Systems and methods for photometric normalization in array cameras

Abstract

Systems and methods for performing photometric normalization in an array camera in accordance with embodiments of this invention are disclosed. The image data of scene from a reference imaging component and alternate imaging components is received. The image data from each of the alternate imaging components is then translated to so that pixel information in the image data of each alternate imaging component corresponds to pixel information in the image data of the reference component. The shifted image data of each alternate imaging component is compared to the image data of the reference imaging component to determine gain and offset parameters for each alternate imaging component. The gain and offset parameters of each alternate imaging component is then applied to the image data of the associate imaging to generate corrected image data for each of the alternate imaging components.

Description

FIELD OF THE INVENTION

The present invention generally relates to digital cameras and more specifically to systems and methods for evaluating imaging conditions.

BACKGROUND

The quality of an image captured by a digital camera can be influenced by factors including the exposure and focal plane settings of the camera and the dynamic range within a scene. The exposure (duration of time which light is sampled by pixels in an image sensor) impacts the color shades and tone and the focal plane settings impact a captured image's sharpness.

The dynamic range within a scene is the difference in brightness from the darkest to brightest sections of the scene. Likewise, the dynamic range of an image sensor is the difference in brightness from the darkest to brightest sections that the image sensor is able to capture. Depending on the dynamic range within a particular scene, the maximum dynamic range of an image sensor can be many times smaller than the scene's dynamic range. Thus, digital cameras may not be able to accurately capture the full range of brightness in any given scene. Various techniques including auto-exposure, autofocus and high dynamic range imaging have been developed to improve the quality of images captured using digital cameras.

In many image capture devices the sensitivity of the device to light intensity can be adjusted by manipulating pixel integration time, pixel gain, and/or iris/lens aperture. Further, metering and auto-exposure algorithms can be used to optimize the above parameters (some of these parameters may be specified or fixed). Auto-exposure algorithms utilize methods to capture images at optimal mean brightness levels by adjusting the exposure time (or focal plane settings). Such algorithms generally perform an iterative process that captures an image at a known exposure time and based on the characteristics of the captured image, sets the exposure time (or focal plane settings) to capture following images at more optimal mean brightness levels.

Most surfaces reflect incident light with some amount of scattering. Thus, the light intercepted by a camera is roughly isotropic with a small region around the vantage point of the camera. Thus, individual imaging components of an array camera should ideally provide the same numerical representation of an object in the individual images captured by each of the imaging components. However, non-idealities exist in an array camera and its individual imaging components due to manufacturing tolerances and other aberrations.

As such, the numerical representation for the same point in space as captured in the image data of each individual imaging component may differ. The differences may be subtle such as those differences caused by among other things, the differences in focal length, aperture ratios, and image sensor sensitivity in the individual imaging components. Some of these differences can be treated as constants and may be accounted for by correction factors determined through a calibration process.

However, there are some differences that are introduced by the scene being imaged that cannot be compensated for ahead of time by correction factors. One example is veiling glare. Veiling glare occurs when the image projected onto the pixels or sensors of an imaging component by a lens system includes the intended image and an erroneous internally scattered set of photons. The internally scattered set of photons may originate from anywhere in front of the imaging component including both within and outside the Field of View (FoV) of the imaging component. This causes the image projected onto the pixels or sensors of the imaging component at a given point to have more than or less than the expected photons. Additional non-idealities may also exist including, but not limited to, contaminants on a protective window over the array camera installed in a device. The contaminants may change the photo-response function for each of the individual imaging components by scattering or absorbing some of the photons entering the optical system.

It is a problem if the individual imaging components of the array camera do not report the same value for a given point in scene space in their image data. If the values for the same point in space differ in the image data of individual imaging components, the parallax detection between the different images may fail or become erroneous. Also, a noise signal may be introduced into fused images from the local differences in the numerical values of the image data from different imaging components.

SUMMARY OF THE INVENTION

The above and other problems are solved an advance in the art is made by systems and methods for providing photometric normalization for an array camera in accordance with embodiments of this invention. In accordance with embodiments of this invention, one or more of the imaging components of the array camera are designated as a reference imaging component and each of the remaining imaging components in the array camera is an alternate imaging component. Each of the alternate imaging components is associated with at least one of the reference imaging components. In accordance with embodiments of this invention, a photometric normalization process is performed after a scene has been captured by the array camera generating image data from each of the individual imaging components of the array camera.

The following process is performed for each reference imaging component and the alternate imaging components associated with each of the reference imaging components in accordance with embodiments of this invention. A nominal parallax shift is determined to translate pixel information in the imaging data of each alternate imaging component to corresponding pixel information in the imaging data of the associated reference imaging component. A low pass filter is then applied to image data of the reference camera and each of the associated translated imaging components. For each associate imaging device, the pixel information from the translated and low-pass filtered image data of the associate imaging device is compared to the corresponding pixel information of the low-pass filtered reference image to compute a gain and offset parameter transformation, which, when applied to the alternate images will photometrically match the two images, thereby reducing or eliminating the photometric imbalance among the images in the array. The computed gain and offset parameters may then applied to the image data of the associate imaging device to photometrically normalize the image data with respect to the reference imaging device.

One embodiment of the method of the invention includes: receiving image data for a scene captured by the reference imaging component; receiving image data for a scene captured by each of plurality of alternate imaging components; determining a nominal parallax for image data of each of the plurality of alternate imaging components that translate information for a particular pixel in the image data of a particular alternate imaging component to a corresponding pixel in the reference imaging component; applying the nominal parallax of each particular alternate imaging component to the image data of the particular alternate imaging component; applying a low pass filter to the image data from the reference imaging component and the shifted image data of each particular alternate imaging component; and computing gain and offset parameters for each particular alternate imaging components from the low pass filtered shifted image data of the particular alternate imaging component and the low pass filtered image data of the reference imaging component.

A further embodiment also includes applying the gain and offset parameters of each particular alternate imaging component to the image data captured by the particular alternate imaging component to form photometrically normalized image data for each particular alternate imaging component.

Another embodiment also includes determining regions of high contrast in the low pass filtered shifted image data of each particular alternate imaging component.

A still further embodiment includes storing determined the regions of high contrast in the low pass filtered image data of each particular alternate imaging component for further correction processing.

In still another embodiment, the computing of the gain and offset parameters is performed on a pixel by pixel basis for the image data of each of the plurality of alternate imaging components.

In a yet further embodiment, the computing of the gain and offset parameters is performed on regions of pixels for the image data of each of the plurality of alternate imaging components.

Yet another embodiment also includes: comparing each gain parameter and each offset parameter for each of the plurality of alternate imaging component to a threshold value; and setting each gain parameter and each offset parameter determined to at least meet the threshold value to a predetermined value.

An embodiment of a system of the invention includes: an array camera including a plurality of imaging components that capture image data of a scene including a reference imaging component and plurality of alternate imaging components; a memory; and a processor that is configured by instructions stored in the memory to: receive image data for a scene captured by the reference imaging component, receive image data for a scene captured by each of plurality of alternate imaging components, determine a nominal parallax for image data of each of the plurality of alternate imaging components that translate information for a particular pixel in the image data of a particular alternate imaging component to a corresponding pixel in the reference imaging component, apply the nominal parallax of each particular alternate imaging component to the image data of the particular alternate imaging component, apply a low pass filter to the image data from the reference imaging component and the shifted image data of each particular alternate imaging component, and compute gain and offset parameters for each particular alternate imaging components from the low pass filtered shifted image data of the particular alternate imaging component and the low pass filtered image data of the reference imaging component.

In a further embodiment, the processor is further configured by the instructions to apply the gain and offset parameters of each particular alternate imaging component to the image data captured by the particular alternate imaging component to form photometrically normalized image data for each particular alternate imaging component.

In another embodiment, the processor is further configured by the instructions to determine regions of high contrast in the low pass filtered shifted image data of each particular alternate imaging component.

In a still further embodiment, the processor is further configured by the instructions to store the determined regions of high contrast in the low pass filtered image data of each particular alternate imaging component for further correction processing.

In still another embodiment, the computing of the gain and offset parameters is performed on a pixel by pixel basis for the image data of each of the plurality of alternate imaging components.

In a yet further embodiment, the computing of the gain and offset parameters is performed on regions of pixels for the image data of each of the plurality of alternate imaging components.

In yet another embodiment, the processor is further configured by the instructions to: compare each gain parameter and each offset parameter for each of the plurality of alternate imaging component to a threshold value; and set each gain parameter and each offset parameter determined to at least meet the threshold value to a predetermined value.

Another further embodiment of the invention includes: receiving image data for a scene captured by the reference imaging component; receiving image data for a scene captured by each of plurality of alternate imaging components; determining a nominal parallax for image data of each of the plurality of alternate imaging components that translate information for a particular pixel in the image data of a particular alternate imaging component to a corresponding pixel in the reference imaging component; applying the nominal parallax of each particular alternate imaging component to the image data of the particular alternate imaging component; applying a low pass filter to the image data from the reference imaging component and the shifted image data of each particular alternate imaging component; and computing gain and offset parameters for each particular alternate imaging components from the low pass filtered shifted image data of the particular alternate imaging component and the low pass filtered image data of the reference imaging component.

In still another further embodiment, the method further comprises applying the gain and offset parameters of each particular alternate imaging component to the image data captured by the particular alternate imaging component to form photometrically normalized image data for each particular alternate imaging component.

In yet another further embodiment, the method further comprises determining regions of high contrast in the low pass filtered shifted image data of each particular alternate imaging component.

In another further embodiment again, the method further comprises storing determined the regions of high contrast in the low pass filtered image data of each particular alternate imaging component for further correction processing.

In another further additional embodiment, the computing of the gain and offset parameters is performed on a pixel by pixel basis for the image data of each of the plurality of alternate imaging components.

In still yet another further embodiment, the computing of the gain and offset parameters is performed on regions of pixels for the image data of each of the plurality of alternate imaging components.

In still another further embodiment again, the method further comprises: comparing each gain parameter and each offset parameter for each of the plurality of alternate imaging components to a threshold value; and setting each gain parameter and each offset parameter determined to at least meet the threshold value to a predetermined value.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of an array camera in accordance with an embodiment of the invention.

FIG. 2 conceptually illustrates an optic array and an imager array in an array camera module in accordance with an embodiment of the invention.

FIG. 3 is an architecture diagram of an imager array in accordance with an embodiment of the invention.

FIG. 4 is a high level circuit diagram of pixel control and readout circuitry for a plurality of focal planes in an imager array in accordance with an embodiment of the invention.

FIG. 5 conceptually illustrates a layout of color filters and the location of a reference imaging component and an alternate imaging component in an array camera module in accordance with an embodiment of the invention.

FIG. 6 is a flow chart illustrating a process for performing photometric normalization for an array camera in accordance with embodiments of the invention.

DETAILED DISCLOSURE OF THE INVENTION

Turning now to the drawings, systems and methods for measuring scene information while capturing images using array cameras in accordance with embodiments of the invention are illustrated. Array cameras including camera modules that can be utilized to capture image data from different viewpoints (i.e. light field images) are disclosed in U.S. patent application Ser. No. 12/935,504 entitled “Capturing and Processing of Images using Monolithic Camera Array with Heterogeneous Imagers” to Venkataraman et al. In many instances, fusion and super-resolution processes such as those described in U.S. patent application Ser. No. 12/967,807 entitled “Systems and Methods for Synthesizing High Resolution Images Using Super-Resolution Processes” to Lelescu et al., can be utilized to synthesize a higher resolution 2D image or a stereo pair of higher resolution 2D images from the lower resolution images in the light field captured by an array camera. The terms high or higher resolution and low or lower resolution are used here in a relative sense and not to indicate the specific resolutions of the images captured by the array camera. The disclosures of U.S. patent application Ser. No. 12/935,504 and U.S. patent application Ser. No. 12/967,807 are hereby incorporated by reference in their entirety.

In accordance with embodiments of this invention, a photometric normalization is performed on image data captured by an array camera. The photometric normalization is performed to determine local offset and gain coefficients for the image data from alternate imaging components with respect to a reference imaging component. The gain and offset coefficients correct the image data of the alternate imaging component to account for differences introduced by the scene being imaged. In particular, the gain coefficient corrects for the resultant attenuation of photons caused by veiling glare and other scene related issues and the offset coefficient corrects for the resultant spurious or additional photons introduced by veiling glare or other scene independent issues. Systems and methods for performing photometric normalization of image data captured by an array camera in accordance with embodiments of the invention are discussed further below.

Array Cameras

Array cameras in accordance with embodiments of the invention can include a camera module and a processor. An array camera in accordance with an embodiment of the invention is illustrated in FIG. 1. The array camera 100 includes a camera module 102 with an array of individual imaging components 104 where an array of individual imaging components refers to a plurality of imaging components in a particular arrangement, such as (but not limited to) the square arrangement utilized in the illustrated embodiment. The camera module 102 is connected to the processor 106 and the processor 106 is connected to a memory 108. Although a specific array camera is illustrated in FIG. 1, any of a variety of different array camera configurations can be utilized in accordance with many different embodiments of the invention.

Array Camera Modules

Camera modules in accordance with embodiments of the invention can be constructed from an imager array and an optic array. A camera module in accordance with an embodiment of the invention is illustrated in FIG. 2. The camera module 200 includes an imager array 230 including an array of focal planes 240 along with a corresponding optic array 210 including an array of lens stacks 220. Within the array of lens stacks, each lens stack 220 creates an optical channel that forms an image of the scene on an array of light sensitive pixels within a corresponding focal plane 240. Each pairing of a lens stack 220 and focal plane 240 forms a single camera 104 within the camera module. Each pixel within a focal plane 240 of a camera 104 generates image data that can be sent from the camera 104 to the processor 108. In many embodiments, the lens stack within each optical channel is configured so that pixels of each focal plane 240 sample the same object space or region within the scene. In several embodiments, the lens stacks are configured so that the pixels that sample the same object space do so with sub-pixel offsets to provide sampling diversity that can be utilized to recover increased resolution through the use of super-resolution processes.

In several embodiments, color filters in individual imaging components can be used to pattern the camera module with it filter groups as further discussed in U.S. Provisional Patent Application No. 61/641,165 entitled “Camera Modules Patterned with pi Filter Groups” filed May 1, 2012, the disclosure of which is incorporated by reference herein in its entirety. The use of a color filter pattern incorporating it filter groups in a 4×4 array is illustrated in FIG. 5. These imaging components can be used to capture data with respect to different colors, or a specific portion of the spectrum. In contrast to applying color filters to the pixels of the individual imaging components, color filters in many embodiments of the invention are included in the lens stack. For example, a green color imaging component can include a lens stack with a green light filter that allows green light to pass through the optical channel. In many embodiments, the pixels in each focal plane are the same and the light information captured by the pixels is differentiated by the color filters in the corresponding lens stack for each filter plane. Although a specific construction of a camera module with an optic array including color filters in the lens stacks is described above, camera modules including it filter groups can be implemented in a variety of ways including (but not limited to) by applying color filters to the pixels of the focal planes of the camera module similar to the manner in which color filters are applied to the pixels of a conventional color camera. In several embodiments, at least one of the imaging components in the camera module can include uniform color filters applied to the pixels in its focal plane. In many embodiments, a Bayer filter pattern is applied to the pixels of one of the imaging components in a camera module. In a number of embodiments, camera modules are constructed in which color filters are utilized in both the lens stacks and on the pixels of the imager array.

In several embodiments, an array camera generates image data from multiple focal planes and uses a processor to synthesize one or more images of a scene. In certain embodiments, the image data captured by a single focal plane in the sensor array can constitute a low resolution image (the term low resolution here is used only to contrast with higher resolution images), which the processor can use in combination with other low resolution image data captured by the camera module to construct a higher resolution image through Super Resolution processing.

Although specific array cameras are discussed above, many different array cameras are capable of utilizing π filter groups in accordance with embodiments of the invention. Imager arrays in accordance with embodiments of the invention are discussed further below.

Imager Arrays

An imager array in which the image capture settings of a plurality of focal planes or imaging components can be independently configured in accordance with an embodiment of the invention is illustrated in FIG. 3. The imager array 300 includes a focal plane array core 302 that includes an array of focal planes 304 and all analog signal processing, pixel level control logic, signaling, and analog-to-digital conversion (ADC) circuitry. The imager array also includes focal plane timing and control circuitry 306 that is responsible for controlling the capture of image information using the pixels. In a number of embodiments, the focal plane timing and control circuitry utilizes reset and read-out signals to control the integration time of the pixels. In other embodiments, any of a variety of techniques can be utilized to control integration time of pixels and/or to capture image information using pixels. In many embodiments, the focal plane timing and control circuitry 306 provides flexibility of image information capture control, which enables features including (but not limited to) high dynamic range imaging, high speed video, and electronic image stabilization. In various embodiments, the imager array includes power management and bias generation circuitry 308. The power management and bias generation circuitry 308 provides current and voltage references to analog circuitry such as the reference voltages against which an ADC would measure the signal to be converted against. In many embodiments, the power management and bias circuitry also includes logic that turns off the current/voltage references to certain circuits when they are not in use for power saving reasons. In several embodiments, the imager array includes dark current and fixed pattern (FPN) correction circuitry 310 that increases the consistency of the black level of the image data captured by the imager array and can reduce the appearance of row temporal noise and column fixed pattern noise. In several embodiments, each focal plane includes reference pixels for the purpose of calibrating the dark current and FPN of the focal plane and the control circuitry can keep the reference pixels active when the rest of the pixels of the focal plane are powered down in order to increase the speed with which the imager array can be powered up by reducing the need for calibration of dark current and FPN.

In many embodiments, a single self-contained chip imager includes focal plane framing circuitry 312 that packages the data captured from the focal planes into a container file and can prepare the captured image data for transmission. In several embodiments, the focal plane framing circuitry includes information identifying the focal plane and/or group of pixels from which the captured image data originated. In a number of embodiments, the imager array also includes an interface for transmission of captured image data to external devices. In the illustrated embodiment, the interface is a MIPI CSI 2 output interface (as specified by the non-profit MIPI Alliance, Inc.) supporting four lanes that can support read-out of video at 30 fps from the imager array and incorporating data output interface circuitry 318, interface control circuitry 316 and interface input circuitry 314. Typically, the bandwidth of each lane is optimized for the total number of pixels in the imager array and the desired frame rate. The use of various interfaces including the MIPI CSI 2 interface to transmit image data captured by an array of imagers within an imager array to an external device in accordance with embodiments of the invention is described in U.S. Pat. No. 8,305,456, entitled “Systems and Methods for Transmitting Array Camera Data”, issued Nov. 6, 2012, the disclosure of which is incorporated by reference herein in its entirety.

Although specific components of an imager array architecture are discussed above with respect to FIG. 3, any of a variety of imager arrays can be constructed in accordance with embodiments of the invention that enable the capture of images of a scene at a plurality of focal planes in accordance with embodiments of the invention. Independent focal plane control that can be included in imager arrays in accordance with embodiments of the invention are discussed further below.

Independent Focal Plane Control

Imager arrays in accordance with embodiments of the invention can include an array of focal planes or imaging components that can independently be controlled. In this way, the image capture settings for each focal plane in an imager array can be configured differently. As is discussed further below, the ability to configure active focal planes using difference image capture settings can enable different cameras within an array camera to make independent measurements of scene information that can be combined for use in determining image capture settings for use more generally within the camera array.

An imager array including independent control of image capture settings and independent control of pixel readout in an array of focal planes in accordance with an embodiment of the invention is illustrated in FIG. 4. The imager array 400 includes a plurality of focal planes or pixel sub-arrays 402. Control circuitry 403, 404 provides independent control of the exposure timing and amplification gain applied to the individual pixels within each focal plane. Each focal plane 402 includes independent row timing circuitry 406, 408, and independent column readout circuitry 410, 412. In operation, the control circuitry 403, 404 determines the image capture settings of the pixels in each of the active focal planes 402. The row timing circuitry 406, 408 and the column readout circuitry 410, 412 are responsible for reading out image data from each of the pixels in the active focal planes. The image data read from the focal planes is then formatted for output using an output and control interface 416.

Although specific imager array configurations are discussed above with reference to FIG. 4, any of a variety of imager array configurations including independent and/or related focal plane control can be utilized in accordance with embodiments of the invention including those outlined in U.S. patent application Ser. No. 13/106,797, entitled “Architectures for Imager Arrays and Array Cameras”, filed May 12, 2011, the disclosure of which is incorporated by reference herein in its entirety. The use of independent focal plane control to capture image data using array cameras is discussed further below.

Photometric Normalization for an Array Camera

In accordance with many embodiments of this invention, a photometric normalization is performed on image data captured by an array camera. The photometric normalization is performed to determine local offset and gain coefficients for the image data from alternate imaging components with respect to a reference imaging component. The gain and offset coefficients correct the image data of the alternate imaging component to account for differences introduced by the scene being imaged. In particular, the gain coefficient corrects for the resultant attenuation of photons caused by veiling glare and other scene related issues and the offset coefficient corrects for the resultant spurious or additional photons introduced by veiling glare or other scene independent issues.

The normalization performed is based on the fact the one of the properties of the veiling glare phenomenon and other scene related errors is that its effect on the photo-response of each of the individual imaging components is typically low in spatial frequency. Thus, the photo-response of the imaging component does not change rapidly within an image area. Instead, the photo-response is relatively slow changing. As the scene related errors may cause the image projected on the imaging components to include either more or less photons than predicted by a flat-field calibration, some areas of the image of the individual image sensor may be brighter or darker versus the image from other imaging components.

These scene related errors may be corrected for or normalized out by computing the above described gain and offset coefficients. These coefficients can be determined because the response of an imaging component in the raw domain is designed to be linear. Thus, the typical y=mx+c formula may be used to define the response in the following manner:y_i,j=m_i,jx_i,j+c

Where:

• • y_i,j=numerical output value of the sensor for a given position in the image;
  • x_i,j=photon input to the sensor at a given position;
  • m_i,j=conversion gain of the sensor at a given position as determined by calibration; and
  • c=pedestal black level of the sensor.

Scene related errors can be thought of as resulting in the following modification to the formula:y_i,j=m_i,jx_i,j(Gvg_i,j)+C+(Ovg_i,j)

Where:

• • Gvg is the gain coefficient representing the resultant attenuation of photons; and
  • Ovg is the offset coefficient representing the resultant spurious or additional photons.

To normalize the image data from alternate imaging components with respect to the image data of a reference imaging component, the gain and offset coefficients for the alternate imaging components with respect to the reference imaging component can be computed and applied to the imaging data of the alternate imaging component to negate their effects. The use of a color filter pattern incorporating it filter groups in a 4×4 array is illustrated in FIG. 5. In the array camera 500, a first imaging component 504 configured to capture green light can be selected as a reference imaging component and a second imaging component 506 configured to capture green light can be selected as an alternate imaging component. As can readily be appreciated, any pair of cameras configured to capture the same frequency of light can be selected as a reference imaging component and an associate imaging component. A process for performing this photometric normalization in accordance with embodiments of this invention is illustrated in FIG. 6.

Process 600 includes obtaining the image data for a scene from a reference imaging component and the alternate imaging components associated with the reference imaging component (605). This may be done by capturing an image of the scene with an array camera causing the reference and alternate imaging components to each generate image data of the scene. Alternatively, the image data may have been previously captured and is read from a memory.

If the array camera includes more than one reference imaging component, a reference imaging component is selected to perform the normalization (605). A low pass filter is then applied to the image data of the reference imaging component (610). The low pass filter removes any high frequency components in the reference image data.

The following process is then performed to normalize the image data from each of the alternate imaging components associated with reference imaging component. An alternate imaging component is selected (615) and the image data for the alternate imaging component is retrieved. A nominal parallax between the selected alternate imaging component and the reference imaging component is determined (620). The nominal parallax may be read from memory if it was previously stored or may be computed at the time of use.

In accordance with some embodiments of this invention, the nominal parallax may be determined by metering a region-of-interest within the field of view and performing a coarse parallax estimate to determine a parallax shift that satisfies the metered region-of-interest. In accordance with some other embodiments, a nominal parallax shift corresponding to typical shooting distances may be used. In accordance with still other embodiments, the depth map from a previously fully computed frame may be used to specify the nominal parallax shift.

The nominal parallax shift is then applied to the image data of the alternate camera to translate the pixel information in the image data to correspond with corresponding pixel information in the image data of the reference imaging component (625). A low pass filter is then applied to the shifted image data of the alternate imaging component (630). The shifted, low passed filtered image data of the alternate image component is aligned with the low passed filtered image data from the reference imaging component in a “strong” blurred alignment. A “strong” blurred alignment is when the images are aligned on surviving high-gradient edges in low frequency such that the images appear be aligned even if there some spatial misalignment due to error in the alignment information. Errors in alignment information may be due to many factors, including, but not limited tom, taking the parallax at an incorrect fixed distance.

In accordance with some embodiments, high contrast components in the image data from the alternate imaging component may be detected. The high contrast components in the shifted image data are typically in areas where alignment errors caused by using the nominal parallax shift between the shifted image data from the alternate imaging component and the image data from the reference imaging component are apparent. These high-contrast edges may still cause differences even after the low pass filter is applied. Thus, these high contrast components are optionally detected and stored as a data set, map, or other data structure (635). As these components have a greater probability of being erroneous even after the subsequent correction values are applied, the data set or map may be used to indicate components of the shifted image data from the alternate imaging component where later correction processes can be applied modulate the corrected data if needed and/or desired.

The low pass filtered shifted image data of the alternate imaging component is then compared to the low pass filtered image data of the reference imaging component to compute the gain and offset parameters for the image data from the alternate imaging component (640). The low passed filtered shifted image data is used to determine the gain and offset parameter because most photometric imbalances occur in low frequency. Thus, the gain and offset parameters to locally correct the photometric imbalance determined using the low pass filtered image data will correct photometric imbalance in the original image data as the photometric imbalance if in the lower frequency.

In accordance with some embodiments, the gain and offset parameters are calculated on a per pixel basis. In accordance with these embodiments, the gain and offset parameters are calculated based on a region surrounding each pixel. For example, a region of 9×9 pixels surrounding a pixel may be analyzed to determine the distribution of values within the region. In other embodiments, any of a variety of fixed or adaptive regions can be utilized including regions that have different shapes in different regions of the image. A level of contrast exists in the region within the image data of the alternate imaging component. The goal of the computation is to determine gain and offset parameters for the pixel in shifted image data that matches the value of the pixel data to the value of the pixel data of the reference imaging component. This may be achieved by comparing the mean and variance of the data for the pixel area to the mean and variance of the data for a corresponding pixel area in the reference image data.

In accordance with some embodiments of this invention, the following equation may be used to perform the comparisons and determine the gain and offset parameters:

$\hat{a}=\frac{\left[\sum _i\sum _{j}y\left(i,j\right)x\left(i,j\right)\right]-N_{1N_2\stackrel{\_}{y}\stackrel{\_}{x}}}{\left[\sum _i\sum _j{x}^2\left(i,j\right)\right]-N_1{N}_2{\stackrel{\_}{x}}^2},\hat{b}=\stackrel{\_}{y}-\hat{a}\stackrel{\_}{x},\mathrm{where}$ $\stackrel{\_}{x}=\frac{1}{N_1{N}_2}\sum _i\sum _{j}x\left(i,j\right),\stackrel{\_}{y}=\frac{1}{N_1{N}_2}\sum _i\sum _{j}y\left(i,j\right),$

Where:

• • x=the image to be corrected
  • y=the reference image
  • N₁, N₂=number of pixels horizontally and vertically of the analyzed region around the pixel being computed.
  • i,j are indices into the images within the bounds defined by N₁ and N₂.
  • â=Gvg_i,j, gain coefficient computed for a specific value of i and j.
  • {circumflex over (b)}=Ovg_i,j, offset term computed for a specific value of i and j.

In accordance with some embodiments, limits may be applied to the computation such that values computed for the gain and offset parameters are constrained in some way. In accordance with some of these embodiments, the gain and offset parameters may be prevented from being too large by being compared to a threshold and being set to a predetermined value if the threshold is at least met.

In accordance with other embodiments, the gain and offset parameters may be determined for regions of the associate image data instead of a per pixel basis by using a sparse grid. The subsequent spatially varying values of the gain and offset parameters may be interpolated to yield the correct value for each pixel. One skilled in the art will recognize that still other methods of determining the gain and offset parameters may be used without departing from the embodiments of this invention.

The determined gain and offset parameters for each pixel are then applied to the corresponding information for each pixel in the original image data of the associate reference component (645). In accordance with some embodiments, the map or data set of high contrast regions may be used to determine regions where the calculations may be erroneous and additional processes may need to be performed to normalize the data.

The process (615-645) for alternate imaging components associated the selected reference is then repeated until normalization is performed for each alternate imaging component associated with the selected reference imaging component (650). The process is likewise repeated for each reference imaging component in the array camera (655).

While the above description contains many specific embodiments of the invention, these should not be construed as limitations on the scope of the invention, but rather as an example of one embodiment thereof. It is therefore to be understood that the present invention may be practiced otherwise than specifically described, without departing from the scope and spirit of the present invention. Thus, embodiments of the present invention should be considered in all respects as illustrative and not restrictive.

Claims

1. A method of manufacturing an array camera device, the method comprising: assembling an array of cameras comprising a plurality of imaging components that capture images of a scene from different viewpoints, where the components include:a set of one or more reference imaging components, each having a reference viewpoint, anda set of one or more associate imaging components;configuring the array of cameras to communicate with at least one processor;configuring the processor to communicate with at least one display;configuring the processor to communicate with at least one type of memory; andperforming a photometric normalization process for the array of cameras, where the photometric normalization process comprises:receiving image data for a scene captured by a first one of the plurality of imaging components;receiving image data for a scene captured by a second one of the plurality of imaging components;determining a nominal parallax for image data of the second one of the plurality the of imaging components that translate information for a particular pixel in the image data of the second one of the plurality of imaging components to a corresponding pixel in the first one of the plurality of imaging components;applying the nominal parallax of the second one of the plurality of imaging components to the image data of the second one of the plurality of imaging components;applying a low pass filter to the image data from the first one of the plurality of imaging components and the shifted image data of the second one of the plurality of imaging components; andcomputing gain and offset parameters for the second one of the plurality of imaging components from the low pass filtered shifted image data of the second one of the plurality of imaging components and the low pass filtered image data of the first one of the plurality of imaging components.

2. The method of claim 1 further comprising applying the gain and offset parameters of the second one of the plurality of imaging components to the image data captured by the second one of the plurality of imaging components to form photometrically normalized image data for the second one of the plurality of imaging components.

3. The method of claim 2 further comprising determining regions of high contrast in the low pass filtered shifted image data of the second one of the plurality of imaging components.

4. The method of claim 3 further comprising storing the determined regions of high contrast in the low pass filtered image data of the second one of the plurality of imaging components for further correction processing.

5. The method of claim 1 wherein the computing of the gain and offset parameters is performed on a pixel by pixel basis for the image data of the second one of the plurality of imaging components.

6. The method of claim 1 wherein the computing of the gain and offset parameters is performed on regions of pixels for the image data of the second one of the plurality of imaging components.

7. The method of claim 1 further comprising: comparing each gain parameter and each offset parameter for the second one of the plurality of imaging components to a threshold value; and setting each gain parameter and each offset parameter determined to at least meet the threshold value to a predetermined value.

8. A method of manufacturing an array camera device, the method comprising: assembling an array of cameras comprising a plurality of imaging components that capture images of a scene from different viewpoints, where the components include:a set of one or more reference imaging components, each having a reference viewpoint, anda set of one or more associate imaging components;configuring the array of cameras to communicate with at least one processor;configuring the processor to communicate with at least one display;configuring the processor to communicate with at least one type of memory; andperforming a photometric normalization process for the array of cameras, where the photometric normalization process comprises:receiving image data for a scene captured by the reference imaging component;receiving image data for a scene captured by each of plurality of alternate imaging components;determining a nominal parallax for image data of each of the plurality of alternate imaging components that translate information for a particular pixel in the image data of a particular alternate imaging component to a corresponding pixel in the reference imaging component;applying the nominal parallax of each particular alternate imaging component to the image data of the particular alternate imaging component;applying a low pass filter to the image data from the reference imaging component and the shifted image data of each particular alternate imaging component; andcomputing gain and offset parameters for each particular alternate imaging components from the low pass filtered shifted image data of the particular alternate imaging component and the low pass filtered image data of the reference imaging component.

9. The method of claim 8 further comprising applying the gain and offset parameters of each particular alternate imaging component to the image data captured by the particular alternate imaging component to form photometrically normalized image data for each particular alternate imaging component.

10. The method of claim 9 further comprising determining regions of high contrast in the low pass filtered shifted image data of each particular alternate imaging component.

11. The method of claim 10 further comprising storing determined the regions of high contrast in the low pass filtered image data of each particular alternate imaging component for further correction processing.

12. The method of claim 8, wherein the computing of the gain and offset parameters is performed on a pixel by pixel basis for the image data of each of the plurality of alternate imaging components.

13. The method of claim 8, wherein the computing of the gain and offset parameters is performed on regions of pixels for the image data of each of the plurality of alternate imaging components.

14. The method of claim 8 further comprising: comparing each gain parameter and each offset parameter for each of the plurality of alternate imaging component to a threshold value; and setting each gain parameter and each offset parameter determined to at least meet the threshold value to a predetermined value.