WO2011145297A1 - 撮像装置、画像処理装置、画像処理方法、および画像処理プログラム - Google Patents
撮像装置、画像処理装置、画像処理方法、および画像処理プログラム Download PDFInfo
- Publication number
- WO2011145297A1 WO2011145297A1 PCT/JP2011/002637 JP2011002637W WO2011145297A1 WO 2011145297 A1 WO2011145297 A1 WO 2011145297A1 JP 2011002637 W JP2011002637 W JP 2011002637W WO 2011145297 A1 WO2011145297 A1 WO 2011145297A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- blur kernel
- blur
- unit
- kernel
- Prior art date
Links
- 238000012545 processing Methods 0.000 title claims abstract description 101
- 238000003672 processing method Methods 0.000 title claims description 9
- 238000003384 imaging method Methods 0.000 claims description 100
- 238000000034 method Methods 0.000 claims description 83
- 238000006243 chemical reaction Methods 0.000 claims description 51
- 230000008569 process Effects 0.000 claims description 48
- 238000001514 detection method Methods 0.000 claims description 15
- 230000002194 synthesizing effect Effects 0.000 claims description 5
- 230000015572 biosynthetic process Effects 0.000 claims description 4
- 238000003786 synthesis reaction Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 description 18
- 230000008859 change Effects 0.000 description 15
- 238000010586 diagram Methods 0.000 description 14
- 230000033001 locomotion Effects 0.000 description 13
- 238000011156 evaluation Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 11
- 239000000203 mixture Substances 0.000 description 9
- 239000002131 composite material Substances 0.000 description 7
- 239000011159 matrix material Substances 0.000 description 7
- 238000012937 correction Methods 0.000 description 6
- 230000007246 mechanism Effects 0.000 description 6
- 238000005457 optimization Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 5
- 230000006837 decompression Effects 0.000 description 4
- 230000004069 differentiation Effects 0.000 description 4
- 230000006866 deterioration Effects 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/68—Control of cameras or camera modules for stable pick-up of the scene, e.g. compensating for camera body vibrations
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/68—Control of cameras or camera modules for stable pick-up of the scene, e.g. compensating for camera body vibrations
- H04N23/681—Motion detection
- H04N23/6811—Motion detection based on the image signal
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/68—Control of cameras or camera modules for stable pick-up of the scene, e.g. compensating for camera body vibrations
- H04N23/682—Vibration or motion blur correction
- H04N23/683—Vibration or motion blur correction performed by a processor, e.g. controlling the readout of an image memory
Definitions
- the present invention relates to an imaging apparatus, an image processing apparatus, an image processing method, and an image processing program that generate a restored image in which blur caused by camera shake during imaging is reduced.
- noise When an image is acquired (captured) with a digital camera, noise may be added to the image due to the characteristics of the CCD (Charge-Coupled Device) or CMOS readout circuit and the characteristics of the transmission path.
- image blur due to out of focus (out-of-focus) at the time of imaging or blur due to camera shake or the like occurs.
- the image is deteriorated due to the noise caused by the artificial operation at the time of photographing being added to the noise due to the characteristic unique to the captured image.
- the blurring of the image due to the movement of the camera during shooting (exposure) may be referred to as “motion blur”.
- the noise is also amplified. Therefore, the signal is buried in the noise, and the image occupies most of the noise in many cases.
- the exposure time by increasing the exposure time, a lot of light generated on the spot is accumulated and an image with less noise can be obtained. In this case, the signal is not buried with noise, but there is a problem that the image is likely to be blurred due to camera shake.
- optical camera shake correction such as lens shift and sensor shift.
- the other is a method (restoration method by signal processing) in which the direction / magnitude of blur is obtained by the obtained image or sensor and the image is restored therefrom by signal processing.
- the restoration method by signal processing is disclosed in, for example, Patent Document 1, Patent Document 2, Non-Patent Document 1, and Non-Patent Document 2.
- Patent Document 1 Patent Document 1
- Patent Document 2 Non-Patent Document 1
- Non-Patent Document 2 Non-Patent Document 2.
- the luminance distribution of the image formed on the imaging surface of the imaging element is represented by I (x, y).
- the coordinates (x, y) are two-dimensional coordinates indicating the position of the pixel (photosensitive cell) on the imaging surface.
- x and y are integers that satisfy the relationship of 0 ⁇ x ⁇ M ⁇ 1 and 0 ⁇ y ⁇ N ⁇ 1, respectively.
- the position of each pixel constituting the image can be specified by coordinates (x, y).
- the origin of coordinates (0, 0) is placed at the upper left corner of the image, and the x axis extends in the vertical direction and the y axis extends in the horizontal direction.
- the method of taking the coordinates is arbitrary.
- the luminance distribution of an image without blur is L (x, y), and PSF that defines blur, that is, a “Point Spread Function” is PSF (x, y).
- PSF a “Point Spread Function”
- the noise is n (x, y)
- the following formula 1 is established.
- the symbol “*” indicates a two-dimensional convolution operation (convolution).
- the camera shake point spread function PSF (x, y) depends on the camera shake trajectory during shooting (exposure). Since the locus of camera shake differs for each camera shot, PSF (x, y) also changes for each camera shot.
- the PSF convolution operation that defines blur due to camera shake is performed by a linear filter.
- a linear filter for a two-dimensional convolution operation is usually represented by an integration kernel (kernel) consisting of coefficient rows having a size of N ⁇ N pixels.
- N is an integer of 3 or more.
- the PSF that defines blur can be expressed by a blur kernel. In order to restore an image with reduced blur (including an image from which blur is removed) from a blurred image, it is necessary to estimate a blur kernel that defines the blur.
- Non-Patent Document 1 discloses that a multi-scale estimation scheme is used when a blur kernel and a restored image are estimated from a blurred image. ing.
- a blur kernel having a size of 3 ⁇ 3 pixels is first estimated using a degraded image with low resolution. Then, by gradually increasing the resolution of the deteriorated image used for estimation, the resolution of the blur kernel is also increased.
- FIGS. 19A to 19H are diagrams for explaining the method disclosed in Non-Patent Document 1.
- FIG. The upper part of FIG. 19A shows a 3 ⁇ 3 pixel blur kernel, and the lower part shows a restored image having a corresponding resolution. The same applies to FIGS. 19B to 19H, but the resolution is gradually improved.
- Patent Document 1 discloses a method for obtaining a restored image with reduced blur from two images acquired in succession.
- a blur kernel is estimated from each of the first image and the second image acquired by exposure that is shorter than the first image.
- One restored image is generated from the synthesized blur kernel obtained by synthesizing the two estimated blur kernels and the synthesized image obtained by synthesizing the first image and the second image.
- the restoration process is performed only when the degree of blur included in the first image is large, and the restoration process is not performed when the degree of blur is small.
- the final blur kernel size is preset to a fixed value. Therefore, in order to correctly perform restoration processing on a degraded image with a large degree of blurring, it is necessary to set the blur kernel size to a large value. Degraded images that contain large blurs that exceed the size of the set blur kernel cannot be restored correctly.
- An object of the present invention is to reduce the amount of calculation required for restoration processing when the degree of blurring of a degraded image is small, and to obtain a restored image correctly even when the degree of blurring of the degraded image is large. It is to provide an image restoration technique that can be used.
- An imaging apparatus is an imaging apparatus that generates a restored image in which blur caused by camera shake at the time of imaging is reduced.
- a first image and an exposure time shorter than the first image And an image processing unit that performs a restoration process using the first image and the second image acquired by the imaging unit.
- the image processing unit generates a restored image using a blur kernel determining unit that determines a blur kernel that defines blur due to camera shake in the first image acquired by the imaging unit, and the determined blur kernel
- An image restoration unit, and the blur kernel determination unit changes a size of the blur kernel in accordance with an exposure time when the first image is acquired.
- the blur kernel determining unit increases the size of the blur kernel as the exposure time increases.
- the imaging apparatus further includes a conversion table that defines a correspondence relationship between the exposure time and the size of the blur kernel, and the blur kernel determination unit is configured to determine the blur kernel based on the conversion table. Determine the size.
- the conversion table defines a correspondence relationship between the exposure time and a magnification from a reference size of the blur kernel.
- the blur kernel determination unit determines a second blur kernel that defines blur due to camera shake in the second image
- the image restoration unit Generates the restored image using the first blur kernel and the second blur kernel.
- the image processing unit includes an image combining unit that combines the first image and the second image to generate a third image
- the restoration unit includes the first blur.
- the restored image is generated from the third image using a kernel and the second blur kernel.
- the image processing unit includes an image synthesis unit that synthesizes a plurality of images, and the image restoration unit uses the first blur kernel to perform a first restoration process from the first image. And generating an image of the second restoration process from the second image using the second blur kernel, and the image composition unit includes the image of the first restoration process and the second image of the second restoration process.
- the restored image is generated by combining the images in the restoration process.
- the imaging apparatus further includes a blur detection unit that obtains a degree of blur of the first image by comparing the first image and the second image acquired by the imaging unit, When the degree of blur of the first image obtained by the blur detection unit is larger than a reference value, a restoration process is performed in the image processing unit, and the degree of blur of the first image is greater than the reference value. If it is smaller, the restoration processing in the image processing unit is not performed.
- the blur kernel determining unit further changes the size of the blur kernel in accordance with a zoom magnification at the time of acquiring the first image.
- the image processing apparatus is configured to acquire the first image and the first image acquired by the imaging apparatus that acquires a second image acquired with an exposure time shorter than the first image in one shooting.
- the restoration process is performed using the second image and the second image.
- the image processing apparatus includes: an image acquisition unit that acquires an exposure time when acquiring the first image, the second image, and the first image; and the first image acquired by the imaging unit.
- a blur kernel determining unit that determines a blur kernel that defines blur due to camera shake in an image; and an image restoring unit that generates a restored image using the determined blur kernel, wherein the blur kernel determining unit includes: The size of the blur kernel is changed according to the exposure time when acquiring one image.
- the image processing program is the first image acquired by the imaging device that acquires the first image and the second image acquired with an exposure time shorter than the first image in one shooting.
- the program acquires, from a computer, information indicating the first image, the second image, and an exposure time when acquiring the first image, and an exposure time when acquiring the first image. Determining a size of a blur kernel that defines blur due to camera shake in the first image based on: determining the blur kernel; generating a restored image using the determined blur kernel; Is executed.
- the image processing method is based on the step of acquiring the first image, the second image, and information indicating an exposure time at the time of acquiring the first image, and an exposure time at the time of acquiring the first image. Determining a blur kernel size defining blur due to camera shake in the first image, determining the blur kernel, and generating a restored image using the determined blur kernel .
- the amount of calculation required for the restoration process can be suppressed when the degree of blurring of the deteriorated image is small, and a restored image can be obtained correctly even when the degree of blurring of the deteriorated image is large. it can.
- FIG. 2 It is a figure which shows an example of schematic structure of the imaging device of this invention. It is a figure which shows an example of schematic structure of the image processing apparatus of this invention. It is a flowchart which shows the outline of the flow of the image restoration process by an image processing part or an image processing apparatus.
- A) is a figure which shows an example of the blur kernel which has a size of 3x3 pixel
- (b) is a figure which shows the example of an arrangement
- (A) is a figure which shows an example of the coefficient matrix of a 9 * 9 pixel blur kernel
- (b) is a black value which fills the element which has a non-zero value among the coefficients shown to (a), and is a zero value
- (A) is a diagram showing that the size of the blur kernel is set to a small value when the assumed camera shake is small, and (b) is the size of the blur kernel when the assumed camera shake is large.
- FIG. 1 is a block diagram illustrating a schematic configuration of an imaging apparatus according to an embodiment of the present invention.
- FIG. 2 is a diagram schematically illustrating a schematic configuration of an imaging element 10 and a photographing lens 20 in the imaging unit 100. 2 is a schematic top view of an imaging surface 10a of the imaging element 10.
- FIG. It is a figure which shows the example of the time change of the electric charge amount accumulate
- FIG. It is a block diagram which shows the structural example of the image process part 220 shown in FIG. (A) is a figure which shows an example of the conversion table in embodiment, (b) is a figure which shows the other example of the conversion table in embodiment.
- FIG. 1 is a figure which shows an example of the conversion table regarding the combination of zoom magnification and exposure time
- (b) is a figure which shows the other example of the conversion table regarding the combination of zoom magnification and exposure time.
- (A)-(h) is a figure for demonstrating the method currently disclosed by the nonpatent literature 1.
- size of an image or blur kernel is used in the same meaning as “number of pixels” or “pixel size”.
- FIG. 1A is a diagram illustrating an example of a schematic configuration of an imaging apparatus of the present invention.
- the imaging apparatus includes an imaging unit 100 that acquires an image, and an image processing unit 220 that performs a restoration process using the image acquired by the imaging unit 100.
- the imaging unit 100 acquires a first image and a second image acquired with an exposure time shorter than that of the first image in one shooting.
- the image processing unit 220 includes a blur kernel determining unit 226 that determines a blur kernel that defines blurring of the first image acquired by the imaging unit 100, and an image restoring unit 224 that generates a restored image.
- the blur kernel determination unit 226 can change the size of the blur kernel in accordance with the exposure time when acquiring the first image.
- FIG. 1B is a diagram showing an example of a schematic configuration of the image processing apparatus of the present invention.
- the image processing device 250 uses the two images acquired by the imaging device that acquires the first image and the second image acquired with an exposure time shorter than the first image in one shooting. Perform the restoration process.
- the image processing apparatus 250 includes an image acquisition unit that acquires information indicating an exposure time when acquiring the first image, the second image, and the first image, and a blur kernel that defines blurring of the first image
- the blur kernel determining unit can change the size of the blur kernel in accordance with the exposure time when acquiring the first image.
- FIG. 1C is a flowchart showing an outline of the flow of image restoration processing by the image processing unit 220 or the image processing device 250.
- step S101 two deteriorated images and exposure time information acquired by the imaging device are acquired.
- step 102 a blur kernel size that defines the degree of blurring of the degraded image is determined.
- step S103 a blur kernel is determined.
- step 104 a restored image is generated using the determined blur kernel.
- FIG. 2A shows an example of a blur kernel having a size of 3 ⁇ 3 pixels.
- This blur kernel defines an example of blurring when the camera moves by 3 pixels in the horizontal direction due to camera shake during exposure.
- FIG. 2B is a diagram illustrating an arrangement example of pixel values of an image without blur. This unblurred image is assumed to have a size of 5 ⁇ 5 pixels.
- the 25 numerical values shown in FIG. 2B are examples of luminance values of pixels constituting the image.
- the degraded image is subjected to a two-dimensional convolution operation by the blur kernel of FIG. 2A on the image of FIG. It is obtained by doing.
- an image (degraded image) I acquired by photographing is expressed by the following Expression 2.
- Three of the nine coefficients in the blur kernel have values of 0.2, 0.4, and 0.4, while the other six elements have values of zero.
- the coefficients of 0.2, 0.4, and 0.4 are the positions (2, 1), (2, 2), and (2, 3) in the image, respectively.
- the sum is obtained by multiplying by the pixel value.
- the coefficients of 0.2, 0.4, and 0.4 are respectively located at positions (2, 2), (2, 3), and (2, 4) in the image. The sum is obtained by multiplying the pixel values.
- an image after the convolution operation that is, an image having blur due to camera shake (degraded image) ) Will be determined.
- FIG. 4A shows an example of a 9 ⁇ 9 pixel blur kernel coefficient matrix. Of this coefficient matrix, the sum of coefficients that are not zero is standardized to be equal to one.
- FIG. 4B is a diagram in which elements having non-zero values among the coefficients shown in FIG. 4A are painted black and elements having zero values are painted white.
- the set of black elements in FIG. 4B corresponds to a hand movement locus.
- the set of black elements in FIG. 4B has different patterns depending on the hand movement trajectory during the exposure.
- the trajectory is constituted by a straight line or a curve connecting the start point and the end point.
- a “curve” included in a part thereof is also obtained by connecting two pixels with a straight line.
- the size of the blur kernel is not set in advance before shooting, but is adaptively determined based on the exposure time at the time of shooting.
- FIG. 5 is a graph illustrating the relationship between the degree of image blur and the exposure time. As illustrated, generally, the longer the exposure time, the greater the degree of image blurring. This is because as the exposure time is longer, more blur due to camera shake is accumulated.
- FIG. 6A shows the size of the blur kernel in an image with a relatively small degree of blurring. This corresponds to a case where the exposure time is relatively short. At this time, the size of the blur kernel is set to a relatively small size.
- FIG. 6B shows an example of the size of the blur kernel in an image with a relatively large degree of blur. This corresponds to the case where the exposure time is relatively long. At this time, the size of the blur kernel is set to a relatively large size.
- the size of the blur kernel suitable for the degree of blur estimated from the exposure time is set. As a result, it is possible to prevent the calculation amount from being increased more than necessary in the restoration process for an image with a small degree of blur. Also, unlike the prior art, since the size of the blur kernel is not fixed in advance, it can be correctly restored even for an image with a large degree of blur.
- FIG. 7 is a block diagram illustrating a schematic configuration of the imaging apparatus according to the present embodiment.
- the imaging apparatus of the present embodiment is a digital electronic camera having a camera shake correction function.
- the imaging apparatus includes an imaging unit 100, a signal processing unit 200 that performs various signal processing, an imaging display unit 300 that displays an image acquired by imaging, a recording medium 400 that records image data, and a system that controls each unit. And a control unit 500.
- the imaging unit 100 includes an imaging device (image sensor) 10 including a plurality of photosensitive cells (photodiodes) arranged on the imaging surface, a shutter 15 having a diaphragm function, and an image on the imaging surface of the imaging device 10. And a photographic lens 20 for forming.
- a typical example of the image sensor 10 is a CCD or CMOS sensor.
- the photographic lens 20 in the present embodiment has a known configuration, and is actually a lens unit that includes a plurality of lenses.
- the shutter 15 and the photographing lens 20 are driven by a mechanism (not shown), and operations necessary for optical zooming, automatic exposure (AE: Auto Exposure), and automatic focus (AF: Auto Focus) are executed.
- the imaging unit 100 includes an imaging element driving unit 30 that drives the imaging element 10.
- the image sensor drive unit 30 is configured by an LSI such as a CCD driver, for example.
- the image sensor drive unit 30 drives the image sensor 10 to read an analog signal from the image sensor 10 and convert it into a digital signal.
- the signal processing unit 200 includes an image processing unit (image processor) 220, a memory 240, an interface (IF) unit 260, and a conversion table 280.
- the conversion table 280 is a table that defines the relationship between the exposure time at the time of shooting and the size of the blur kernel. Based on the information in the conversion table 280, the image processing unit 220 changes the size of the blur kernel.
- the conversion table 280 may be stored in the memory 240 or another recording medium. Hereinafter, the information recorded in the conversion table is referred to as “conversion table information”.
- the signal processing unit 200 is connected to a display unit 300 such as a liquid crystal display panel and a recording medium 400 such as a memory card. The recording medium can be removed from the imaging device.
- the image processing unit 220 performs various signal processes necessary for operations such as color correction, resolution change, automatic exposure, automatic focus, and data compression, and also performs a degraded image restoration process according to the present invention.
- the image processing unit 220 is preferably realized by a combination of hardware such as a known digital signal processor (DSP) and software that executes image processing including image restoration processing according to the present invention.
- DSP digital signal processor
- the memory 240 is configured by a DRAM or the like. The memory 240 records the image data obtained from the imaging unit 100 and temporarily records the image data subjected to various image processing by the image processing unit 220 and the compressed image data. These image data are converted into analog signals and then displayed on the display unit 300 or recorded on the recording medium 400 via the interface unit 260 as digital signals.
- a system control unit 500 including a central processing unit (CPU) (not shown) and a flash memory.
- CPU central processing unit
- the imaging apparatus of the present embodiment may include known components such as a viewfinder, a power source (battery), and a flashlight, but a description thereof is omitted because it is not particularly necessary for understanding the present invention.
- FIG. 8 schematically shows a schematic configuration of the imaging element 10 and the lens 20 in the imaging unit 100.
- the image sensor 10 has an imaging surface 10a.
- a plurality of photosensitive cells are arranged on the imaging surface 10a.
- FIG. 9 is a top view schematically showing the imaging surface 10a.
- the photosensitive cells 11 are arranged in a matrix.
- sequence form of the photosensitive cell 11 is not restricted to the example shown in figure,
- the planar shape of each photosensitive cell is not limited to a square.
- a primary color filter or a complementary color filter is typically disposed on each photosensitive cell 11.
- a configuration in which three image sensors are provided and light is separated into three colors of RGB, for example, and then each color light is received by separate image sensors may be employed.
- the imaging unit 100 can acquire an image by imaging.
- the imaging unit 100 according to the present embodiment is configured to acquire two images with different exposure times in one shooting when performing camera shake correction.
- FIG. 10 is a graph illustrating an image acquisition method in this embodiment. In the illustrated graph, the horizontal axis represents time, and the vertical axis represents the amount of accumulated charge in one photosensitive cell. In FIG. 10, only the amount of accumulated charge corresponding to the above-described two exposures is shown, and the amount of accumulated charge corresponding to the image used for preview display is not shown.
- T S a relatively short time
- a first image short-time exposure image
- the second exposure is performed with a time T L longer than T S , and a second image (long exposure image) is acquired.
- a second image long exposure image
- the order of acquisition of the short-time exposure image and the long-time exposure image may be reversed. Since the long-time exposure image obtained in this way is acquired with an exposure time longer than that of the short-time exposure image, the image has a high luminance level and a high degree of blur due to camera shake.
- the short-time exposure image is an image having a low luminance level and a small degree of blur due to camera shake.
- FIG. 11 is a block diagram illustrating a schematic configuration of the image processing unit 220.
- the image processing unit 220 acquires a long-time exposure image, a short-time exposure image, exposure time information, and conversion table information, and outputs a restored image in which blur due to camera shake is reduced compared to the long-time exposure image.
- the image processing unit 220 includes an image acquisition unit 222, a blur detection unit 224, an image synthesis unit 225, a blur kernel determination unit 226, an image restoration unit 228, and a parameter update unit 229.
- the blur kernel determination unit 226 includes a kernel size determination unit 226a, an initial kernel setting unit 226b, and a kernel estimation unit 226c.
- the image acquisition unit 222 acquires a long exposure image, a short exposure image, and exposure time information from the imaging unit 100.
- the blur detection unit 224 compares the long-time exposure image acquired by the image acquisition unit 222 with the short-time exposure image, and obtains the degree of blur included in the long-time exposure image.
- an evaluation value indicating the degree of blur can be obtained by obtaining a motion vector of each corresponding point in two images using a known representative point matching method or block matching method.
- a method for evaluating blur based on the ratio of the intensities of high-frequency components of two images, or a method for evaluating blur by determining the correlation between two images by template matching can be used. These evaluation methods are disclosed in, for example, Japanese Patent Application Laid-Open No. 2009-111596.
- image restoration processing described later is performed only when the evaluation value indicating the degree of blur obtained by the blur detection unit 224 is greater than a preset threshold value.
- the threshold value can be set to an arbitrary value.
- the image composition unit 225 composes the long exposure image and the short exposure image to generate one composite deteriorated image.
- the image composition unit 225 generates a composite deteriorated image by adding luminance values of corresponding pixels of the long-time exposure image and the short-time exposure image.
- the corresponding pixels may not be simply added, but may be added after weighting so that the luminance levels of the long-time exposure image and the short-time exposure image are approximately the same.
- the kernel size determination unit 226 a determines the size of the blur kernel based on the exposure time information acquired from the imaging unit 100 and the conversion table information recorded in the conversion table 280. Specifically, the “blur kernel magnification” corresponding to the exposure time is obtained from the conversion table, and a value obtained by multiplying the obtained blur kernel magnification by the reference size is determined as the size of the blur kernel.
- the “exposure time” means the total exposure time T L + T S when a long exposure image and a short exposure image are acquired.
- the “reference size” may be set in advance in the imaging apparatus as a fixed value, or may be manually set by the user.
- the size of the blur kernel is manually set by the user, it is possible to set a more preferable reference size in consideration of the difference in the degree of blur due to the difference in user. Alternatively, a different size may be determined for each shooting based on various shooting parameters at the time of image acquisition. Details of the conversion table will be described later.
- the initial kernel setting unit 226b sets an initial blur kernel necessary for the restoration process.
- the initial blur kernel may be set manually or may be set in advance as a fixed coefficient matrix. Alternatively, a different initial blur kernel may be determined for each shooting based on various shooting parameters at the time of image acquisition. From the viewpoint of reducing the time required for image processing, the initial blur kernel is preferably close to the actual blur kernel, but the restoration process is possible even if it is not necessarily close.
- the size of the initial blur kernel can be set to a different value depending on the restoration algorithm described later. For example, when the algorithm disclosed in Non-Patent Document 1 is used, the size of the initial blur kernel is set to a relatively small value such as 3 ⁇ 3 pixels. When an algorithm that does not change the size of the blur kernel is used in the restoration process, the size of the initial blur kernel is set to a size determined by the kernel size determination unit 226a.
- the image restoration unit 228 generates a restored image from the synthesized degraded image generated by the image synthesis unit 225 using the initial blur kernel.
- the kernel estimation unit 226c estimates a blur kernel based on the synthesized degraded image and the restored image generated by the image restoration unit 224.
- the parameter update unit 229 updates the initial blur kernel with the blur kernel estimated by the kernel estimation unit 226c.
- the updated initial blur kernel is given to the image restoration unit 228, and the above processing by the image restoration unit 228, the kernel estimation unit 226c, and the parameter update unit 229 is repeatedly executed.
- FIG 11 shows an example of a functional block of the image processing unit 220, and the image processing unit 220 can be divided into other functional blocks.
- the image processing unit 220 is also preferably realized by, for example, incorporating image processing software into known hardware.
- FIG. 12A is a diagram illustrating an example of the conversion table in the present embodiment.
- the conversion table defines a correspondence between the exposure time range and the blur kernel magnification.
- the kernel size determination unit 226a should set the blur kernel magnification to A1 when the exposure time (T L + T S ) is within the range from T1 to T2, for example.
- FIG. 12B is a diagram illustrating another example of the conversion table. As in this example, it may be specified that the size of the blur kernel is directly determined according to the exposure time.
- the size of the blur kernel is set to N2 ⁇ N2 (N2: integer of 3 or more).
- N2 integer of 3 or more
- the conversion table is not limited to the above, and any table may be used as long as it indicates the correspondence between the exposure time and the size of the blur kernel.
- the conversion table is created in advance as advance information.
- an example of a conversion table creation method will be described.
- an object such as a chart or a point light source that can measure the movement distance in the image
- the exposure time becomes a certain value
- the movement amount is measured respectively.
- the same measurement is performed by changing other exposure times, and the average of the movement amounts is obtained for each exposure time.
- Information indicating the relationship between the average movement amount obtained in this way and the exposure time is recorded in the conversion table.
- the imaging apparatus may have a function of learning a tendency of image blurring in shooting performed by the user in the past and rewriting the conversion table according to the user.
- the user points the imaging device at the subject, and when the user presses the shutter button halfway, the subject is focused by the autofocus operation.
- “exposure” is started (step S1).
- an image of the subject is formed on the imaging surface of the image sensor 10.
- the exposure is temporarily stopped, signal charges are read by the image sensor 10, and a short-time exposure image is acquired (step S2).
- the second exposure starts and the exposure continues until a relatively long time T L elapses.
- T L elapses the exposure ends, the signal charge is read again by the image sensor 10, and a long-time exposure image is acquired (step S3).
- the order of steps S2 and S3 may be reversed.
- the imaging unit 100 inputs the acquired two images and exposure time information to the image processing unit 220 in the signal processing unit 200.
- the image processing unit 220 performs image restoration processing based on the input information, and generates a restored image in which blur due to camera shake is reduced compared to the long-time exposure image (step S4).
- the generated restored image is sent to the display unit 300 or the recording medium 400 and displayed or recorded (step S5).
- step S41 the image acquisition unit 222 acquires a short-time exposure image and a long-time exposure image.
- step S42 the blur detection unit 224 obtains a blur evaluation value indicating the degree of blur in the two acquired images.
- the blur evaluation value is obtained by using a known method as described above.
- step S43 the size of the blur is determined based on the blur evaluation value obtained by the blur detection unit 224.
- the restoration process is not performed and the process ends.
- the long exposure image is stored in the memory 240 as a restoration result.
- the following steps S44 to S50 are executed.
- step S44 the image acquisition unit 222 acquires exposure time information.
- step S45 the image composition unit 225 combines the short-time exposure image and the long-time exposure image to generate a composite deterioration image.
- step S46 the size of the blur kernel is determined by the kernel size determination unit 226a.
- the kernel size determination unit 226a refers to the conversion table information to determine the size of the blur kernel corresponding to the exposure time (T L + T S ).
- step S47 the initial blur kernel setting unit 222 determines an initial blur kernel.
- step S48 the image restoration unit 228 executes an image restoration process using the synthesized deteriorated image generated in step S45 and the initial blur kernel set in step S47, thereby generating a restored image.
- This image restoration process is executed by a known restoration algorithm.
- the image restoration unit 228 once records the obtained restored image in the memory 240.
- the kernel estimation unit 226c estimates a blur kernel from the restored image, and the parameter update unit 228 updates the previous blur kernel with the estimated blur kernel.
- step S50 it is determined whether or not the blur kernel before and after the update and the change in the restored image before and after the update are smaller than a predetermined threshold value. If the change is greater than or equal to the threshold value, the process of step S48 is performed again, and thereafter steps S48 to S50 are repeated until the change becomes smaller than the threshold value. If the change is smaller than the threshold, it is determined that the process has converged, and the restoration result is stored in the memory 240.
- step S45 may be performed prior to the acquisition of the exposure time information in step S44.
- Non-Patent Document 2 an image restoration method based on a signal processing method disclosed in Non-Patent Document 2 will be described.
- first image restoration is performed based on the initial blur kernel set in step S47.
- the blur kernel at this stage does not necessarily match the true blur kernel (correct answer), but the restored result is closer to the original image than the degraded image.
- the blur kernel is estimated from the first restored image that is the result of the first image restoration. Since the first restored image is closer to the original image than the degraded image, the estimated blur kernel approaches the correct answer.
- the second image restoration is performed using the blur kernel estimated here, that is, the initial blur kernel is updated with the next blur kernel.
- the image restoration unit 228 restores an image based on a given blur kernel (initial value first, update value next) and a deteriorated image.
- I is a degraded image
- L is an unblurred image L
- f is a blur kernel.
- the variables wk, ⁇ 1, and ⁇ 2 are “weights” set manually.
- ⁇ is a set of operators that define what differentiation is applied to the image. Specifically, ⁇ is the sum of zero differentiation, one differentiation (each in x and y directions), two differentiations (twice in the x direction, twice in the y direction, and once in the x and y directions).
- d * is a differential operator.
- ⁇ ⁇ d 0 , d x , dy , d xx , d xy , d yy ⁇ .
- M is a two-dimensional mask, and has a “1” element in a pixel included in a flat area in an image, that is, a local smooth area ( ⁇ ), and a “0” element in other pixels. is doing.
- p is a p-norm operator.
- ⁇ (x) is a function that approximately represents the relationship between the luminance gradient x in a naturally observed image and its distribution density (logarithmic display).
- the first term on the right side of Equation 3 is a term indicating the difference (distance) between the image obtained by convolving the restored image L and the blur kernel f and the degraded image I.
- Equation 3 The second term on the right side of Equation 3 is a term indicating the property of the luminance gradient in the image (called “heavy tail”).
- ⁇ (d x L) and ⁇ (d y L) are histograms of the luminance gradient of the restored image, a sharp peak of the appearance probability appears near the gradient, and the appearance probability decreases as the gradient increases. It has a statistical property.
- the distance from the distribution showing the above statistical properties is calculated for each of the gradient in the x direction and the gradient in the y direction. This statistical property is also used in the method disclosed in Non-Patent Document 1.
- Equation 3 The third term on the right side of Equation 3 is a term for evaluating the flatness using the mask M, the differentially degraded image, and the differentially restored image.
- the gradient value of the luminance is close between the degraded image and the restored image. Therefore, an error in gradient values in the x and y directions is used as an evaluation value.
- the restored image L can be obtained by obtaining L that minimizes the right side of Equation 3 (optimization of L).
- a specific calculation method for optimization of L is disclosed in Non-Patent Document 2.
- the blur kernel estimation is a problem of estimating the blur kernel f using the restored image L and the degraded image I obtained by the image restoration unit 228. By determining f so as to minimize the right side of Equation 4 below, the blur kernel f can be obtained (optimization of f).
- the first term on the right side of Equation 4 corresponds to the first term on the right side of Equation 3, and gives an evaluation criterion indicating whether or not the convolution of the restored image L and the blur kernel f is close to the degraded image I.
- the second term on the right side of Equation 4 is the 1 norm of the blur kernel f.
- the second term is a term based on a concept called sparse coding. This optimization term is used because most elements in the row example of the blur kernel f are 0 (no motion). In the present embodiment, optimization by “interor point method” is performed as in Non-Patent Document 2, and overall optimization can be realized.
- image restoration procedure is not limited to the above example.
- the method disclosed in Non-Patent Document 1 and other blind deconvolution methods can be used.
- An important point in the image restoration in the present embodiment is to change the size of the blur kernel in accordance with the change in the exposure time at the time of shooting. Accordingly, the size of the blur kernel is set to a large value at the time of shooting by long-time exposure in which image blurring is likely to occur, so that the size of the blur kernel can be prevented from exceeding a preset size. On the other hand, the size of the blur kernel is set to a small value at the time of photographing with a short exposure, so that it is possible to prevent the calculation amount from being increased more than necessary.
- processing flow in the image restoration processing S4 in FIG. 13 may have various modes other than the flow shown in FIG.
- any of the flows shown in FIGS. 15 to 17 can be employed.
- each flow will be described.
- step S51 to S54 are the same as steps S41 to S44 in FIG.
- step S55 the blur kernel size in the long exposure image is determined.
- the kernel size determination unit 226a obtains a blur kernel magnification corresponding to the exposure time T L when the long exposure image is acquired from the conversion table.
- the subsequent processing from step S56 to step S59 is the same as the processing from step S46 to step S49 in FIG.
- the restored image is generated by performing the processing from step S56 to step S59 on the long-time exposure image.
- the image composition unit 225 is not necessary.
- step S61 to S64 are the same as steps S41 to S44 in FIG.
- step S65a to step S69a estimation of a blur kernel and generation of a restored image are performed for the long-time exposure image.
- step S65b to step S69b estimation of a blur kernel for a short-time exposure image and generation of a restored image are performed.
- steps S65a to S69a and steps S65b to S69b are the same as the processes in steps S46 to S50 in FIG.
- the blur kernel size in the long exposure image is determined from the conversion table based on the exposure time T L when the long exposure image is acquired.
- the blur kernel size in the short exposure image is determined from the conversion table based on the exposure time T S when the short exposure image is acquired.
- the image composition unit 225 combines the restored image for the long-time exposure image and the restored image for the short-time exposure image to generate a composite restored image.
- the generated composite restored image is recorded in the memory 240 as a final restoration result.
- the image composition unit 225 performs the restored image composition process in step S70 without composing the long exposure image and the short exposure image.
- a blur kernel is estimated for each of the long-exposure image and the short-exposure image, and a composite blur kernel is generated from the two estimated blur kernels.
- a restored image is generated by restoring the synthesized image using the synthesized blur kernel.
- steps S71 to S75 are the same as steps S41 to S45 in FIG.
- steps S76a and S77a the blur kernel size in the long exposure image is determined and the blur kernel is estimated.
- steps S76b and S77b determination of the blur kernel size and estimation of the blur kernel in the long exposure image are performed.
- the blur kernel size in the long exposure image is determined from the conversion table based on the exposure time T L when the long exposure image is acquired.
- the blur kernel size in the short exposure image is determined from the conversion table based on the exposure time T S when the short exposure image is acquired.
- a blur kernel estimation method in steps S77a and S77b for example, a known method such as estimating a blur kernel from the magnitude of a camera shake obtained by a camera shake detection mechanism provided in the imaging apparatus can be used.
- step S78 the obtained two blur kernels are synthesized.
- the image restoration unit 228 generates a restored image by restoring the synthesized image obtained in step S75 using the synthesized blur kernel.
- a restored image can be obtained using a non-blind deconvolution method such as a known Wiener filter method or Richardson-Lucy (RL) method.
- the blur kernel size can be adaptively changed according to the exposure time, and the effect of this embodiment can be obtained.
- the order of individual processes may be different from the order shown in the possible range.
- the imaging apparatus acquires a long-time exposure image first, and does not acquire a short-time exposure image when the magnitude of camera shake at that time is smaller than a predetermined threshold. It may be configured. With such a configuration, a short-time exposure image is acquired and an image restoration process is performed only when camera shake correction is necessary, so that the processing time can be shortened.
- the image processing unit 220 does not necessarily need to include the blur detection unit 224.
- the blur detection unit 224 is not provided, the processing in steps S42 and S43 in FIG. 14 or the corresponding steps in FIGS.
- the same processing can be performed by using a function representing the relationship between the exposure time and the blur kernel size.
- a function representing the relationship between the exposure time and the blur kernel size is used.
- the acquired data is plotted in a two-dimensional space with the exposure time as the horizontal axis and the blur kernel size as the vertical axis.
- linear regression or curve fitting is performed on the plotted data, and the relationship between exposure time and blur kernel size is expressed as a multidimensional function.
- the multidimensional function is obtained in this way, the blur kernel size can be obtained by inputting the exposure time obtained at the time of photographing into the obtained multidimensional function. According to such a method, the conversion table is unnecessary.
- the blur kernel size is determined only by the exposure time at the time of shooting.
- the blur kernel size may be determined in consideration of other information that affects the degree of blurring of the image.
- the degree of image blur is affected not only by the exposure time but also by the zoom magnification at the time of shooting. That is, when the zoom magnification (or focal length) of the optical system is increased in order to shoot a distant subject, the degree of blurring of the image is greater than when the zoom magnification is small. Therefore, when the zoom magnification is increased in order to photograph a distant scene, it is preferable to increase the size of the blur kernel.
- zoom magnification means the ratio of the focal length at the time of photographing to the minimum (wide-angle side) focal length of the optical system in the imaging apparatus.
- the imaging apparatus changes the blur kernel size according to changes in both the exposure time and the zoom magnification.
- both the conversion table regarding the zoom magnification and the conversion table regarding the exposure time may be recorded in the imaging apparatus. For example, if a conversion table similar to the conversion table shown in FIG. 12A is prepared for the zoom magnification, the size of the blur kernel can be determined by multiplying the reference size by the magnification obtained from each table.
- FIG. 18 is a diagram illustrating an example of a table in which these pieces of information are integrated.
- FIG. 18A shows an example of a conversion table that defines the correspondence between the combination of zoom magnification and exposure time and the blur kernel magnification.
- FIG. 18B shows an example of a conversion table that defines the correspondence between the combination of zoom magnification and exposure time and blur kernel size.
- the optimum blur kernel size can be determined in consideration of both the zoom magnification and the exposure time. For example, when the conversion table shown in FIG. 18A is used, when the zoom magnification is in the range from z2 to z3 and the exposure time is in the range from T2 to T3, the blur kernel magnification is Set to C22.
- the image restoration process in the above embodiment can be executed not only by the image processing unit built in the imaging apparatus but also by an image processing apparatus independent of the imaging apparatus.
- each image acquired by the imaging apparatus, exposure time information, and the like are input to the image processing apparatus, and for example, a program defining the processing shown in any of FIGS. 14 to 15 is executed by a computer in the image processing apparatus.
- the restored image can also be obtained.
- the imaging apparatus of the present invention has high industrial applicability when applied to an imaging apparatus in which blurring may occur due to camera shake. Since both the PSF and the restored image can be estimated under a situation where the PSF is unknown, it is possible to obtain an image with less blur without providing a special camera shake prevention mechanism or together with such a camera shake prevention mechanism. become.
- the image processing apparatus of the present invention does not need to be built in the imaging apparatus, and can be configured to receive and process image data acquired by the imaging apparatus.
- Shutter 20 which has a diaphragm function Image pick-up lens 30 Image pick-up element drive part 100 Image pick-up part 200 Signal processing part 220 Image processing part 222 Image acquisition part 224 Blur detection part 225 Image composition part 226 Blur kernel determination part 226a Kernel size determination unit 226b Initial kernel setting unit 226c Kernel estimation unit 228 Image restoration unit 229 Parameter update unit 240 Memory 260 Interface (IF) 280 Conversion table 300 Display unit 400 Recording medium 500 System control unit
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
- Image Processing (AREA)
- Adjustment Of Camera Lenses (AREA)
Abstract
Description
図7は、本実施形態の撮像装置の概略構成を示すブロック図である。本実施形態の撮像装置は、手振れ補正機能を備えるデジタル式の電子カメラである。撮像装置は、撮像部100と、各種信号処理を行う信号処理部200と、撮像によって取得した画像を表示する撮像表示部300と、画像のデータを記録する記録媒体400と、各部を制御するシステム制御部500とを備えている。
ここで、Iは劣化画像、Lはぼやけのない画像L、fはブラーカーネルである。変数wk、λ1、λ2は手動で設定される「重み」である。Θは、どのような微分を画像に施すかを規定する演算子のセットである。Θは、具体的には、0回微分、1回微分(x、y方向それぞれ)、2回微分(x方向に2回、y方向に2回、xとy方向に1回ずつ)の合計6個の微分パラメータを持つ。d*は、微分演算子である。d*を用いてΘを表現すると、Θ={d0、dx、dy、dxx、dxy、dyy}と表される。d*により、輝度情報とエッジ情報の両方を用いた処理を行うことが可能になり、輝度だけでは得られない情報も得ることができる。Mは、2次元のマスクであり、画像中の平坦な領域、すなわち局所的なスムーズな領域(Ω)に含まれる画素では「1」の要素、それ以外の画素では「0」の要素を有している。||・||pは、pノルム演算子である。Φ(x)は、自然に観察される画像中の輝度勾配xとその分布密度(対数表示)との関係を近似的に示す関数である。
11 光感知セル
15 絞り機能を有するシャッタ
20 撮影レンズ
30 撮像素子駆動部
100 撮像部
200 信号処理部
220 画像処理部
222 画像取得部
224 ぼやけ検出部
225 画像合成部
226 ブラーカーネル決定部
226a カーネルサイズ決定部
226b 初期カーネル設定部
226c カーネル推定部
228 画像復元部
229 パラメータ更新部
240 メモリ
260 インターフェース(IF)
280 変換テーブル
300 表示部
400 記録媒体
500 システム制御部
Claims (12)
- 撮像時の手振れによるぼやけを減少させた復元画像を生成する撮像装置であって、
1回の撮影において、第1の画像、および前記第1の画像よりも短い露光時間で取得される第2の画像を取得する撮像部と、
前記撮像部によって取得された前記第1の画像および前記第2の画像を用いて復元処理を行う画像処理部と、
を備え、
前記画像処理部は、
前記撮像部によって取得された前記第1の画像における手振れによるぼやけを規定するブラーカーネルを決定するブラーカーネル決定部と、
決定された前記ブラーカーネルを用いて復元画像を生成する画像復元部と、
を有し、
前記ブラーカーネル決定部は、前記第1の画像を取得する際の露光時間に応じて前記ブラーカーネルのサイズを変化させる、
撮像装置。 - 前記ブラーカーネル決定部は、前記露光時間が長いほど前記ブラーカーネルのサイズを大きくする、請求項1に記載の撮像装置。
- 前記露光時間と前記ブラーカーネルのサイズとの間の対応関係を規定する変換テーブルをさらに備え、
前記ブラーカーネル決定部は、前記変換テーブルに基づいて前記ブラーカーネルのサイズを決定する、請求項1または2に記載の撮像装置。 - 前記変換テーブルは、前記露光時間と、前記ブラーカーネルの基準サイズからの倍率との間の対応関係を規定する、請求項3に記載の撮像装置。
- 前記ブラーカーネルを第1のブラーカーネルとするとき、
前記ブラーカーネル決定部は、前記第2の画像における手振れによるぼやけを規定する第2のブラーカーネルを決定し、
前記画像復元部は、前記第1のブラーカーネルおよび前記第2のブラーカーネルを用いて前記復元画像を生成する、請求項1から4のいずれかに記載の撮像装置。 - 前記画像処理部は、前記第1の画像および前記第2の画像を合成して第3の画像を生成する画像合成部を有し、
前記復元部は、前記第1のブラーカーネルおよび前記第2のブラーカーネルを用いて前記第3の画像から前記復元画像を生成する、請求項5に記載の撮像装置。 - 前記画像処理部は、複数の画像を合成する画像合成部を有し、
前記画像復元部は、前記第1のブラーカーネルを用いて前記第1の画像から第1の復元過程の画像を生成し、前記第2のブラーカーネルを用いて前記第2の画像から第2の復元過程の画像を生成し、
前記画像合成部は、前記第1の復元過程の画像および前記第2の復元過程の画像を合成することによって前記復元画像を生成する、請求項5に記載の撮像装置。 - 前記撮像部によって取得された前記第1の画像および前記第2の画像を比較することによって前記第1の画像のぼやけの程度を求めるぼやけ検出部をさらに備え、
前記ぼやけ検出部によって求められた前記第1の画像のぼやけの程度が予め定められた基準値よりも大きい場合には前記画像処理部における復元処理を行い、前記第1の画像のぼやけの程度が前記基準値よりも小さい場合には前記画像処理部における復元処理を行わない、請求項1から7のいずれかに記載の撮像装置。 - 前記ブラ―カーネル決定部は、さらに、前記第1の画像を取得する際のズーム倍率に応じて、前記ブラ―カーネルのサイズを変化させる、請求項1から8のいずれかに記載の撮像装置。
- 1回の撮影において、第1の画像、および前記第1の画像よりも短い露光時間で取得される第2の画像を取得する撮像装置によって取得された前記第1の画像および前記第2の画像を用いて復元処理を行う画像処理装置であって、
前記第1の画像、前記第2の画像、および前記第1の画像を取得する際の露光時間を取得する画像取得部と、
前記撮像部によって取得された前記第1の画像における手振れによるぼやけを規定するブラーカーネルを決定するブラーカーネル決定部と、
決定された前記ブラーカーネルを用いて復元画像を生成する画像復元部と、
を有し、
前記ブラーカーネル決定部は、前記第1の画像を取得する際の露光時間に応じて前記ブラーカーネルのサイズを変化させる、
画像処理装置。 - 1回の撮影において、第1の画像、および前記第1の画像よりも短い露光時間で取得される第2の画像を取得する撮像装置によって取得された前記第1の画像および前記第2の画像を用いて復元処理を行うためのプログラムであって、
コンピュータに対し、
前記第1の画像、前記第2の画像、および前記第1の画像取得時の露光時間を示す情報を取得するステップと、
前記第1の画像取得時の露光時間に基づいて前記第1の画像における手振れによるぼやけを規定するブラーカーネルのサイズを決定するステップと、
前記ブラーカーネルを決定するステップと、
決定された前記ブラーカーネルを用いて復元画像を生成するステップと、
を実行させるプログラム。 - 1回の撮影において、第1の画像、および前記第1の画像よりも短い露光時間で取得される第2の画像を取得する撮像装置によって取得された前記第1の画像および前記第2の画像を用いて復元処理を行う画像処理方法であって、
前記第1の画像、前記第2の画像、および前記第1の画像取得時の露光時間を示す情報を取得するステップと、
前記第1の画像取得時の露光時間に基づいて前記第1の画像における手振れによるぼやけを規定するブラーカーネルのサイズを決定するステップと、
前記ブラーカーネルを決定するステップと、
決定された前記ブラーカーネルを用いて復元画像を生成するステップと、
を含む画像処理方法。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201180024954.0A CN102907082B (zh) | 2010-05-21 | 2011-05-11 | 摄像装置、图像处理装置、图像处理方法 |
EP11783236.0A EP2574039B1 (en) | 2010-05-21 | 2011-05-11 | Image pickup device, image processing device, image processing method, and image processing program |
JP2012515735A JP5756099B2 (ja) | 2010-05-21 | 2011-05-11 | 撮像装置、画像処理装置、画像処理方法、および画像処理プログラム |
US13/697,686 US9036032B2 (en) | 2010-05-21 | 2011-05-11 | Image pickup device changing the size of a blur kernel according to the exposure time |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2010-117676 | 2010-05-21 | ||
JP2010117676 | 2010-05-21 | ||
JP2010119384 | 2010-05-25 | ||
JP2010-119384 | 2010-05-25 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2011145297A1 true WO2011145297A1 (ja) | 2011-11-24 |
Family
ID=44991418
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2011/002636 WO2011145296A1 (ja) | 2010-05-21 | 2011-05-11 | 撮像装置、画像処理装置、画像処理方法、および画像処理プログラム |
PCT/JP2011/002637 WO2011145297A1 (ja) | 2010-05-21 | 2011-05-11 | 撮像装置、画像処理装置、画像処理方法、および画像処理プログラム |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2011/002636 WO2011145296A1 (ja) | 2010-05-21 | 2011-05-11 | 撮像装置、画像処理装置、画像処理方法、および画像処理プログラム |
Country Status (5)
Country | Link |
---|---|
US (2) | US9071754B2 (ja) |
EP (2) | EP2574038B1 (ja) |
JP (2) | JP5756099B2 (ja) |
CN (2) | CN102907082B (ja) |
WO (2) | WO2011145296A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2013192227A (ja) * | 2012-03-13 | 2013-09-26 | Samsung Electronics Co Ltd | タイル単位に基づいて大きい入力映像の不均一モーションブラーを除去する方法及び装置 |
JP2017070590A (ja) * | 2015-10-08 | 2017-04-13 | キヤノン株式会社 | 画像処理装置及びその制御方法、コンピュータプログラム |
JP2021027491A (ja) * | 2019-08-06 | 2021-02-22 | キヤノン株式会社 | 回転半径演算装置および回転半径演算方法 |
Families Citing this family (38)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2574038B1 (en) * | 2010-05-21 | 2018-10-24 | Panasonic Intellectual Property Corporation of America | Image capturing apparatus, image processing apparatus, image processing method, and image processing program |
KR101896026B1 (ko) * | 2011-11-08 | 2018-09-07 | 삼성전자주식회사 | 휴대 단말기에서 움직임 블러를 생성하는 장치 및 방법 |
KR101844332B1 (ko) * | 2012-03-13 | 2018-04-03 | 삼성전자주식회사 | 블러 영상 및 노이즈 영상으로 구성된 멀티 프레임을 이용하여 비균일 모션 블러를 제거하는 방법 및 장치 |
US9143687B2 (en) * | 2012-03-14 | 2015-09-22 | University Of Dayton | Method of analyzing motion blur using double discrete wavelet transform |
US8917330B1 (en) * | 2012-06-01 | 2014-12-23 | Gopro, Inc. | Cinematic image blur in digital cameras based on exposure timing manipulation |
CN104662435A (zh) | 2012-09-27 | 2015-05-27 | Metaio有限公司 | 确定与用于捕获至少一个图像的捕获装置关联的装置的位置和方向的方法 |
CN104969545B (zh) * | 2013-02-05 | 2018-03-20 | 富士胶片株式会社 | 图像处理装置、摄像装置、图像处理方法以及程序 |
WO2014144820A1 (en) | 2013-03-15 | 2014-09-18 | The Regents Of The University Of Colorado | 3-d localization and imaging of dense arrays of particles |
EP3028220A4 (en) * | 2013-07-29 | 2017-04-05 | Nikon Corporation | Multiple phase method for image deconvolution |
WO2015015966A1 (ja) * | 2013-08-01 | 2015-02-05 | 富士フイルム株式会社 | 撮像装置、撮像方法及び画像処理装置 |
US9918021B2 (en) * | 2013-08-30 | 2018-03-13 | Nikon Corporation | Image processing device that changes extent of image altering by first and second image processing |
US9066014B2 (en) * | 2013-10-11 | 2015-06-23 | Facebook, Inc. | Applying video stabilization to a multimedia clip |
US9299132B2 (en) * | 2013-10-23 | 2016-03-29 | Adobe Systems Incorporated | Automatically determining the size of a blur kernel |
CN103839233B (zh) * | 2014-01-20 | 2017-05-10 | 左旺孟 | 一种相机抖动造成的模糊图像复原方法 |
WO2015191998A1 (en) * | 2014-06-12 | 2015-12-17 | Duke University | Systyem and method for improved computational imaging |
US9646365B1 (en) | 2014-08-12 | 2017-05-09 | Amazon Technologies, Inc. | Variable temporal aperture |
US9749532B1 (en) | 2014-08-12 | 2017-08-29 | Amazon Technologies, Inc. | Pixel readout of a charge coupled device having a variable aperture |
US9787899B1 (en) | 2014-08-12 | 2017-10-10 | Amazon Technologies, Inc. | Multiple captures with a variable aperture |
US9749534B2 (en) * | 2014-12-31 | 2017-08-29 | Canon Kabushiki Kaisha | Devices, systems, and methods for estimation of motion blur from a single image |
US9684970B2 (en) * | 2015-02-27 | 2017-06-20 | Qualcomm Incorporated | Fast adaptive estimation of motion blur for coherent rendering |
CN106033595B (zh) * | 2015-03-13 | 2021-06-22 | 中国科学院西安光学精密机械研究所 | 一种基于局部约束的图像盲去模糊方法 |
KR101596203B1 (ko) | 2015-06-16 | 2016-02-22 | 중앙대학교 산학협력단 | 모션 블러 이미지 복원 방법 및 장치 |
US10007990B2 (en) * | 2015-12-24 | 2018-06-26 | Intel Corporation | Generating composite images using estimated blur kernel size |
US10311560B2 (en) * | 2016-09-07 | 2019-06-04 | Huazhong University Of Science And Technology | Method and system for estimating blur kernel size |
CN108665417B (zh) * | 2017-03-30 | 2021-03-12 | 杭州海康威视数字技术股份有限公司 | 一种车牌图像去模糊方法、装置及系统 |
CN108881703B (zh) * | 2017-05-09 | 2020-07-21 | 杭州海康威视数字技术股份有限公司 | 防抖控制方法和装置 |
CN107395961A (zh) * | 2017-07-07 | 2017-11-24 | 青岛海信移动通信技术股份有限公司 | 一种图像数据的复原方法和装置 |
JP2020115607A (ja) * | 2019-01-17 | 2020-07-30 | 日本電気株式会社 | 画像処理装置、画像処理方法、プログラム |
US11475571B2 (en) * | 2019-03-13 | 2022-10-18 | Canon Kabushiki Kaisha | Apparatus, image processing apparatus, and control method |
US11120538B2 (en) * | 2019-12-27 | 2021-09-14 | Zoox, Inc. | Sensor degradation detection and remediation |
TW202143101A (zh) * | 2020-05-11 | 2021-11-16 | 大陸商上海耕岩智能科技有限公司 | 用於成像系統的模糊函數處理方法及裝置、圖像採集設備、儲存介質、圖像處理方法及裝置、成像設備 |
US20220012853A1 (en) * | 2020-07-09 | 2022-01-13 | Technoimaging, Llc | Joint minimum entropy method for simultaneous processing and fusion of multi-physics data and images |
US11449979B2 (en) * | 2020-10-09 | 2022-09-20 | Applied Materials Israel Ltd. | Measuring a pattern |
CN113347352B (zh) * | 2021-05-21 | 2023-10-03 | Oppo广东移动通信有限公司 | 拍摄参数调整方法及装置、电子设备、计算机存储介质 |
WO2023003556A1 (en) * | 2021-07-22 | 2023-01-26 | Google Llc | Joint video stabilization and motion deblurring |
CN113870134B (zh) * | 2021-09-28 | 2023-04-07 | 南京诺源医疗器械有限公司 | 手持式摄像镜数据处理方法及装置 |
US11425304B1 (en) * | 2021-12-06 | 2022-08-23 | Black Sesame Technologies Inc. | Reducing global motion and rolling shutter in a dual camera system |
CN114928694A (zh) * | 2022-04-25 | 2022-08-19 | 深圳市慧鲤科技有限公司 | 图像获取方法和装置、设备、介质 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005064851A (ja) * | 2003-08-12 | 2005-03-10 | Fuji Photo Film Co Ltd | デジタルカメラ及び固体撮像装置 |
JP2008061217A (ja) * | 2006-07-31 | 2008-03-13 | Ricoh Co Ltd | 画像処理装置、撮像装置、画像処理方法、および画像処理プログラム |
JP2008099025A (ja) | 2006-10-12 | 2008-04-24 | Kyocera Corp | 撮像装置、ぶれ補正方法及び撮像方法 |
JP2009111596A (ja) | 2007-10-29 | 2009-05-21 | Sanyo Electric Co Ltd | 撮像装置 |
Family Cites Families (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6429895B1 (en) * | 1996-12-27 | 2002-08-06 | Canon Kabushiki Kaisha | Image sensing apparatus and method capable of merging function for obtaining high-precision image by synthesizing images and image stabilization function |
JP4298752B2 (ja) | 2004-09-14 | 2009-07-22 | 富士通株式会社 | 画像処理装置、画像処理方法及び画像処理プログラム |
JP2006129236A (ja) | 2004-10-29 | 2006-05-18 | Sanyo Electric Co Ltd | リンギング除去装置およびリンギング除去プログラムを記録したコンピュータ読み取り可能な記録媒体 |
US7643062B2 (en) | 2005-06-08 | 2010-01-05 | Hewlett-Packard Development Company, L.P. | Method and system for deblurring an image based on motion tracking |
CN101366270B (zh) * | 2005-11-29 | 2011-07-13 | 京瓷株式会社 | 成像设备和成像方法 |
JP4916862B2 (ja) * | 2005-12-27 | 2012-04-18 | 京セラ株式会社 | 撮像装置およびその方法 |
JP2008011424A (ja) | 2006-06-30 | 2008-01-17 | Eastman Kodak Co | 画像処理装置及び画像処理プログラム |
CN100550996C (zh) * | 2006-07-31 | 2009-10-14 | 株式会社理光 | 图像处理装置,成像装置以及图像处理方法 |
US7796872B2 (en) * | 2007-01-05 | 2010-09-14 | Invensense, Inc. | Method and apparatus for producing a sharp image from a handheld device containing a gyroscope |
US8306348B2 (en) | 2007-04-24 | 2012-11-06 | DigitalOptics Corporation Europe Limited | Techniques for adjusting the effect of applying kernels to signals to achieve desired effect on signal |
US20090086174A1 (en) | 2007-09-28 | 2009-04-02 | Sanyo Electric Co., Ltd. | Image recording apparatus, image correcting apparatus, and image sensing apparatus |
JP4661922B2 (ja) * | 2008-09-03 | 2011-03-30 | ソニー株式会社 | 画像処理装置、撮像装置、固体撮像素子、画像処理方法およびプログラム |
KR101574733B1 (ko) * | 2008-11-19 | 2015-12-04 | 삼성전자 주식회사 | 고화질 컬러 영상을 획득하기 위한 영상 처리 장치 및 방법 |
JP5179398B2 (ja) * | 2009-02-13 | 2013-04-10 | オリンパス株式会社 | 画像処理装置、画像処理方法、画像処理プログラム |
JP5499050B2 (ja) * | 2009-12-22 | 2014-05-21 | パナソニック株式会社 | 画像処理装置、撮像装置、及び画像処理方法 |
EP2574038B1 (en) * | 2010-05-21 | 2018-10-24 | Panasonic Intellectual Property Corporation of America | Image capturing apparatus, image processing apparatus, image processing method, and image processing program |
-
2011
- 2011-05-11 EP EP11783235.2A patent/EP2574038B1/en not_active Not-in-force
- 2011-05-11 JP JP2012515735A patent/JP5756099B2/ja not_active Expired - Fee Related
- 2011-05-11 JP JP2012515734A patent/JP5596138B2/ja not_active Expired - Fee Related
- 2011-05-11 WO PCT/JP2011/002636 patent/WO2011145296A1/ja active Application Filing
- 2011-05-11 EP EP11783236.0A patent/EP2574039B1/en not_active Not-in-force
- 2011-05-11 US US13/697,624 patent/US9071754B2/en not_active Expired - Fee Related
- 2011-05-11 CN CN201180024954.0A patent/CN102907082B/zh not_active Expired - Fee Related
- 2011-05-11 WO PCT/JP2011/002637 patent/WO2011145297A1/ja active Application Filing
- 2011-05-11 CN CN201180025062.2A patent/CN102907083B/zh not_active Expired - Fee Related
- 2011-05-11 US US13/697,686 patent/US9036032B2/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005064851A (ja) * | 2003-08-12 | 2005-03-10 | Fuji Photo Film Co Ltd | デジタルカメラ及び固体撮像装置 |
JP2008061217A (ja) * | 2006-07-31 | 2008-03-13 | Ricoh Co Ltd | 画像処理装置、撮像装置、画像処理方法、および画像処理プログラム |
JP2008099025A (ja) | 2006-10-12 | 2008-04-24 | Kyocera Corp | 撮像装置、ぶれ補正方法及び撮像方法 |
JP2009111596A (ja) | 2007-10-29 | 2009-05-21 | Sanyo Electric Co Ltd | 撮像装置 |
Non-Patent Citations (3)
Title |
---|
QI SHAN; JIAYA JIA; ASEEM AGARWALA: "High-quaiity Motion Deblurring from a Single Image", SIGGRAPH, 2008 |
ROB FERGUS ET AL.: "SIGGRAPH", 2006, article "Removing Camera Shake from a Single Image" |
See also references of EP2574039A4 |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2013192227A (ja) * | 2012-03-13 | 2013-09-26 | Samsung Electronics Co Ltd | タイル単位に基づいて大きい入力映像の不均一モーションブラーを除去する方法及び装置 |
KR101810876B1 (ko) * | 2012-03-13 | 2018-01-26 | 삼성전자주식회사 | 타일 단위를 기반으로 큰 입력 영상의 비균일 모션 블러를 제거하는 방법 및 장치 |
JP2017070590A (ja) * | 2015-10-08 | 2017-04-13 | キヤノン株式会社 | 画像処理装置及びその制御方法、コンピュータプログラム |
JP2021027491A (ja) * | 2019-08-06 | 2021-02-22 | キヤノン株式会社 | 回転半径演算装置および回転半径演算方法 |
JP7292145B2 (ja) | 2019-08-06 | 2023-06-16 | キヤノン株式会社 | 回転半径演算装置および回転半径演算方法 |
Also Published As
Publication number | Publication date |
---|---|
WO2011145296A1 (ja) | 2011-11-24 |
CN102907083B (zh) | 2016-09-28 |
US20130057714A1 (en) | 2013-03-07 |
JPWO2011145297A1 (ja) | 2013-07-22 |
JP5756099B2 (ja) | 2015-07-29 |
EP2574039B1 (en) | 2018-05-02 |
JP5596138B2 (ja) | 2014-09-24 |
EP2574039A4 (en) | 2014-04-23 |
EP2574039A1 (en) | 2013-03-27 |
US9036032B2 (en) | 2015-05-19 |
US9071754B2 (en) | 2015-06-30 |
CN102907082A (zh) | 2013-01-30 |
CN102907082B (zh) | 2016-05-18 |
EP2574038A4 (en) | 2014-04-23 |
US20130063616A1 (en) | 2013-03-14 |
JPWO2011145296A1 (ja) | 2013-07-22 |
CN102907083A (zh) | 2013-01-30 |
EP2574038B1 (en) | 2018-10-24 |
EP2574038A1 (en) | 2013-03-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5756099B2 (ja) | 撮像装置、画像処理装置、画像処理方法、および画像処理プログラム | |
JP5468404B2 (ja) | 撮像装置および撮像方法、ならびに前記撮像装置のための画像処理方法 | |
JP5237978B2 (ja) | 撮像装置および撮像方法、ならびに前記撮像装置のための画像処理方法 | |
JP5909540B2 (ja) | 画像処理表示装置 | |
JP5484631B2 (ja) | 撮像装置、撮像方法、プログラム、及びプログラム記憶媒体 | |
JP5744614B2 (ja) | 画像処理装置、画像処理方法、および、画像処理プログラム | |
JP2007135182A (ja) | 撮像装置および撮像方法 | |
JP5184574B2 (ja) | 撮像装置、画像処理装置、および画像処理方法 | |
WO2019124289A1 (ja) | 装置、制御方法および記憶媒体 | |
US9007471B2 (en) | Digital photographing apparatus, method for controlling the same, and computer-readable medium | |
JP5115297B2 (ja) | 画像処理装置、撮像装置、画像処理方法およびプログラム | |
JP6245847B2 (ja) | 画像処理装置および画像処理方法 | |
JP2011135379A (ja) | 撮像装置、撮像方法及びプログラム | |
JP6548409B2 (ja) | 画像処理装置、その制御方法、および制御プログラム、並びに撮像装置 | |
JP2009089228A (ja) | 撮像装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201180024954.0 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11783236 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2012515735 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13697686 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2011783236 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |