One chip, low light level color camera

The present invention relates to low light level imaging devices. In particular, the present invention relates to electronic low light level imaging devices that are sensitive to infrared wavelengths and that are capable of providing color information.

BACKGROUND OF THE INVENTION

Low light level imaging devices are useful in a wide variety of applications. For example, low light level imaging devices are useful in nighttime surveillance activities. In low light level conditions, it is important for sensors to take advantage of every available photon. Approaches to providing human perceptible images in low light conditions have included the use of image intensifiers. Other approaches to providing images in low light conditions have utilized the detection of light having wavelengths outside of the visible spectrum. Although such approaches have had success in providing human perceptible images of scenes that could not otherwise be viewed by the human eye, such approaches have been unable to provide chrominance information in combination with high sensitivity at low light levels.

Image intensifiers are generally formed using microchannel plates. In particular, a photocathode is positioned near a honeycomb of small channels (or microchannels) upon which a voltage gradient is imposed. When a photon collides with the photocathode, an electron is released and is accelerated along a channel. The electron is focused on a phosphorous screen, which produces photons in response to being bombarded by the electrons ejected by the microchannel. The resulting image on the phosphorous screen may be viewed directly, or may be converted into digital information by a imaging device, such as a charge coupled device (CCD).

The amplification of light provided by an image intensifier is effective in providing views of scenes at low light levels. However, the use of a phosphorous screen results in a monochromatic image. In addition, the limited resolution of the microchannel element in turn limits the image resolution available at the phosphorous screen. Also, a "halo" effect can occur when electrons bounce off the mouth of a channel and hit a neighbor channel. Furthermore, image intensifiers require a relatively high voltage for operation, and have a finite life span.

Another approach to providing high sensitivity imaging devices in low light conditions is to utilize image sensors that are capable of detecting light falling outside of the normal range of human vision. For example, typical nighttime scenes are relatively rich in infrared light wavelengths. Therefore, by detecting infrared wavelengths and providing the detected infrared information as luminance (or brightness) information to a human perceptible display, high sensitivity may be obtained. However, systems utilizing imagers that are sensitive to infrared wavelengths do not provide information regarding the colors of objects present in the imaged scene.

As a further alternative, imaging devices or cameras utilizing three image sensors or chips for detecting color information, and a fourth chip for detecting luminance information have been proposed. However, multichip designs are difficult to manufacture and implement. In addition, the ability of such designs to provide high sensitivity is compromised, by splitting the light gathered by the device's lens system among four different imagers. Furthermore, the use of four separate image sensors results in an overall package that is relatively large and expensive to produce.

In consumer and military applications, it is desirable to provide imaging devices that are relatively small and light, and that use relatively little power. Accordingly, most consumer imaging devices and many imaging devices designed for military applications utilize a single image sensor. As a result, existing imaging devices that provide color information are relatively insensitive in low light conditions, while imaging devices optimized for high sensitivity and low light conditions typically provide a monochromatic image.

The visual tasks of detection and recognition can be greatly aided if color information is provided to the viewer. Imaging devices capable of providing color information typically do so by separately sampling light having bands of color centered on the red, green, and blue portions of the spectrum. However, because filtering light requires the rejection of at least some components of the light incident on the filter, filtering reduces the sensitivity that might otherwise be available from an imaging device. One approach to increasing the light sensitivity of an image sensor used in connection with a color imaging device is described by Bayer in <patcit id="pcit0001" dnum="US3971065A"><text>U.S. Patent No. 3,971,065</text></patcit>. The filter, known as a Bayer filter, disclosed therein establishes pixel subsets distributed across the entire array of pixels in an image sensor. Each subset of pixels consists of one pixel having a filter that admits red light, one pixel having a filter that admits blue light, and two pixels having filters that admit green light. The Bayer filter favors green filter elements because green is the main contributor to the luminance information in a scene. This preference for luminance (or brightness) information over chrominance information provides an image sensor with greater sensitivity and resolution. However, because at least some filtering is performed before photons reach the pixels of the image sensor, the sensitivity of devices employing Bayer filtration could be improved.

It would be desirable to provide an electronic color imaging device that is capable of providing high sensitivity. In addition, it would be advantageous to provide such a device that utilized a single image sensor. Furthermore, it would be desirable to provide such a device that was relatively inexpensive to implement and easy to manufacture.

<patcit id="pcit0002" dnum="WO8601678A"><text>WO 86/01678 A, published 27 March 1986</text></patcit>, discusses a single chip color image sensor having an array of color filters with luminance sensing elements interspersed with chrominance sensing elements R and B in which the luminance sensing elements comprise 75 percent of the sensing elements of the chip and each chrominance sampling element R or B is totally surrounded by luminance sampling elements. The sensor is associated with signal processing means for providing interpolated values between chrominance sample values by interpolating hue values.

<patcit id="pcit0003" dnum="JP63062492A"><text>JP 63 062492 A, published 18 March 1988</text></patcit>, discusses a solid state image pickup element for color picture with a micro filter arranged on an optical receiving part with luminance exclusive lines and color exclusive lines. By this arrangement of luminance exclusive lines and color exclusive I ines, the occurrence of a moire is suppressed.

<patcit id="pcit0004" dnum="WO0057651A"><text>WO 00/57651 A, published 28 September 2000</text></patcit>, discusses a color image pickup apparatus in which a color filter pattern for a semiconductor imaging sensor is described. The imaging sensor includes light sensitive elements that are each sensitive to photon energy in a spectral region or color band associated with the light sensitive elements. Select light sensitive elements in the array are sensitive to energy in a wide band spectral region or "white" color band.

<patcit id="pcit0005" dnum="US4437112A"><text>US4437112</text></patcit> discloses an imaging device composed of a solid-state charge-coupled device having a cells each of which has at least two potential wells and color filter having a combination of color filter elements positioned corresponding to the wells so as to provide different spectral sensitivities to the two potential wells.

<patcit id="pcit0006" dnum="US4404586A"><text>US 4404586</text></patcit> discloses a solid-state color imager comprised of a solid-state base comprised of a plurality of charge switching elements and photodiode elements associated with some of the switching elements arranged in sets having superimposed thereon a photosensor layer comprised of photoconductor segments which can detect and absorb light.

SUMMARY OF THE INVENTION

In accordance with the present invention, a one chip, low light level imaging device, including a color camera, is provided. More than half of the pixels comprising the device's image sensor receive unfiltered light. Furthermore, less than half of the pixels of the image sensor receive filtered light, to provide color information to the user. The pixels of the image sensor that receive unfiltered image information are sensitive to light having wavelengths extending across the visible spectrum and into the near infrared and infrared wavelengths.

In accordance with an embodiment of the present invention, the number of pixels receiving filtered light is small, as compared to the number of pixels receiving unfiltered light. For example, in accordance with an embodiment of the present invention, less than about ten percent of the pixels of the image sensor are provided with filtered light. In accordance with another embodiment of the present invention, less than about four percent of the pixels of the image sensor are provided with filtered light.

In accordance with another embodiment of the present invention, color information obtained by pixels receiving filtered light is assigned to pixels receiving unfiltered light through interpolation. In accordance with still another embodiment of the present invention, color information is associated with pixels receiving unfiltered light by convoluting the signals obtained by pixels receiving filtered light with a suitable kernel.

In accordance with still another embodiment of the present invention, the pixels receiving filtered light are distributed among the pixels of the image sensor. For example, the image sensor may be considered as a plurality of pixel subsets. Each pixel subset comprises a group of pixels having n rows and m columns. Within each subset, a group of pixels is provided with filtered light. For example, a first pixel may be provided with red light, a second with green light, and a third with blue light. As a further example, a first and a second of the pixels may be provided with green light, a third with red light, and a fourth with blue light. The color value sensed by the pixels receiving filtered light may then be assigned to all of the pixels included in the subset of pixels. In accordance with an embodiment of the present invention, the pixels receiving filtered light are positioned at or near the center of the subset of pixels. In accordance with a further embodiment of the present invention, the subsets of pixels are arranged about the image sensor in a staggered configuration, to reduce aliasing of color features in an observed scene. In another embodiment of the present invention, pixels receiving filtered information are distributed randomly or pseudo-randomly about the image sensor.

In accordance with an embodiment of the present invention, the light provided to a small proportion of pixels included as part of an image sensor is filtered to obtain a color signal. The color signal thus obtained is assigned to areas of the image corresponding to pixels that receive unfiltered image information. Additional advantages of the present invention will become readily apparent from the following description, particularly when taken together with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

<ul id="ul0001" list-style="none" compact="compact"><li><figref idref="f0001"><b>Fig. 1</b></figref> illustrates an electronic imaging device configuration;</li><li><figref idref="f0001"><b>Fig. 2</b></figref> is a block diagram depicting major components of an electronic imaging device;</li><li><figref idref="f0002"><b>Fig. 3</b></figref> depicts the imaging surface of an image sensor;</li><li><figref idref="f0002"><b>Fig. 4</b></figref> depicts a pixel of an image sensor;</li><li><figref idref="f0002"><b>Fig. 5</b></figref> depicts a filter in accordance with the prior art;</li><li><figref idref="f0002"><b>Fig. 6</b></figref> depicts a filter element in accordance with the prior art;</li><li><figref idref="f0003"><b>Fig. 7A</b></figref> depicts an image sensor in accordance with an embodiment of the present invention;</li><li><figref idref="f0003"><b>Fig. 7B</b></figref> depicts a subset of pixels in accordance with an embodiment of the present invention;</li><li><figref idref="f0003"><b>Fig. 7C</b></figref> depicts a subset of pixels in accordance with another embodiment of the present invention;</li><li><figref idref="f0003"><b>Fig. 7D</b></figref> depicts a subset of pixels in accordance with another embodiment of the present invention;</li><li><figref idref="f0004"><b>Fig. 8</b></figref> depicts an image sensor in accordance with another embodiment of the present invention;</li><li><figref idref="f0005"><b>Fig. 9</b></figref> depicts an image sensor in accordance with another embodiment of the present invention;</li><li><figref idref="f0005"><b>Fig. 10</b></figref> depicts an image sensor in accordance with another embodiment of the present invention;</li><li><figref idref="f0006"><b>Fig.11</b></figref> depicts the determination of color information in accordance with an embodiment of the present invention;</li><li><figref idref="f0007"><b>Fig. 12</b></figref> is a chromaticity diagram of an ideal color imaging device;</li><li><figref idref="f0008"><b>Fig. 13</b></figref> is a chromaticity diagram for a color imaging device in accordance with an embodiment of the present invention; and</li><li><figref idref="f0009"><b>Fig. 14</b></figref> depicts aspects of an interlaced imaging device sensor in accordance with an embodiment of the present invention.</li></ul>

DETAILED DESCRIPTION

According to the present invention, a one chip, low light level color imaging device or camera is provided.

In <figref idref="f0001"><b>Fig.1</b></figref><b>,</b> a digital imaging device <b>100</b> configuration suitable for use in connection with the present invention is illustrated. As shown in <figref idref="f0001"><b>Fig. 1</b></figref><b>,</b> an object <b>104</b> reflects light <b>108</b> that is focused by a lens or lens system <b>112</b> onto an image sensor <b>116</b> as an image object <b>120.</b> The image sensor <b>116</b> is comprised of a plurality of photosensitive elements or pixels <b>124.</b> The pixels <b>124</b> are distributed over the surface of the image sensor <b>116</b> in a two-dimensional array. The pixels <b>124</b> generally convert the light focused onto the surface of the image sensor <b>116</b> by the lens <b>112</b> (<i>i.e.</i> the image object <b>120</b>) into electronic signals.

In a conventional color imaging device, each of the pixels <b>124</b> of the image sensor <b>116</b> receive light within one of three overlapping frequency bands. The relative intensity of the light received by individual pixels <b>124</b> included in a group of pixels in which each of the different frequency bands are represented enables the image sensor <b>116</b> to provide color information. Furthermore, the image information received by pixels <b>124</b> that are particularly sensitive to luminance information (<i>i.e</i>., pixels receiving green light) allows the image sensor <b>116</b> to provide a signal corresponding to details of the image object <b>120.</b> As will be explained in detail below, an imaging device <b>100</b> in accordance with the present invention allows a larger proportion of the pixels <b>124</b> to contributed luminance information than a conventional device, increasing the sensitivity of the device, while continuing to provide color information.

With reference now to <figref idref="f0001"><b>Fig. 2</b></figref><b>,</b> a block diagram depicting the major functional components of an imaging device <b>100</b> are illustrated. As noted above in connection with <figref idref="f0001"><b>Fig. 1</b></figref><b>,</b> a digital imaging device or camera <b>100</b> typically includes a system for focusing light reflected by images onto the image sensor <b>116</b> in the form of a lens or lens system <b>112.</b> In addition to the lens or lens system <b>112</b> and the image sensor <b>116</b> as shown in <figref idref="f0001"><b>Fig. 2</b></figref><b>,</b> a digital image device <b>100,</b> including a digital imaging device or camera <b>100,</b> in accordance with the present invention, may include a controller <b>204,</b> a preamplifier <b>208,</b> and a storage and/or display device <b>212.</b> In general, the controller <b>204</b> controls the operation of the electronic components (<i>e.g</i>., the image sensor <b>116,</b> preamplifier <b>208</b> and storage and/or display <b>212</b>) included in the device <b>100.</b> With respect to the image sensor <b>116,</b> the controller <b>204</b> controls the frame rate at which image information is converted into electronic form by the image sensor <b>116.</b> The electronic information regarding the image object <b>120</b> produced by the image sensor <b>116</b> is provided to the preamplifier <b>208</b> for amplification. The controller <b>204</b> may control the amount of gain imparted by the preamplifier <b>208</b> to the electronic image signal <b>216</b> received by the preamplifier <b>208.</b> The amplified image signal <b>220</b> may then be provided to the storage device and/or display device <b>212.</b> The image information stored or displayed by a storage device and/or display device <b>212,</b> in accordance with the present invention, includes information calculated by the controller <b>204.</b> Such information includes color information related to pixels. In particular, the controller <b>204</b> is capable of determining color information related to areas of the image sensor <b>116</b> with respect to which color information is not provided as part of the electronic image signal <b>216.</b> The function of the controller <b>204</b> in determining this color information will be described in greater detail below.

With reference now to <figref idref="f0002"><b>Fig. 3</b></figref><b>,</b> the imaging surface <b>304</b> of an image sensor <b>116</b> in accordance with an embodiment of the present invention is illustrated. As shown in <figref idref="f0002"><b>Fig. 3</b></figref><b>,</b> the pixels <b>124</b> comprising the imaging surface <b>304</b> of the image sensor <b>116</b> are distributed across the imaging surface <b>304</b> in rows and columns. However, it should be appreciated that other arrangements of pixels <b>124</b> may be utilized. For example, the position of pixels in a first row may be staggered from the position of pixels in a second row to create a pattern that does not result in the pixels being arranged in continuous columns.

With reference now to <figref idref="f0002"><b>Fig. 4</b></figref><b>,</b> a cross-section of an individual pixel <b>124</b> is depicted. The pixel <b>124</b> shown in <figref idref="f0002">Fig. <b>4</b></figref> is a charge coupled device. A charge coupled device type pixel <b>124</b> converts light <b>404</b> incident on the active surface of the pixel <b>124</b> into an electronic charge. This electronic charge is collected in a potential well <b>408</b> that is established by applying a voltage to the silicone substrate <b>412</b> through <b>a</b> gate electrode <b>416</b> to which a positive voltage is applied. The electrons are confined in the well <b>408</b> during the exposure period by a potential barrier <b>420</b> created by applying negative voltages to barrier electrodes <b>424.</b> In a typical device, the electrodes <b>416, 424</b> are insulated from the substrate <b>412</b> by a layer of silicone dioxide <b>428.</b>

Also illustrated in <figref idref="f0002"><b>Fig. 4</b></figref> is a filter element or microfilter <b>432.</b> A filter element <b>432</b> is associated with selected pixels <b>124</b> to allow the image sensor <b>116</b> to provide color information. The filter element <b>432</b> generally functions to allow one of three spectral regions (<i>i.e</i>. colors) to pass through to the pixel <b>124.</b>

After the desired period of exposure has elapsed, the barrier <b>420</b> is lowered, and the collected charge is transferred to a serial register. In general, charge collected by rows of pixels <b>124</b> are shifted into the serial register. Alternatively, an entire array of pixels <b>124</b> may be shifted into an identically sized storage array during readout of the image.

In <figref idref="f0002">Fig. <b>5</b></figref><b>,</b> a filter <b>504</b> for providing color information in connection with an image sensor <b>116</b> in accordance with the prior art is illustrated. In general, the filter <b>504</b> includes a plurality of microfilter elements <b>508.</b> In general, a single filter element <b>508</b> is provided for each pixel <b>124</b> (see <figref idref="f0002"><b>Fig. 3</b></figref>) of the image sensor <b>116.</b> Accordingly, each pixel <b>124</b> receives light within a selected frequency band. According to alternative methods for providing color information, a filter <b>504</b> may operate in sequence to provide a first color to the pixels <b>124</b> (<i>e.g</i>., green) and the information provided while that first frequency band is in effect read out. Next, the filter <b>504</b> may provide filtered light within a second frequency band (<i>e.g</i>., red), and the information obtained during operation of the filter mode providing the second frequency band read from the pixels <b>124.</b> Finally, a third filter mode may be entered by the filter <b>504</b> to provide the pixels <b>124</b> with a third frequency band (<i>e.g</i>., blue). Under either scheme, it can be appreciated that the pixels <b>124</b> of the image sensor <b>116</b> always receive filtered light information in a color camera in accordance with the prior art. Therefore, during operation of such a camera, some available photons are removed from the light that might otherwise be available to the pixels comprising the image sensor <b>116.</b>

The filter scheme illustrated in <figref idref="f0002"><b>Fig. 5</b></figref> utilizes subgroups of pixels <b>124.</b> Such a subgroup <b>604</b> is generally illustrated in <figref idref="f0002"><b>Fig. 6</b></figref><b>.</b> From <figref idref="f0002"><b>Fig. 6</b></figref><b>,</b> it is apparent that each subgroup <b>604</b> contains four filter elements <b>508.</b> One filter element allows red light to pass, and another filter element <b>508</b> allows blue light to pass. In addition, two filter elements <b>508</b> allow green light to pass. Because of the importance of green light in providing luminance information, and because color information can be adequately conveyed using a relatively small portion of the available bandwidth, the use of two green filter elements <b>508</b> per subgroup <b>604</b> tends to improve the sensitivity of the image sensor <b>116,</b> while maintaining a relatively large amount of color information.

The inventor of the present invention has recognized that the sensitivity of an image sensor <b>116</b> could be further improved for use in low light applications if the number of pixels <b>124</b> receiving filtered light was reduced, thereby making more of the photons reflected from images in a scene available to the pixels <b>124.</b> Furthermore, the inventor has recognized that color information that is adequate for aiding in object recognition can be supplied by providing filtered light to a very small proportion of pixels 124 of an image sensor 116.

With reference now to <figref idref="f0003">Fig. 7A</figref>, an arrangement of pixels 124 across the imaging surface of an image sensor 116 in accordance with an embodiment of the present invention is depicted. In general, the pixels 124 are arranged in a plurality of subsets 704. For purposes of clarity, it should be appreciated that only some of the pixels 124 of the imaging sensor 116 are depicted in <figref idref="f0003">Fig. 7</figref>.

With reference now to <figref idref="f0003">Fig. 7B</figref>, a subset of pixels 704 is illustrated. As shown in <figref idref="f0003">Fig. 7B</figref>, a group of color information pixels 708 is associated with color filters or color filter elements 432. Alternatively, pixels 124 that are responsive to light confined to select frequency bands (i.e. colors) may comprise the group of color information pixels <b>708.</b> In the example of <figref idref="f0003"><b>Fig. 7B</b></figref><b>,</b> the group of pixels <b>708</b> consists of three pixels, one of which receives light passing through a red filter <b>432,</b> another that receives light passing through a green filter <b>432,</b> and a third that receives light passing through a blue filter <b>432.</b> Furthermore, it should be appreciated that the majority of pixels <b>124</b> included in the subset of pixels <b>704</b> do not receive filtered light. For instance, in the example subset of pixels <b>704</b> illustrated in <figref idref="f0003"><b>Fig. 7B</b></figref><b>,</b> there are 81 pixels, 79 of which receive unfiltered light, and 3 of which (or less than 4%) receive filtered light. Accordingly, the majority of pixels <b>124</b> in the subset of pixels <b>704</b> are provided with unfiltered light. In this example then, 96% of the pixels <b>124</b> in the subset of pixels <b>704</b> are sensitive to light of full spectral distribution, with the result that the sensitivity of the imaging device <b>100</b> is minimally impacted by the group of color information pixels <b>708.</b>

In accordance with an embodiment of the present invention, the pixels <b>124</b> that are not among the group of pixels <b>708</b> receiving filtered light are sensitive to light having wavelengths from about 400nm to about 1050nm. In accordance with a further embodiment of the present invention, the color information pixel included in the group of the pixels <b>708</b> receiving red filtered light is provided with light having a wavelength from about 540nm to about 780nm, the pixel receiving green filtered light is provided with light having a wavelength from about 470nm to about 610nm, and the pixel receiving blue filtered light is provided with light having a wavelength from about 380nm to about 510nm.

In <figref idref="f0003"><b>Fig. 7C</b></figref><b>,</b> a subset of pixels <b>704</b> having a group of color information pixels <b>708</b> in accordance with another embodiment of the present invention is illustrated. The subset of pixels <b>704</b> in <figref idref="f0003"><b>Fig. 7C</b></figref> are arranged in a 9x9 array of pixels <b>124,</b> with each of the color information pixels having an associated color filter element <b>432.</b> Alternatively, pixels <b>124</b> that are responsive to light confined to select frequency bands may comprise the group of color information pixels <b>708.</b> In the example of <figref idref="f0003"><b>Fig. 7C</b></figref><b>,</b> the group of pixels <b>708</b> consists of one pixel <b>124</b> that is responsive to red light, a second pixel <b>124</b> that is responsive to green light, and a third pixel <b>124</b> that is responsive to blue light. Furthermore, the pixels <b>124</b> receiving color information are located diagonally from their neighbor color information pixel <b>124.</b> As in the example subset of pixels <b>704</b> of <figref idref="f0003"><b>Fig. 7B</b></figref><b>,</b> the majority of the pixels in the subset of pixels <b>704</b> of <figref idref="f0003"><b>Fig. 7C</b></figref> receive light that has not been color filtered. Accordingly, the majority of the pixels <b>124</b> provide unfiltered, or full spectrum, luminance information, for high sensitivity. In the embodiment shown in <figref idref="f0003"><b>Fig. 7C</b></figref><b>,</b> less than 4% of the pixels <b>124</b> in the subset of pixels <b>704</b> receive filtered light.

With reference now to <figref idref="f0003"><b>Fig. 7D</b></figref><b>,</b> a group of color information pixels <b>708</b> included in a subset of pixels <b>704</b> in accordance with another embodiment of the present invention is illustrated. The group of color information pixels <b>708</b> illustrated in <figref idref="f0003"><b>Fig. 7D</b></figref> differs from the group of color information pixels <b>708</b> illustrated in <figref idref="f0003"><b>Fig. 7B</b> and <b>7C</b></figref> in that the group of color information pixels <b>708</b> illustrated in <figref idref="f0003"><b>Fig. 7D</b></figref> implements a Bayer filter. Color information regarding the imaged object <b>120</b> is obtained only from those pixels in the group of color information pixels <b>708.</b> Therefore, the group of color information pixels <b>708</b> receiving light filtered to within selected frequency bands represents a small proportion (<i>i.e</i>. less than half) of the total pixels <b>124</b> within a subset of pixels <b>704.</b> The remainder of the pixels <b>124</b> provide luminance information. This arrangement provides high sensitivity, by limiting the number of pixels <b>124</b> receiving filtered light as compared to prior art systems.

As can be appreciated, a group of color information pixels <b>708</b> preferably includes at least one pixel <b>124</b> sensitive to green light, at least one pixel <b>124</b> sensitive to red light, and at least one pixel <b>124</b> sensitive to blue light. As can further be appreciated, it is generally desirable to group the individual pixels <b>124</b> of a group of color information pixels <b>708</b> in close proximity to one another. However, the particular spacing and geometric relationship of the pixels <b>124</b> in a group of color information pixels <b>708</b> may be varied.

With the addition of a switch in the controller <b>204</b> or other processing electronics associated with the imaging device <b>100,</b> the chrominance signals red, green and blue can be turned off and the system will perform almost identically to that of a monochrome camera having no color filtering. Such a switch can serve to provide color information to the displayed image only when color discrimination is desired or merited. A switch capable of disabling the chrominance signals can also be used to reduce noise. For example, when light levels are very low and signal gain for the chrominance signal is high, the noise associated with the high gain levels can be removed by turning off the signals from the color information pixels <b>708.</b> When the chrominance signal is turned off, luminance information for the affected pixels can continue to be supplied by neighboring pixels <b>124</b> receiving a full spectral distribution.

Although the description set forth herein contains examples of color information pixels <b>708</b> comprising additive color schemes, such as RGB color schemes, it should be appreciated that the present invention may also utilize a complementary color scheme. For example, the color information pixels <b>708</b> may comprise a subtractive color scheme, such as Cyan Magenta Yellow (CMY). A subset of pixels <b>704</b> utilizing a subtractive color scheme may include a group of color information pixels <b>708</b> in which at least one pixel <b>124</b> is sensitive to cyan light, at least one pixel <b>124</b> is sensitive to magenta light, and at least one pixel <b>124</b> is sensitive to yellow light. The remaining pixels <b>124,</b> which comprise the majority of the pixels included in the subset of pixels <b>704,</b> provide luminance information.

In <figref idref="f0004"><b>Fig. 8</b></figref><b>,</b> an alternative arrangement of pixel subsets are depicted. In particular, the image sensor <b>116</b> illustrated in <figref idref="f0004"><b>Fig. 8</b></figref> comprises square pixel subsets <b>804,</b> and rectangular pixel subsets <b>808.</b> The rectangular pixel subsets <b>808</b> are the result of staggering the subsets of pixels <b>804</b> with respect to one another, so that continuous columns of pixel subsets <b>804</b> are not formed. The subsets of pixels <b>804, 808</b> each contain groups of pixels <b>708</b> that receive color filtered light. The groups of pixels <b>708</b> represent a small proportion (<i>i.e</i>. less than half) of the total number of pixels <b>124</b> available within a subset of pixels <b>804, 808.</b> By staggering the subsets of pixels <b>804, 808,</b> the color information obtained by the groups of pixels <b>708</b> can reduce the aliasing that can occur as a result of the relatively sparse color information obtained by an image sensor <b>116</b> in accordance with the present invention.

With reference now to <figref idref="f0005"><b>Fig. 9</b></figref><b>,</b> the surface of an image sensor <b>116</b> in accordance with yet another embodiment of the present invention is illustrated. In <figref idref="f0005"><b>Fig. 9</b></figref><b>,</b> groups of color information pixels <b>708</b> comprising one pixel provided with red filtered light, one pixel provided with green filtered light, and one pixel provided with blue filtered light are distributed about diagonal lines across the surface of the image sensor <b>116.</b> This arrangement can also reduce the effects of the aliasing of color information.

With reference now to <figref idref="f0005"><b>Fig.10</b></figref><b>,</b> still another image sensor <b>116</b> in accordance with an embodiment of the present invention is illustrated. In <figref idref="f0005"><b>Fig. 10</b></figref><b>,</b> groups of color information pixels <b>708</b> receiving color filtered light are arranged about the surface of the image sensor <b>116</b> randomly. In accordance with an alternative embodiment of the present invention, the groups of color information pixels are distributed pseudo-randomly about the surface of the image sensor <b>116.</b> In general, the distribution of groups of color information pixels <b>708</b> in horizontal and vertical lines should be minimized to minimize aliasing of the dominant horizontal and vertical lines found in most objects. In accordance with an embodiment of the present invention, less than 10% of the pixels <b>124</b> of the image sensor <b>116</b> receive color filtered light. In accordance with a further embodiment of the present invention, less than 5% of the pixels <b>124</b> of the image sensor <b>116</b> receive color filtered light.

The present invention undersamples color information because the majority of pixels <b>124</b> receive full spectrum light. <figref idref="f0006"><b>Fig.11</b></figref> is a schematic representation of how color is assigned to areas of the image <b>120</b> that coincide with areas of the image sensor <b>116</b> comprising pixels <b>124</b> that receive full spectrum light (<i>i.e</i>. that do not receive color filtered light). In order to suppress visible artifacts that result from the invention's undersampling of color, the color signals that are received should be filtered or smoothed along both the vertical and horizontal axes of the image sensor <b>116.</b> Filtering can be realized by convoluting the color channel signals with a suitable kernel. A less computationally intense approach is to provide a form of linear interpolation of the color signals. In doing so, advantage can be taken of the relatively low resolution color, compared to the limiting resolution of the entire system, in that color information can be delayed without major degradation due to image motion smear. That is, if image motion is great enough that a frame delay of the color or chrominance information is significantly impacted, the luminance signal would be degraded to the point of questionable usefulness.

One implementation of a linear interpolation approach is illustrated in <figref idref="f0006"><b>Fig.11</b></figref><b>.</b> In <figref idref="f0006"><b>Fig.11</b></figref><b>,</b> a single color channel and associated pixels <b>1104</b> are illustrated for purposes of clarity. The color samples are taken on a square grid having a pitch (P) equal to the dimensions of a subset of pixels <b>704.</b> That is, the pitch as it relates to color information is equal to the distance between adjacent groups of color information pixels <b>708</b> (represented by single color information pixels <b>1104</b>). The digitized central values for each group <b>708</b> are illustrated as A(i, j), A(i + p, j). For two adjacent groups 708 in the horizontal direction, and A(i, j), A(i, j+p) in the vertical direction. In a progressive scan device, as the image is being read from the image sensor 116, the color value for columns of pixels 124 located between the group of pixels 708 providing the central color values are estimated by suitable algorithm. For example, the color value for pixels located in columns between A(i, j) and A(i, j+p) can be estimated by: <maths id="math0001"><math display="block"><mi mathvariant="normal">A</mi><mfenced><mi mathvariant="normal">i</mi><mi mathvariant="normal">j</mi></mfenced><mo mathvariant="normal">=</mo><msubsup><mrow><mo mathvariant="normal">|</mo></mrow><mrow><mi mathvariant="normal">n</mi><mo mathvariant="normal">=</mo><mi mathvariant="normal">j</mi><mo mathvariant="normal">+</mo><mfenced><mi mathvariant="normal">p</mi><mo mathvariant="normal">-</mo><mn mathvariant="normal">1</mn></mfenced></mrow><mrow><mi mathvariant="normal">n</mi><mo mathvariant="normal">=</mo><mi mathvariant="normal">j</mi><mo mathvariant="normal">+</mo><mn mathvariant="normal">1</mn></mrow></msubsup><mo>⁢</mo><mfenced open="[" close="]"><mi mathvariant="normal">A</mi><mfenced><mi mathvariant="normal">i</mi><mi mathvariant="normal">j</mi></mfenced><mo mathvariant="normal">+</mo><mfenced><mi>n x</mi><mspace width="1em" /><msub><mi mathvariant="normal">δ</mi><mi mathvariant="normal">j</mi></msub></mfenced></mfenced></math><img file="EP1530873B1_D0001.tif" /></maths> where <maths id="math0002"><math display="block"><msub><mi mathvariant="normal">δ</mi><mi mathvariant="normal">j</mi></msub><mo mathvariant="normal">=</mo><mfrac><mrow><mi mathvariant="normal">A</mi><mo>⁢</mo><mfenced><mi mathvariant="normal">i</mi><mo mathvariant="normal">,</mo><mi mathvariant="normal">j</mi><mo mathvariant="normal">+</mo><mi mathvariant="normal">p</mi></mfenced><mo mathvariant="normal">-</mo><mi mathvariant="normal">A</mi><mfenced><mi mathvariant="normal">i</mi><mi mathvariant="normal">j</mi></mfenced></mrow><mi mathvariant="normal">p</mi></mfrac></math><img file="EP1530873B1_D0002.tif" /></maths> Accordingly, the column values between A(i, j) and A(i, j+p) are calculated using the algorithm. Likewise, the column values between A(i+p, j) and A(i+p, j+p) are also determined using the above algorithm.

As the next frame is read, all of the color values between columns containing pixels providing color information are estimated by a similar algorithm. That is, the values between A(i+1, j) and A(i+(p-1), j) where the index for i ranges from 1 to p - 1, thus denoting any line value between A(i+1, j) and A(i+1, j+(p-1)). As these row color values are interpolated, the column values of the current frame being read are also calculated. In general, interpolation is carried out for each color channel in parallel and used as the smoothed or filtered chrominance signal. The edge values for the outside blocks use the nearest neighbor computed or digitized central value as appropriate. The color values of the pixels <b>124</b> that do not receive filtered light are used to provide the high resolution luminance signal. The luminance signal values for the locations occupied by the color sensing pixels <b>708</b> (<i>i.e</i>. the pixels receiving filtered light) can be interpolated from neighbor luminance values.

As an example of the improved sensitivity of a color imaging device <b>100</b> in accordance with the present invention, as compared to a conventional color imaging device, consider an image sensor <b>116</b> comprising 1000 rows of pixels and 1000 columns of pixels. Furthermore, assume that 100 independent color samples are taken in both the horizontal and vertical axes. For purposes of the present example, a Bayer filter type group of pixels <b>708</b> may be used for obtaining color information. Accordingly, there will be a total of 100 x 100 = 10,000 pixels receiving color information. If those pixels receiving color information are considered to provide no contribution to the luminescence signal provided by the image sensor <b>116,</b> only four percent of the total photon capture area provided by the image sensor <b>116</b> has been lost. Accordingly, the sensitivity of an image sensor <b>116</b> in accordance with the present invention is comparable to the sensitivity of a monochromatic image sensor. However, the present invention provides color information. In addition, an image sensor in accordance with the present invention is capable of providing luminance information from light that extends into infrared wavelengths, again providing increased sensitivity as compared to conventional color imaging systems. Furthermore, by obtaining color information from the groups of color sensing pixels, and using the remaining pixels as sources of luminance information, an imaging device <b>100</b> in accordance with the present invention provides the correct hue, although color saturation will often be inaccurate.

In accordance with an embodiment of the present invention, luminance signal values are provided with respect to pixels <b>124</b> sensitive to selected spectral bands or colors from neighboring pixels. For example, a luminance signal for a pixel <b>124</b> receiving filtered light may be interpolated from one or more adjacent pixels that receive unfiltered light. For instance, in connection with the pixel subset <b>704</b> arrangement shown in <figref idref="f0003"><b>Fig. 7C</b></figref><b>,</b> the luminance value for each color information pixel may be derived from the luminance values for the pixels <b>124</b> on the left and right using the equation {L_left + L_right)/2, where L is the luminance value. As a further example, the luminance value for the pixel <b>124</b> preceding a color information pixel <b>124</b> can simply be repeated using a sample and hold circuit.

In <figref idref="f0007"><b>Fig. 12</b></figref><b>,</b> a chromaticity diagram for an ideal color imaging system is illustrated. The display of the ideal system is set as if the scene illumination was that of a black body having a color temperature of 6500 Kelvin. The displaying primaries are as defined in specification Rec. 709 for HDTV. The gamut of colors that can be reproduced by the example ideal system fall within the triangle <b>1204.</b> In <figref idref="f0007"><b>Fig. 12</b></figref><b>,</b> the hue and saturation of 13 test colors are shown by marks numbered 1-13. The output provided by this ideal system for each of the 13 test colors corresponds exactly with those test colors, with the exception of test sample no. 5, which is located outside of the range of colors that can be reproduced by the ideal system, given the reproducing primaries. Accordingly, the color of sample no. 5 is reproduced at the edge of the triangle, in line with reference white <b>1208,</b> as shown by the vector emanating from the color's actual chromaticity coordinates, as delineated by the "+" next to reference no. 5. That is, the color is reproduced at the end of the vector. This is the best representation that the ideal system can provide for that color.

In the ideal system referred to with respect to <figref idref="f0007"><b>Fig. 12</b></figref><b>,</b> the spectral range of the test colors used in the simulation is from 400nm to 1100nm, covering both the visible and near infrared portions of the spectrum. The camera's Red-Green-Blue (RGB) taking characteristics cover only the visible portion of the spectrum, being nil for wavelengths longer than 780nm. The luminance signal for the system is derived by summing the RGB signals in the proper proportions, as done in an NTSC TV broadcast signal.

In <figref idref="f0008"><b>Fig. 13</b></figref><b>,</b> the performance of an imaging system <b>100</b> in accordance with the present invention that provides a luminance signal generated with full spectrum (<i>e.g</i>. 400nm to 1100nm) responsitivity. For example, such responsitivity can be obtained using a thinned, rear illuminated CCD with an epitaxial thickness of 14 microns that is provided with light that has not been color filtered. For purposes of the test results shown in <figref idref="f0008"><b>Fig.13</b></figref><b>,</b> the gain of the chrominance signal is several times that of the luminance signal. The gamut of colors that can be reproduced by the imaging system <b>1100</b> fall within the triangle <b>1304.</b> It will be noted that most of the test colors 1-13 are reproduced by the imaging system <b>100</b> in error. However, it will also be noted that the test colors are reproduced by the imaging system <b>100</b> with the correct hue. This is shown by the orientation of the vectors associated with each of the test colors; the vectors are in line with reference white <b>1208.</b> The errors generally occur with respect to the saturation of the test colors. Nonetheless, by providing the correct hue, an imaging system 100 in accordance with the present invention is capable of providing color information sufficient to aid in the task of object recognition, while providing good low light level performance (<i>i.e</i>. high sensitivity). Furthermore, the provision of color information is obtained, even while providing pixels operating at full spectral bandwidth (<i>i.e</i>. including infrared portions of the spectrum) to provide even higher sensitivity.

<figref idref="f0009"><b>Fig. 14</b></figref> depicts aspects of an interlaced imaging device <b>100</b> in accordance with an embodiment of the present invention. In such a device, the center of collected information is shifted between the first and second fields comprising a frame of image data. Because of the shift in the center of the signal collection site, the requirements for the placement of the color microfilters needed in connection with pixels providing color information is potentially problematic.

In <figref idref="f0009"><b>Fig. 14</b></figref><b>,</b> one solution for providing color information in connection with an embodiment of the present invention utilizing a pseudo-interlaced system is illustrated. In <figref idref="f0009"><b>Fig. 14</b></figref><b>,</b> a four phase device is used for illustration purposes. Field 1 <b>1404</b> uses phase 1 <b>1408</b> as the barrier phase <b>1412</b> for the signal <b>1416</b> that is collected by phases 2 <b>1420,</b> 3 <b>1424</b> and 4 <b>1428.</b> The color microfilter <b>432</b> is centered on the field 1 <b>1404</b> collection site in the vertical axis. Above and below the microfilter <b>432,</b> there is no spectral filter, and therefore the full spectrum falls on the other pixels. In field 2 <b>1436</b>, the placement of the microfilter <b>432</b> results in two collection sites <b>1440</b> and <b>1444</b> that do not receive full spectrum information. Accordingly, collection sites <b>1440</b> and <b>1444</b> are discarded from the signal provided in connection with field 2 <b>1436,</b> and color samples are only taken with respect to field 1 <b>1404.</b>

In accordance with an embodiment of the present invention, the amount of chrominance signal gain can be adjusted according to the light levels available in a scene. For example, at very low light levels, the chrominance signal gain can be reduced to reduce noise. The variation in signal gain can be implemented in connection with an automatic gain control (AGC) circuit.

In addition, the relatively low resolution chrominance signal, as compared to the luminance signal, provided by the present invention, allows a frame delay with respect to color channel information, with little or no deterioration in the provided image. The implementation of a frame delay can provide additional time for signal processing and for spatial filtering of the chrominance signal. The chrominance signals may also be averaged over a number of frames to improve the signal to noise ratio of the color information.

The foregoing discussion of specific embodiments of the invention has been presented, for purposes of illustration and description. Further, the description of specific embodiments is not intended to limit the invention. Consequently, variations and modifications of the specific embodiments commensurate with the above teachings, within the skill and knowledge of the relevant art, within the scope of the appended claims are within the scope of the present invention. The embodiments described hereinabove are further intended to explain the best mode presently known of practicing the invention and to enable others skilled in the art to utilize the invention.