Multilinear array sensor with an infrared line

Information

  • Patent Grant
  • 6590679
  • Patent Number
    6,590,679
  • Date Filed
    Tuesday, January 26, 1999
    27 years ago
  • Date Issued
    Tuesday, July 8, 2003
    23 years ago
Abstract
Surface defect correction technology for photographic images requires an infrared scan along with a conventional color scan. In the present invention, the additional infrared scan needed for surface defect correction is obtained by adding a line of sensors specific to infrared light to a conventional multilinear color sensor array. The invention teaches a practical mode of distinguishing infrared light using a dichroic prism placed over the sensor. This mode has the additional advantage of placing the infrared-specific sensor line in a displaced focus plane to match conventional lenses. Adding a sensor line to a conventional trilinear sensor array requires a quadrilinear array topology. In addition to the direct quadrilinear topology, the invention teaches a method of obtaining full color image information with only two linear sensor lines by interstitially mixing red and blue sensors on a single sensor line, which, in conjunction with the additional infrared line, results in a conventional trilinear sensor topology with a different filter arrangement.
Description




TECHNICAL FIELD OF THE INVENTION




This invention relates to the scanning of photographic images, and more particularly in a primary application, to scanning in infrared and visible light in order to prepare for correction of surface defects.




BACKGROUND OF THE INVENTION





FIG. 1

shows a prior art trilinear film scanner, and also introduces some terms that will be used in this application. A lamp


102


transilluminates a filmstrip


104


containing an image


106


to be scanned. Normally the light from the lamp


102


would be diffused or directed by additional optics, not shown, positioned between the lamp


102


and film


104


in order to illuminate the image


106


more uniformly. The image


106


on the film


104


is focused by lens


108


onto a sensor line


110


in a circuit package


112


. The sensor line


110


projects back through lens


108


as a line


116


across the image


106


. This line


116


is composed of many individual points, or pixels. To scan the entire image


106


, the film


104


is moved perpendicularly to the line


116


to scan a two dimensional area, such as image


106


. Because the sensors of the sensor line


110


are positioned in lines, this arrangement is called a linear, or line, sensor.




The sensor line


110


may be of a form known in the art as a “trilinear”, or three line, array. As shown magnified at


120


, the sensor line


110


actually consists of three parallel lines of sensors. In this prior art embodiment, one line of sensors


122


is behind a line of red filters


124


. This arrangement could consist of a series of independent filters, but is normally a single long red filter


124


which covers all of the sensors of line


122


. Another line of sensors


126


is behind a green filter line


128


, and a third line of sensors


130


is behind a blue filter line


132


.




As the film


104


is moved, the three lines


122


,


126


, and


130


each provide an individual image of the film seen with a different color of light. The data from the circuit package


112


is sent along cable


136


to supporting electronics and computer storage and processing means, shown together as computer


138


. Inside computer


138


the data for each color image is grouped together, and the three images are registered as the three color planes


140


,


142


, and


144


of a full color image. Each of these color planes


140


,


142


and


144


consists of pixels describing with a number the intensity of the light at each point in the film. For example, pixel


150


of the red color plane


144


may contain the number “


226


” to indicate a near white light intensity at point


152


on the film


104


, as measured at a specific sensor


154


in the array


110


, shown enlarged in circle


120


as sensor


156


behind the red filter line.




In

FIG. 1

it is noted that there is a spacing between sensor lines


122


,


126


and


130


, and therefore the same point on the film


104


is not sensed by all three color lines at the same point in time.

FIG. 2

illustrates this registration problem in more detail.




In

FIG. 2

there is a trilinear array (not shown) with red, green, and blue sensor lines


202


,


204


, and


206


. These lines are projected onto a substrate (not shown) which is moved in the direction of the arrow to scan out regions of the image on the substrate. The region seen by each line is different from the region seen by the other lines. For example, at the beginning of an arbitrary time interval, sensor


210


of the blue line


206


may see point


212


of the substrate, while at the end of the time interval, it may see point


214


. It is apparent that each of the different sensors


210


,


220


and


230


sees a different area during the same time interval. For example, at the end of the time interval, sensor


220


of the red sensor array


202


sees point


222


, which is different than point


214


seen by the blue sensor


210


at the same end time. However, if the time interval is long enough, there will exist a region of overlap


224


over which all array lines have passed. If the interval between measurements is an integer submultiple of the spacing between the arrays, then there exists a time at which sensor


230


of the green line


204


sees the same point


232


on the substrate as


214


, and another time at which sensor


220


of the red line


202


sees point


234


, the same as point


214


, which in turn will be seen by the blue sensor


210


at a later time. The computer system


138


receiving the information from the scans made by the trilinear array registers the data representing the three color images by shifting the data an amount corresponding to the distance between sensor lines, and discarding the part of each color record outside the full color range overlap


224


.




Although this illustration has presented a so-called transmission, or film, scanner, a reflection, or print, scanner uses the same principles except that the source light is reflected from the same side as the imaging lens. As is explained later, there are uses for the present invention in both transmission and reflection scanners.




The conventional scanners described above scan in the three visible colors, exclusive of the invisible infrared. There are several reasons that it would be useful to add an infrared record registered to the conventional colored records. For example, examination of old documents under infrared with a reflection scanner is proving useful in examination of historic works, such as the Dead Sea Scrolls, to disclose alterations. Another potential use presented here without admission that it is known in the art, is to distinguish the “K” or black channel from the cyan, magenta, and yellow channels in a four color print. Currently a major commercial use of infrared plus visible scans is a technology called infrared surface defect correction, as explained in FIG.


3


. Current applications of infrared surface defect correction are limited to transmission scanners, although it may be extended to reflection scanners, and therefore the specific illustration of a transmission scanner given below is not to be considered a limitation.




In

FIG. 3

, a lamp


302


transilluminates filmstrip


304


containing an image


306


. An electronic camera


308


views the image


306


and outputs red, green, and blue digitized records


310


,


312


, and


314


. In addition the electronic camera


308


outputs an infrared record


316


. There are several ways a conventional camera can be made responsive to selectively visible and infrared light. One way is to provide a filter wheel


320


with four filters: red


322


, green


324


, blue


326


, and infrared


328


. If the camera


308


is a monochrome camera whose sensitivity extends into infrared, then the three visible colors and infrared may be captured at four different times, each time illuminating the film with a different filter in the filter wheel


320


.




The cyan, magenta, and yellow dyes that create the image


306


are all transparent to infrared light, and therefore the film


304


appears clear to camera


308


when viewed under infrared light. On the other hand, surface defects such as dust, scratches, and fingerprints refract the light passing through the film


304


away from the camera


308


, and therefore appear as darkened points under both visible and infrared light. Because refraction under infrared light is nearly equal to refraction under visible light, the defects appear nearly as dark in the infrared as in the visible spectrum.




Therefore infrared record


316


is effectively of a clear piece of film including defects, and image


310


contains the same defects plus the red image. The infrared image


316


provides a pixel by pixel “norming” for the effect of defects. For example, defect-free pixel


340


in the red record


310


may contain a 50% brightness measurement. The corresponding defect-free pixel


342


in the infrared record


316


contains 100% brightness because no defect has attenuated the light. Function block


344


divides the 50% brightness level from the red record


310


by the norming 100% brightness level from the infrared record


316


to give a 50% brightness measurement for corrected pixel


346


. On the other hand, pixel


350


under scratch


352


in the red record


310


may contain a 40% brightness measurement. The corresponding pixel


354


in the infrared record


316


seeing the same scratch may contain 80% brightness because the scratch has refracted 20% of the light. When function block


344


divides 40% by 80%, a corrected brightness value of 50% is determined for pixel


356


. Note that corrected pixels


346


and


356


within the same background area of the image now both contain the same brightness value of 50%, so the effect of the scratch has disappeared. This division is repeated for each pixel to produce the corrected red record


360


; and the same division by infrared is applied to the green record


312


, and blue record


314


, to produce the corrected green and blue records


362


and


364


, resulting in a full color corrected image.




There are several ways of generating an infrared scan in conjunction with a visible scan. One method makes four passes across the original image using a light that changes color between passes, as was shown in FIG.


3


. Unfortunately, this can take four times as long as a single pass scanner. Alternately, one can make a single pass while flashing four lights in rapid succession, but again the hardware may need to move at one fourth the speed. None of these prior art methods combines the speed obtained with a single pass multilinear array with the image clarity possible in the prior art attained by making multiple scans. It is apparent that the introduction of such a system would provide an improvement to the state of the art in infrared surface defect correction, as well as to the other uses of combined infrared and visible scans mentioned above.




SUMMARY OF THE INVENTION




The present invention adds a line to a conventional multilinear sensor array. The added line is specific to the infrared scan. In the most direct embodiment, the added line makes what was a trilinear array containing three lines, one for each of three primary colors, into a quadrilinear array. In a second embodiment, the red and blue sensor lines are combined into one line that alternates between red and blue sensors. This second embodiment uses only two lines for sensing full color, allowing the third line of existing trilinear layouts to be devoted to infrared.











BRIEF DESCRIPTION OF THE DRAWINGS




For a more complete understanding of the present invention and for further advantages thereof, reference is now made to the following Description of the Preferred Embodiments taken in conjunction with the accompanying Drawings in which:





FIG. 1

shows a conventional trilinear film scanner;





FIG. 2

illustrates registration of a multilinear array;





FIG. 3

explains the operation of infrared surface defect correction;





FIG. 4

shows the present invention with a quadrilinear line and infrared filter;





FIG. 5

graphs the color transmission of available filters;





FIG. 6

shows the preferred embodiment with a dichroic prism;





FIG. 7

illustrates the infrared focus shift common to imaging lenses;





FIG. 8

charts the filter arrangement of a trilinear infrared sensor;





FIG. 9

charts the filter arrangement of an alternate trilinear infrared sensor; and





FIG. 10

presents missing color recovery used with the sensor shown in FIG.


7


.











DESCRIPTION OF THE PREFERRED EMBODIMENTS





FIG. 4

shows a prior art trilinear array with the addition of the novel fourth sensor line for infrared scanning. In this figure, sensor line


402


contains individual photosites


404


, each behind a red filter. In perspective, it is seen that the line


402


consists of a row of silicon photosensor sites


406


behind red filter material


408


to render the row of photosites


406


responsive primarily to red light. Similarly, the photosites in sensor lines


410


and


412


are made primarily responsive to green and blue light, respectively. Together these three lines


402


,


410


and


412


, in conjunction with their overlaid filters, form a prior art trilinear sensor array. These three silicon lines with their overlaid filters are typically contained in a package


414


under cover glass


416


.




The present invention teaches the addition of another line


420


to the device which is specific to infrared light. Silicon sensor material is inherently sensitive to the lower end of the infrared spectrum, so the layout and construction of the extra sensor line


420


, shown in perspective as line


424


having photosites


422


, may be a copy of one of the other three lines, such as line


428


or


430


. The manufacture of a silicon sensor line, and the duplication of multiple copies of a silicon circuit, called “macros” onto a silicon die, are well known in the art.




Line


420


is made primarily responsive to infrared light by removing visible light reaching the photosensor sites


422


with a line of infrared-passing, visible-absorbing filter material


424


. Infrared filter material


424


would appear black to the human eye. A number of such infrared materials are known in the art. As an example, a double layer filter, the first layer consisting of the filter material


408


printed to make the red line


402


, overlaid with a second layer consisting of the filter material printed to make the green line


428


, would together absorb visible light and transmit infrared. In fact, any two or three of the visible colors combined would absorb visible and pass infrared light. Because these filter materials are already present in the printing of the other two lines, this method would enable manufacture of the infrared filter


424


without requiring any additional dye types in the fabrication process.




Unfortunately, a problem arises when using the device described thus far. This problem is explained using

FIG. 5

wherein the transmission of typical organic colored dyes commonly used to form the colored filter lines of

FIG. 4

are graphed. Although the infrared line is rendered primarily responsive to only infrared light by the infrared filter, the other colors are specific not only to their labeled visible color, but to infrared light as well, as is seen by observing that the labeled color graphs in

FIG. 5

transmit infrared light. When a scan is made in which each of the visible records also contains infrared light, a faded and excessively surface-defect-sensitive image results. Therefore, in the prior art, either a light source was used that had no infrared content, or an infrared blocking filter was placed somewhere in the light path, such as at the light source itself or as a component of the cover glass in the sensor circuit package.




In the present invention, in order to overcome this problem and sense infrared in one sensor line, infrared light cannot be blocked before reaching the sensor package. Therefore, the infrared blocking filter must be combined with the visible filters at the sensors. In one embodiment of the present invention, the visible color filters of lines


408


,


428


, and


430


of

FIG. 4

are manufactured with a process that absorbs or reflects infrared light. Such a process may use multiplayer interference filters commonly known in the art. These filters require many layers, and in order to produce three colors on the same substrate, the substrate would need to be deposited many times resulting in difficult and expensive manufacturing. As depicted in

FIG. 4

, noninfrared (which includes visible light) and infrared light are simultaneously substantially blocked as the noninfrared and infrared light pass through the filters to their respective sensors.




As an improved embodiment,

FIG. 4

teaches the use of a cut piece of infrared blocking material


431


laid over the three visible lines


408


,


428


, and


430


, but having a terminating edge


432


between the last visible color sensor line


430


and the infrared line


424


. This blocking material


431


, shown also as filter


434


in circuit package


414


, could be placed on top of the cover glass


416


, manufactured as part of the cover glass, or ideally placed under the cover glass and directly over the sensor lines.




An infrared absorbing filter, such as is available from Schott Optical of Germany, typically is relatively much thicker than the organic colored filters, and therefore creates a shadowing parallax problem at the edge


432


over sensor lines that are very close. This shadowing may be minimized by moving the thick filter closer to the sensor lines when placing it under the cover glass


416


, and is also made less objectionable by moving the infrared sensor line


424


further from the last visible line


430


. In the arrangement of

FIG. 4

, the spacing S


2


between the visible and infrared lines is made greater than the spacing S


1


between the visible lines. In particular, for maximum step resolution flexibility, the ratio S


2


/S


1


should be an integer. In the specific illustration of

FIG. 4

, the ratio of S


2


/S


1


is two.




A preferred embodiment of the present invention which uses a prism to separate visible from infrared light is shown in FIG.


6


. This embodiment has the added advantage of shifting the infrared focus plane to correct for common chromatic aberrations caused by imaging lenses.




In

FIG. 6

the three visible color sensor lines


602


,


604


, and


606


lie under red, green, and blue filter lines


608


,


610


, and


612


, respectively, as previously illustrated in FIG.


4


. Also novel infrared sensor line


614


is added to practice the present invention, but unlike

FIG. 4

, an infrared filter line over sensor line


614


is optional. All these lines are housed within circuit package


620


under a transparent cover glass


622


. Also included is a dichroic prism


624


which may be mounted on the cover glass


622


, incorporated as a part of the cover glass (shown separately at


626


), or incorporated under the cover glass in contact with the sensor lines as illustrated in the enlarged illustration


600


of FIG.


6


.




In the enlarged illustration


600


, a first prism


630


has a surface


632


coated such that a light ray


634


is split into a transmitted visible component


636


and a reflected infrared component


638


. Such a surface coating is commonly known in the art as a “hot mirror”, because the heat, or infrared, component is reflected. Hot mirrors are available from Edmund Scientific Corporation of Barrington, N.J. Bonded to the first prism


630


is a second prism


640


constructed with parallel surfaces such that the reflected infrared component


638


is further reflected at surface


642


so as to be directed toward infrared sensor line


614


. The reflection surface


642


is preferably coated so as to enhance reflection while preventing light incident from above the second prism


640


from penetrating to the sensor line


614


. A coating material commonly used in mirrors for high infrared reflectivity is gold.




The apparatus of

FIG. 6

has a significant advantage over the apparatus of

FIG. 4

, as explained with reference to FIG.


7


. In

FIG. 7

, it is noted that the focus plane for infrared light is displaced relative to the focus plane for visible light. The focus difference is shown greatly exaggerated in FIG.


7


. In a typical achromatic lens, the focus shift


702


is about 0.25% of the visible focal length for infinity which would be about 0.5% of the combined visible focal distance


704


of the lens in

FIG. 7

operating at a unity magnification. Returning to

FIG. 6

, it is noted that the virtual image


650


of the infrared line


614


is displaced by distance


652


below the visible sensor lines. Distance


652


is equal to D


1




654


divided by the index of refraction to infrared light of prism section


640


. The displacement distance D


1




654


is defined by the characteristics of prism section


640


, and is preferably chosen to be the distance between the infrared sensor line and the middle visible sensor line.




Note that the distance D


1




654


is easily controlled at manufacture as a function of the thickness of prism section


640


and is not affected by misalignments in laying the prism over the sensor lines. Therefore, an otherwise high level of precision is not required in the cleanroom where the circuits are packaged. The displacement distance


652


also introduces a slight magnification of the infrared image which can be corrected through a resize algorithm, such as is commonly known in the art.




Although it is conceptually simple to duplicate a fourth silicon sensor line, in practice it is very expensive to make any change to an existing silicon fabrication process. In addition, more silicon is required for the extra line along with more electronics to support the extra data. Accordingly it would be an advantage if an existing trilinear line could be adapted to the present invention. Such an adaptation is shown in

FIG. 8

which illustrates the application of the invention with other than a quadrilinear array.




In

FIG. 8

, a conventional trilinear silicon sensor is adapted to practice the current invention by altering only the filters deposited over the silicon sensors, and not the silicon layout itself. To practice the invention, one of the sensor lines


802


is rendered specific to infrared scanning by one of the methods discussed above to substantially block noninfrared light. One such method, presented above, uses a prism with an infrared reflecting dichroic coating.




With one of the three sensor lines dedicated to infrared, only two sensor lines remain to receive three visible colors. One line


804


receives a primary color, chosen as green in the preferred embodiment. The last array


806


must therefore sense the remaining two colors. This can be done by alternating the remaining two colors interstitially at pixel boundaries, such as by making sensors receptive to red


810


in even rows such as row


16


and to blue


812


in odd rows such as row


15


, as illustrated in FIG.


8


.




The primary color may be green as presented above; however, in an embodiment more conservative of photons for use in low light, the primary color used for line


804


is chosen to be white, namely the visible light remaining after removal of infrared light with an infrared blocking filter, but with no auxiliary color filter. The two alternating colors in line


806


are then chosen as either cyan and orange or cyan and light red. Such a combination would approximately double the number of photons impinging on the silicon sensors after passing through the colored filters, and therefore the luminance noise would be less due to lowered shot noise. This lowering of luminance noise comes, however, at the expense of weaker color distinction, requiring color amplification and thereby causing increased color noise. In low light this has been found to be a practical trade-off.




The embodiment just described uses row


15


as an example of a row in which each point on a substrate, after scanning, has been sensed with infrared, green, and blue light, and row


16


as an example of a row in which each point on the substrate, after scanning, has been sensed with infrared, red, and green light. It does not matter significantly in what time order the colors are sensed. The color filter topology of

FIG. 9

is seen to produce the same color combinations, and in particular rows


15


and


16


of

FIG. 9

are seen to sense the same colors as lines


15


and


16


of

FIG. 8

, albeit at different times. In fact, the infrared sensors also could obviously be interspersed on the other lines in any of the embodiments illustrated in this application without departing from the scope of the invention. The embodiment shown in

FIG. 9

is still considered to have a green selective line; the line has just been interspersed in other lines.




In the reproduction of an image made with the color topology taught in

FIG. 8

, it is necessary to recreate the missing color in each row. Turning to

FIG. 10

, the sensed colors are named for each pixel of each row, including rows


15


and


16


used in the above examples. The missing colors for each pixel of each row are named in parentheses. It is those colors that need to be estimated using available known color information. The specific color measurements are named using a nomenclature relative to the center pixel: “c” for center, “t” for top, “tr” for top right, “r” for right, “br” for bottom right, “b” for bottom, “bl” for bottom left, “l” for left, and “tl” for top left. In this example Rc, the unmeasured red value for the center pixel, will be estimated. Although this example will estimate the value for one red pixel, the same algorithm could apply to any red pixel within any even row, or the blue value on odd rows by interchanging red and blue wording.




The unmeasured value of Rc will be calculated by combining estimates based on color information from the surrounding pixels. These estimates are named as the six estimates Et, Etr, Ebr, Eb, Ebl, and Etl. In a preferred embodiment, the diagonal estimates contribute to the determination of the unmeasured value just as do the vertical estimates, but at a reduced strength, namely Rc=(Et+Eb)/4+(Etl+Etr+Ebl+Ebr)/8. In another embodiment, more efficient of computer time, the diagonal estimates may be completely ignored, namely Rc=(Et+Eb)/2.




Next, calculation of the estimates using Et and Eb by name will be disclosed. Any of the other estimates Etl, Etr, Ebl, and Ebr may be similarly calculated with the change in nomenclature. Most basically, each estimate can be simply the red value of the adjacent pixel, namely Et=Rt and Eb=Rb. It is possible to improve the estimate by using the green value that is known for all pixels. In particular, it can be assumed, because the real world tends to be rather monochromic, that the red value changes with position about as fast as the green value changes. Therefore, a particular estimate, such as Et, will use not only the red value of the adjacent pixel, but will adjust this value by the amount green changes from the adjacent pixel to the center pixel for which red is being estimated.




For example, Et=Rt+(Gc−Gt) and Eb=Rb+(Gc−Gb) to use the linear change in color. Alternately Et=Rt*Gc/Gt and Eb=Rb*Gc/Gb to use the percent change in color. When the color measurements are directly proportional to lumens, the linear change is found to work better when Rt>Gt or Rb>Gb, and the percent change works better when Rt<Gt or Rb<Gb. The equations produce the same result when Rt=Gt or Rb=Gb. When the color measurements are expressed in the common gamma correct space wherein values are proportional to the square root of lumens, the linear change is found to be acceptable in all cases, but with the resultant values clamped so as not to go negative.




Whereas the present invention has been described with respect to specific embodiments thereof, it will be understood that various changes and modifications will be suggested to one skilled in the art and it is intended to encompass such changes and modifications as fall within the scope of the appended claims.



Claims
  • 1. A system for scanning a substrate in visible and infrared light, the substrate including a plurality of parallel sensor lines, the system including:means for substantially blocking noninfrared light to a first sensor of the plurality of sensor lines; and means for simultaneously substantially blocking infrared light to a second sensor of the plurality of sensor lines.
  • 2. The system of claim 1 wherein said infrared light blocking and noninfrared light blocking means comprise a prism.
  • 3. The system of claim 2 wherein the prism includes a first surface having an interference filter to transmit noninfrared light and reflect infrared light.
  • 4. The system of claim 3 wherein the prism further includes a second surface substantially parallel to the first surface which redirects the reflected infrared light to the first sensor line.
  • 5. The system of claim 1 and further including means for substantially blocking infrared light to a third sensor line of the plurality of sensor lines, wherein the second sensor line is made primarily responsive to visible light of a first color, and the third sensor line is primarily responsive to visible light of a second color.
  • 6. The system of claim 5 and further including:a first filter for passing light of the visible light of the first color, free of infrared light to the second sensor line, and a second filter for passing light of the visible light of the second color, free of infrared light to the third sensor line.
  • 7. The system of claim 5 and further including:means for substantially blocking infrared light to a fourth sensor line of the plurality of sensor lines, wherein the fourth sensor line is primarily responsive to a visible light of a third color.
  • 8. The system of claim 5 and further including:an infrared blocking filter for blocking infrared light to the second and third sensor lines, and a first filter for passing light of the visible light of the first color plus infrared light to the second sensor line, and a second filter for passing light of the visible light of the second color to the third sensor line.
  • 9. The system of claim 8 wherein the infrared blocking filter reflects infrared light toward the first sensor line.
  • 10. The system of claim 8 wherein the infrared blocking filter absorbs infrared light.
  • 11. The system of claim 8 wherein the infrared blocking filter reflects infrared light.
  • 12. The system of claim 5 wherein the third sensor line further comprises a plurality of individual sensors wherein a first individual sensor of the third sensor line is primarily responsive to a high color within the visible light of the second color, and a second individual sensor of the third sensor line is primarily responsive to a low color with the visible light of the second color.
  • 13. The system of claim 12 wherein the color of the visible light of the second color is magenta, the high color is blue, and the low color is red.
  • 14. The system of claim 13 wherein the color of the visible light of the first color is green.
  • 15. The system of claim 12 wherein the color of the visible light of the first color is white.
  • 16. The system of claim 15 wherein the high color is cyan.
  • 17. The system of claim 12 wherein the low color corresponding to the position of the first individual sensor is estimated as a function of the low color measured by the second individual sensor, and the difference in the first color measured by respective individual sensors of the second sensor lines at positions corresponding to the first individual sensor and the second individual sensor.
  • 18. A sensor responsive to visible light and infrared light, comprising:a substrate, a first linear sensor disposed on said substrate and being pninarily responsive to visible light of a first color; a second linear sensor disposed on said substrate and being primarily responsive to visible light of a second color; a third linear sensor disposed on said substrate and being primarily responsive to visible light of a third color; a fourth linear sensor disposed on said substrate and being primarily responsive to infrared light; means for substantially blocking infrared light from said first, second, and third linear sensors; and means for simultaneously substantially blocking visible light from said forth linear sensor.
  • 19. The sensor of claim 18 wherein said linear sensors are spaced apart by a distance, such that the distance between said first and second sensors and the distance between said second and third sensors is less than the distance between said third and fourth sensors.
  • 20. The sensor of claim 18 wherein said blocking means includes a prism.
RELATED APPLICATION

This application relies on U.S. Provisional Application Serial No. 60/073,602, filed Feb. 4, 1998, and entitled “Multilinear Array Sensor With An Infrared Line.”

US Referenced Citations (76)
Number Name Date Kind
4260899 Baker Apr 1981 A
4301469 Modeen et al. Nov 1981 A
4302108 Timson Nov 1981 A
4462860 Szmanda Jul 1984 A
4670779 Nagano Jun 1987 A
4677465 Alkofer Jun 1987 A
4680638 Childs Jul 1987 A
4700229 Herrmann et al. Oct 1987 A
4775238 Weber Oct 1988 A
4796061 Ikeda et al. Jan 1989 A
4845551 Matsumoto Jul 1989 A
4933566 Masaaki et al. Jun 1990 A
4937720 Kirchberg Jun 1990 A
4969045 Haruki et al. Nov 1990 A
4972091 Cielo et al. Nov 1990 A
4989973 Noso et al. Feb 1991 A
4994918 Lingemann Feb 1991 A
5003379 Moore, Jr. et al. Mar 1991 A
5010401 Murakami et al. Apr 1991 A
5047968 Carrington et al. Sep 1991 A
5058982 Katzir Oct 1991 A
5091972 Kwon et al. Feb 1992 A
5097521 Massmann Mar 1992 A
5149960 Dunne et al. Sep 1992 A
5155596 Kurtz et al. Oct 1992 A
5200817 Birnbaum Apr 1993 A
5266805 Edgar Nov 1993 A
5267030 Giorgianni et al. Nov 1993 A
5291286 Murakami et al. Mar 1994 A
5311310 Jozawa et al. May 1994 A
5335086 Kitamura Aug 1994 A
5371542 Pauli et al. Dec 1994 A
5447811 Buhr et al. Sep 1995 A
5448380 Park Sep 1995 A
5452018 Capitant et al. Sep 1995 A
5465155 Edgar Nov 1995 A
5465163 Yoshihara et al. Nov 1995 A
5477345 Tse Dec 1995 A
5509086 Edgar et al. Apr 1996 A
5516608 Hobbs et al. May 1996 A
5552904 Ryoo et al. Sep 1996 A
5561611 Avinash Oct 1996 A
5565931 Girod Oct 1996 A
5568270 Endo Oct 1996 A
5581376 Harrington Dec 1996 A
5582961 Giorgianni et al. Dec 1996 A
5583950 Prokoski Dec 1996 A
5589887 Wischermann Dec 1996 A
5608547 Nakatani et al. Mar 1997 A
5641596 Gray et al. Jun 1997 A
5666443 Kumashiro Sep 1997 A
5673336 Edgar et al. Sep 1997 A
5721624 Kumashiro et al. Feb 1998 A
5726773 Mehlo et al. Mar 1998 A
5729631 Wober et al. Mar 1998 A
5771107 Fujimoto et al. Jun 1998 A
5801373 Oozu Sep 1998 A
5808674 Adams, Jr. et al. Sep 1998 A
5892595 Yamakawa et al. Apr 1999 A
5923042 Mietta et al. Jul 1999 A
5930388 Murakami et al. Jul 1999 A
5963662 Vachtsevanos et al. Oct 1999 A
5969372 Stavely et al. Oct 1999 A
5979011 Miyawaki et al. Nov 1999 A
5982941 Loveridge et al. Nov 1999 A
5982951 Katayama et al. Nov 1999 A
5991444 Burt et al. Nov 1999 A
6005987 Nakamura et al. Dec 1999 A
6057040 Hage May 2000 A
6075905 Herman et al. Jun 2000 A
6078051 Banton et al. Jun 2000 A
6078701 Hsu et al. Jun 2000 A
6094281 Nakai Jul 2000 A
6101273 Matama Aug 2000 A
6128416 Oura Oct 2000 A
6239886 Klasser et al. May 2001 B1
Foreign Referenced Citations (33)
Number Date Country
28 21 868 Nov 1979 DE
196 86 867 Jan 1998 DE
1 547 811 Jun 1979 EP
0 422 220 Apr 1991 EP
0 482 790 Apr 1992 EP
0 527 097 Feb 1993 EP
0 569 142 Nov 1993 EP
0 624 848 Nov 1994 EP
0 699 753 Aug 1995 EP
0 716 538 Jun 1996 EP
0 716 538 Jun 1996 EP
0 751 670 Jan 1997 EP
0 768 621 Apr 1997 EP
0 794 454 Sep 1997 EP
0 816 833 Jan 1998 EP
0 816 833 Aug 1998 EP
0 893 914 Jan 1999 EP
2 283 633 May 1995 GB
4-291139 Oct 1992 JP
11185028 Jul 1999 JP
2000-13604 Jan 2000 JP
2000-196813 Jul 2000 JP
WO 8402019 May 1984 WO
WO 8906890 Jul 1989 WO
WO 9001240 Feb 1990 WO
WO 9109493 Jun 1991 WO
WO 9205469 Apr 1992 WO
WO 9515530 Jun 1995 WO
WO 9716028 May 1997 WO
WO 9831142 Jul 1998 WO
WO 9834397 Aug 1998 WO
WO 9940729 Aug 1999 WO
WO 0148694 Jul 2001 WO
Non-Patent Literature Citations (12)
Entry
New Adaptive Vector Filter Based on Noise Estimate, Mei Yu, et al., IEICE Trans Fundamentals, vol. E82, No. 6, Jun., 1999.
A Robust Method for Parameter Estimation of Signal-Dependent Noise Models in Digital Images, B. Aiazzi, et al., IEEE, 1997, pp. DSP 97—601-604.
A Regularized Iterative Image Restoration Algorithm, Aggelos K. Katsaggelos, et al., IEEE, 1991, pp. 914-929.
Adaptive Fourier Threshold Filtering: A Method to Reduce Noise and Incoherent Artifacts in High Resolution Cardiac Images, M. Doyle, et al., 8306 Magnetic Resonance in Medicine, May 31, 1994, No. 5, Baltimore, MD., pp. 546-550.
Anisotropic Spectral Magnitude Estimation Filters for Noise Reduction and Image Enhancement, Til Aach, et al., IEEE, 1996, pp. 335-338.
Adaptive-neighborhood filtering of images corrupted by signal-dependent noise, Rangaraj M. Rangayyan et al., Applied Optics, vol. 37, No. 20, Jul. 10, 1998, pp. 4477-4487.
Grayscale Characteristics, Photographic Negatives The Nature of Color Images, pp. 163-168.
Digital Imaging Equipment White Papers, Putting Damaged Film on Ice, A Technical Discussion of Advances in Digital Imaging, Nikon Corporation, http://www.nikonusa.com/reference/whitepapers/imaging/ditechdisc.html, Aug. 5, 1999.
Local Cosine Transform—A Method for the Reduction of the Blocking Effect in JPEG, Gil Aharoni, et al., Journal of Mathematical Imaging and Vision, 3, 7-38, 1993.
Digital ICE, Applied Science Fiction, Inc., http://www.asf.com/html/o_products/iceprod.html. Aug. 5, 1999.
About Digital ICE Technology, Applied Science Fiction, Inc., http://www.asf.com/html/o_products/icetech,html. Aug. 5, 1999.
2-D Adaptive Volterra Filter for 2-D Nonlinear Channel Equalisation and Image Restoration, J.N. Lin, et al., Electronic Letters, vol. 28, No. 2, Jan. 16, 1992, pp. 180-182.
Provisional Applications (1)
Number Date Country
60/073602 Feb 1998 US