in the captured image obtained by the imaging means;
Image processing means for obtaining a signed differential value between a plurality of captured images obtained under the same conditions and the same composition by the imaging means;
Combining ratio determining means for determining a different combining ratio between positive and negative based on a noise component by the noise estimating means for the signed difference value obtained by the image processing means;
An image pickup apparatus comprising: an image composition unit configured to perform weighted averaging of the plurality of photographed images based on the composition ratio obtained by the composition ratio determination unit. Wherein the image processing means, the imaging apparatus according to claim 1, characterized in that it comprises a means for performing a noise low reduction process of the plurality of captured images prior to obtaining the signed difference value.   The image processing means includes means for calculating a signed difference value for each corresponding pixel of a plurality of photographed images obtained under the same condition and the same composition by the imaging means, and the image synthesizing means includes: The imaging apparatus according to claim 1, further comprising means for performing synthesis by weighted average for each corresponding pixel of the plurality of captured images.   The noise estimation means includes means for calculating a noise component distribution according to at least one of a set photographing sensitivity, white balance, and peripheral light amount correction amount. The imaging device according to any one of the above.   The synthesis ratio determining means determines a synthesis ratio that is approximately equally weighted in an area where the signed difference value is estimated to be distributed with a large amount of noise components by the noise estimation means, and the signed difference value is The imaging according to any one of claims 1 to 4, further comprising means for determining a composite ratio that gives priority to weighting high luminance in a region on the positive side corresponding to high luminance. apparatus.   The synthesis ratio determining means determines a synthesis ratio that is approximately equally weighted in an area where the signed difference value is estimated to be distributed with a large amount of noise components by the noise estimation means, and the signed difference value is The imaging according to any one of claims 1 to 5, further comprising means for determining a composite ratio that gives priority to weighting of low luminance in a negative large region corresponding to low luminance. apparatus.   The imaging according to any one of claims 1 to 6, wherein the synthesis ratio determining means includes means for independently determining a synthesis ratio for each color of the imaging device of the imaging means. apparatus. The image synthesizing unit is different from a plurality of images as a result of performing the noise reduction process used for calculating a signed difference value in the image processing unit, and a plurality of captured images corresponding to the plurality of images The imaging apparatus according to claim 2, further comprising: a unit that uses a weighted average for synthesis.   9. The imaging apparatus according to claim 8, wherein the image synthesizing unit includes a unit for subjecting a plurality of photographed images not subjected to the noise reduction processing by the image processing unit to image synthesis based on a weighted average.   The plurality of photographed images to be processed by the image processing means and the image composition means are images obtained in time series under the same conditions and the same composition by the imaging means. The imaging device according to any one of claims 1 to 9.

Images