.
`United States Patent
`
`[191
`
`lllllllllllllllllllllllllllIllllllllllllllllllllllllIllllIlllllllllllllllll
`US0053O3O64A
`[11] Patent Number:
`
`5,303,064
`
`t
`
`Johnson et al.
`
`[45] Date of Patent:
`
`Apr. 12, 1994
`
`[54]
`
`IMAGE SCANNER WITH CALIBRATION
`MECHANISM TO OBTAIN FULL DYNAMIC
`RANGE AND COMPENSATED LINEAR
`OUTPUT
`
`Primary Examiner—Edward L. Coles, Sr.
`Assistant Exam1‘ner—Jill Jackson
`Attorney, Agent. or Ft'rm—Flehr, Hohbach, Test,
`Albritton & Herbert
`
`ABSTRACT
`[57]
`A calibrated scanner is disclosed. The calibrated scan-
`ner includes a sensor board with a Charge Coupled
`Device (CCD) having two output signals. The sensor
`board also includes two offset devices, a coarse offset
`and a fine offset, the coarse offset being connected to
`the two output signals,‘the fine offset being connected
`to one of the output signals. The offsets are adjusted to
`provide a nonclipping DC bias for the two output.sig-
`nals. The sensor board also includes two gain devices,
`one gain device being attached to each output signal.
`The 831" d8ViC€5 3“? adjusted 10 balance th¢fW° OUTPU;
`signals. After these adjustments, the saturation level 0
`the CCD is determined. The saturation level informa-
`tion is then used in readjusting the gain and offset de-
`vices. After this calibration, a linear correction may be
`applied to the output signal of the sensor board. The
`linear correction is accomplished by determining dark
`an<1:1hligjht 5/ajlueks oftihle splnsorlboard. Thesg valuels, alpng
`wt
`1 ea
`ar
`an
`ig
`va ues, are use
`0 so ve wo
`linear equations to determine M and b values. The M
`and b values are then used to provide direct linear cor-
`rection to the data leaving the sensor board.
`
`12 Claims, 11 Drawing Sheets
`
`SET COLOR
`
`[76]
`
`[56]
`
`Inventors: Kent Johnson, 1038 Neilson St.,
`Albany, Calif. 94706; Lynn L. Ackler,
`354 walnut 51” Alameda, ca1if_
`94501; David C, Jenkins, 5406 Ridge
`Park Dr., Loomis, Calif. 95650;
`Howard H. Barney, 2237 Carleton
`St., Berkeley, Calif. 94703; Chris A.
`3I'001<S. 1301 Clement AVE-1 131dg-
`27, Mameda. C3115 94501
`[2]] Appl‘ No‘. 658,177
`_
`F°b' 20! 1991
`filed‘
`[221
`Int. Cl.5 ............................................... H04.\' 1/46
`[51]
`
`[52] us, C1, __
`358/406- 353/504;
`358/512; 358/483; 358/513; 358/526; 358/479
`[S8] Field of Search ............... .. 358/406, 76, 479, 482,
`358/483, 487, 213.15, 213.16, 213.18, 213.19,
`213.27, 213-28. 213-31, 445, 512, 504, 513, 514,
`513, 520; H04N 1/46
`References Cited
`U5‘ PATENT DOCUMENTS
`_ 4,307,423 12/1981 Atherton . . . . ..
`.. . .. .. 358/213.15
`4539-034
`5/1936 Y°k°miZ°
`353/433
`4,694,356 9/1987 Constable
`358/527
`5,144,448
`9/1992 Hornblower, III et al.
`358/483
`5,153,421 10/1992 Tandon ei al.
`............... .. 358/213.16
`350
`
`
`
`3501\
`FOCUS OPTICAL
`ASSEMBLY
`
`/' 371
`
` 3Z3
`
`
`
`CALCULATE
`SATURATION
`EXPOSURE
`
`
`
`32?
`
`
`
`N0
`CALCULATE
`CORRECTION
`
`
`
`
`
`Apple 1047
`
`U.S. Pat. 6,470,399
`
`
`
`SET NOMINAL
`
`OFFSET
`
`
`
`DEFINE NOMINAL
`EXPOSURE
`
`CALCULATE
`OFFSETS
`
`CALCULATE
`GAINS
`
`
`
`CALCULATE
`
`SATURATION
`EXPOSURE
`
`CALCULATE
`OFFSETS
`
`CALCULATE
`GAINS
`
`
`
`Apple 1047
`U.S. Pat. 6,470,399
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 1 of 11
`
`5,303,064
`
`
`
`/-'15, 1
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 2 of 11
`
`5,303,064
`
`FIG. 3
`
`%$
`§§
`\§
`§§
`%%
`
`%$\%§§
`
`

`
`tHef.aPS”U
`
`.Apr.12,1994
`
`Sheet3 ofll
`
`460,303,5
`
`a2<4
`
`>4m¢:m
`
`
`
`4mz<mpzomm
`
`gm~
`
`mam
`
`»-
`
`asw
`
`Vbw
`
`mzH4zH>4mm:mmm;om_Eemsa»»Eam
`NNNVNN
`xmaE駧:z8as
`
`
`
`mz<4¢xo<mm>Hmm<¢m4mHh<¢zouh<\u¢
`
`8»
`
`as»»s~
`
`VSm
`
`9284<Em_m
`
`,ms»
`
`oz<omago
`
`
`
`zmooozmmu<Hmm<o
`
`msm
`
`uma
`
`Ema:ohpomzoo
`
`‘|__
`.682I2:“so:£58mos:808
`I9:...as:$5:55:E5:
`u<#_mz_mqsgs
`9:...mzo:.62~55:
`
`
`mobs:mo<Hzz<o
`EmR
`
`msw
`
`arm
`
`M_
`
`NE
`
`..§
`
`
`
` Iom<ommomzmm
`
`mm
`
`
`
`E23.,._e2.:225E5928mg
`
`N3V3
`
`qmfi
`
`
`
`
`

`
`U.S. Patent
`
`Apr. 12,1994
`
`Sheet 4 of 11
`
`5,303,064
`
`ms»om<omago
`
` wq»
`
`vx
`
`m2
`
`
`
`928as5E3zé.meow.mos.E1.928
`Hmommagnum:<mawn
`
`:3B><BEEHHm..“mm:
`
`
`
`momzmm4<mm
`
`om<ommo<zH
`
`:<mmax
`
`
`
`xmHaom<:mzH<wmp<4:o4<o
`
`momhumumomh<4:o4<o
`
`5.
`
`IIIIIzoHHoummou
`m»<4=o4<o
`
`
`
`zoHp<m=»<mm»<4:o4<o
`
`my.mymm
`
`
`
`

`
`..lHef.anr3U
`
`M
`
`5,303,064
`
`
`
`
`
`8.“. mI228AAso<n_EmtmVnm.MAhrm$E:n__2<IxmIt_8aA%<1éfizuwmtaE3
`
`asm-_AzH<ovzxm.u
`05EmmaAA<82>mm
`
`Mmgoouowo\<%pHmu~_mmmmoo<msmwAomhzooaICzocsmzmo
`93Emum~122.v_oo..o
`
`
`
`~_EmH8~_S._.Hmum
`
`mam
`
`,A
`
`m_o<Eo._.Eofizosa
`
`928mommmoozm
`
`4!
`
`Ahmmuuov
`
`M928
`
`
`
`
`
`
`
`Smoz5%<mEmacsS:<dEoo\mEE:n_z<mm._
`o<o»Hmummmm»4Hmmm<¢go;
`
`Sm:\mImam
`TQEm
`
`
`
`
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 6 of 11
`
`5,303,064
`
`350w\
`
`FOCUS OPTICAL
`ASSEMBLY
`
`DEFINE NOMINAL
`
`EXPOSURE
`
`‘
`
`CALCULATE
`
`OFFSETS
`
`A
`
`CALCULATE
`
`GAINS
`
`‘
`
`CALCULATE SATURATION
`
`EXPOSURE
`
`350
`
`353
`
`371
`
`323
`
`SET °°L°
`
`R
`
`SET NOMINAL
`
`OFFSET
`
`SET NOMINAL
`
`GAIN
`
`354
`
`36
`
`CALCULATE
`
` 5
`
`
`
`
`SATURATION
`
`353
`
`EXPOSURE
`
`CALCULATE
`OFFSETS
`
`354
`
`355
`
`CALCULATE
`
`OFFSETS
`
`CALCULATE
`
`GAINS
`
`{I
`
`_
`
`N0
`
`370
`
`CALCULATE
`
`GAINS
`
`356
`
`370
`
`331
`
`3z2
`
`E
`
`NO
`
`I
`/_-I6 7
`
`I CALCULATE
`CORRECTION
`
`-
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 7 of 11
`
`5,303,064
`
`3544
`
`\
`
`401
`
`
`
`
`
`BLOCK LIGHT
`
`INCREMENT
`
`COARSE
`
`OFFSET
`
`STORE VALUE
`
`FIG. 8
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 8 of 11
`
`5,303,064
`
`3543 \
`
`40!
`
`BLOCK LIGHT
`
`424
`
`
`
`
`
`495
`
`4&6
`
`432
`
`434
`
`STORE PIXEL
`
`VALUES
`
`RANGE
`COMPLETE?
`
`NO‘
`
`FIND CLOSEST
`
`BALANCE
`
`
`
`INCREMENT FINE
`OFFSET
`
`
`
`STORE VALUE
`
`
`
`FIG. 9
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 9 of 11
`
`5,303,064
`
`355\
`
`457
`
`
`EVERY PIXEL
`BELOW
`SATURATION?
`
`
`
`N0
`
`476'
`
`STORE VALUE
`
`FIG. 10
`
`EXPOSE LIGHT
`
`
`INCREMENT
`GAIN
`
`
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 10 of 11
`
`5,303,064
`
`
`
`3&Bw\\\
`
`EXPOSE LIGHT
`
`467
`
`502
`
`504
`
`INCREMENT
`
`EXPOSRE TIME
`
`
`
`YES
`
`STORE VALUE
`
`FIG. 11
`
`
`SATURATION?
`
`508
`
`
`
`

`
`U.S. Patent
`
`Apr. 12, 1994
`
`Sheet 11 of 11
`
`310
`
`.\\
`
`401
`
`BLOCK LIGHT
`
`457
`
`EXPOSE
`LIGHT
`
`403
`
`4a3
`
`
`SET MINIMUM
`EXPOSURE FOR
`COLOR
`
`SCAN
`
`STORE VALUE
`XDARK MIN
`
`
`
`SET MAXIMUM
`
`
`
`EXPOSURE
`FOR COLOR
`
`SET MINIMUM
`EXPOSURE FOR COLOR
`
`
`
`SCAN
`
`STORE VALUE
`XLIGHT
`
`N0
`
`511
`
`CALCULATE
`
`XDARK
`
`4
`
`532
`
`534
`
`535
`
`532
`
`537
`
`SOLVE LINEAR
`
`EQUATION
`
`554
`
`//"556
`
`
`
`
`
`SCAN
`
`STORE VALUE
`XDARK MAX
`
`3”
`® F15. 12
`
`NO
`
`

`
`1
`
`5,303,064
`
`IMAGE SCANNER WITH CALIBRATION
`MECHANISM TO OBTAIN FULL DYNAMIC
`RANGE AND COMPENSATED LINEAR OUTPUT
`
`BRIEF DESCRIPTION OF THE INVENTION
`
`This invention relates to scanners which scan film
`and generate digital data representing the film’s image.
`More particularly, this invention relates to a scanner
`which utilizes a calibration procedure to obtain optimal
`digital data of a scanned image.
`
`-
`
`l0
`
`l5
`
`BACKGROUND OF THE INVENTION _
`A number of difficulties are associated with image
`scanner technology. Specifically, the transformation of
`an analog signal of the image into a digital signal of the
`image is accompanied by a number of difficulties. First,
`the analog signal is subject to variation with heat, hu-
`midity, and other external factors which affect the qual-
`ity of the resultant digital data. Next, the analog signal 20
`includes a noise portion which is transferred into the
`digital signal in the absence of proper processing. An-
`other problem with analog signal manipulation in slide
`scanners is the saturation of analog devices within the
`scanner. These problems often result in a signal which
`includes erroneous data in a non-linear form. To correct
`this data requires a great deal of processing power.
`By calibrating the sensor board, or analog electron-
`ics, on a scanner, an improved signal may be obtained.
`As a result, subsequent signal processing is reduced,
`therefore a smaller processor may be employed.
`Calibration of a scanner is difficult since many ele-
`ments associated with the scanner have interdependent
`relationships. Thus, adjusting one element will result in
`changes to other elements. Prior artattempts at calibrat-
`‘ing scanners have relied upon a one-pass approach to
`calibrating elements.
`OBJECTS AND SUMMARY OF THE
`INVENTION
`
`25
`
`30
`
`35
`
`Thus it is a primary object of the present invention to
`provide a calibrated image scanner.
`It is a related object of the present invention to pro-
`vide a scanner which is calibrated to increase the dy-
`namic range of the resultant image.
`It is still another object of the present invention to
`provide a scanner which minimizes the analog manipu-
`lation of the image data.
`It is another object of the present invention to pro-
`vide a scanner which includes a method for correcting
`the resultant signal to obtain a linear response.
`It is yet another object of the present invention to
`provide a scanner which operates below the saturation
`level of its analog devices.
`It is another object of the present invention to pro-
`vide a scanner which calibrates itself through an itera-
`tive procedure.
`These and other objects are obtained by a calibrated
`scanner in accordance with the present invention. The
`calibrated scanner includes a sensor board, or analog
`electronics, with a Charge Coupled Device (CCD)
`having two output signals. The sensor board also in-
`cludes two offset devices, a coarse offset and a fine
`offset, the coarse offset being connected to the two
`output signals, the fine offset being connected to one of
`the output signals. The offsets are adjusted to provide a
`non-clipping DC bias for the two output signals. The ‘
`sensor board also includes two gain devices, one gain
`
`45
`
`50
`
`55
`
`65
`
`2
`device being attached to each output signal. The gain
`devices are adjusted to balance the two output signals.
`After these adjustments,
`the saturation level of the
`CCD is determined. The saturation level information is
`then used in readjusting the gain and offset devices.
`After this calibration, a linear correction may be applied
`to the output signal of the sensor board. The linear
`correction is accomplished by determining dark and
`light values of the sensor board. These values, along
`with ideal dark and light values, are used to solve two
`linear equations to determine M and b values. The M
`and b values are then used to provide direct linear cor-
`rection to the data leaving the sensor board.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`Other objects and advantages of the invention will
`become apparent upon reading the following detailed
`description and upon reference to the drawings,
`in
`which:
`FIG. 1 is a side cross-sectional view of a scanner
`which may be used in accordance with the present
`invention.
`
`FIG. 2 is top cross-sectional view of the scanner of
`FIG. 1.
`FIG. 3 is a front cross-sectional view of the scanner
`of FIG. 1.
`
`FIG. 4 is a block diagram of scanner electronics
`which may be utilized in accordance with the present
`invention.
`
`FIG. 5 is a block diagram of the data flow within a
`scanner utilized in accordance with the present inven-
`tion.
`FIG. 6 is a schematic view of a sensor board of a
`scanner which may be employed in accordance with the
`present invention.
`FIG. 7 is a flow diagram of the calibration method of
`the present invention.
`FIG. 8 is a flow diagram of the method of calculating
`coarse offsets, in accordance with the present invention.
`FIG. 9 is a flow diagram of the method of calculating
`fine offsets, in accordance with the present invention.
`FIG. 10 is a flow diagram of the method of calculat-
`ing gains, in accordance with the present invention.
`FIG. 11 is a flow diagram of the method of calculat-
`ing saturation exposure, in accordance with the present
`invention.
`FIG. 12 is a flow diagram of the method of calculat-
`ing correction, in accordance with the present inven-
`tion.
`
`DETAILED DESCRIPTION OF THE
`INVENTION
`
`Turning now to the drawings, wherein like compo-
`nents are designated by like reference numerals in the
`various figures, attention is initially directed to FIG. 1.
`Depicted therein is a cross sectional view of a slide
`scanner 20 in accordance with a preferred embodiment
`of the present invention. Slide scanner 20 includes a
`housing 22 with an optical region 24 and an electronics
`region 26. A door 28 is positioned on the housing 22
`above the optical region 24. The door includes a light
`source 30 and provides access to a film carriage 32. Film
`33 is placed in film carriage 32. The film carriage 32
`transports the film 33 across the optical assembly 60, in
`accordance with prior art techniques.
`The optical assembly 60 receives the image from the
`film 33 as illuminated by light source 30. The image
`
`

`
`3
`projected by light source 30 is reflected from mirror 62
`into infrared filter 63 and into lens assembly 64, which
`includes a plurality of lenses 65. From the lens assembly
`64, the image passes through a filter of the RGB filter
`assembly 70. The filter assembly 70 positions either a 5
`red, green, or blue filter in the light path, as is known in
`the art. After being filtered, as described, the image
`impacts upon sensor 72, which comprises a plurality of
`photosites (or sensors). Sensor 72 is coupled to sensor
`board 74.
`,
`Preferably, a zoom assembly 66 allows the lens as-
`sembly 64 and sensor 72 to move in reference to film 33.
`Analogously, focus assembly 68 preferably allows the
`lens assembly 64 to move in reference to sensor 72.
`Methods known in the art may be utilized to construct
`zoom assembly 66 and focus assembly 68. As a_result,
`the optical assembly 60 may be adjusted to accommo-
`date a variety of film sizes. Table I provides data defin-
`ing an optimal configuration of optical assembly 60 for
`various film sizes.
`
`TABLE I
`
`Lens Width
`32 mm
`Focal Length
`65 mm
`F-Number
`4.5
`Resolution
`100 line pairs/rnm
`
`
`
` Film Size 35 mm 2.25 mm X 2.25 mm 4" X 5"
`
`
`
`
`
`472 mm
`327 mm
`268 mm
`Object to Image
`375 mm
`218 mm
`132 mm
`Object to Lens
`65 mm
`78 mm
`104 mm
`Lens to Image
`19.5 mm
`21.6 mm
`19.6 mm
`Image Length
`
`Object Length 30 24.9 mm 54.9 mm 99.1 mm
`
`
`
`
`The folded configuration of the optical assembly 60
`allows for a compact scanner 20. The door 28 allows for
`easy access to the film carriage 32. Since only the film
`carriage 32 moves during scanning (while the light 35
`source 30, lens assembly 64, and mirror 62 remain sta-
`ble), jitter and other problems are reduced.
`Turning to FIG. 2, depicted therein is a top view of
`scanner 20. Electronics region 26 preferably includes
`serial board 100, SCSI board 102, Digital Signal Pro-
`cessing (DSP) board 104, motor board 106, CPU board
`108, and hard disk controller 124. The figure also de-
`picts lamp power supply 110 and hard disk 112. Prefera-
`bly, a power supply 144 is positioned beneath the differ-
`ent boards in the electronics region 26.
`Another view of scanner 20 is provided in FIG. 3.
`FIG. 3 is a front view of the scanner, which in addition
`to showing previously described elements, also depicts
`linear encoder 114, as to be more fully described herein.
`Returning to FIG. 2 and the elements depicted
`therein, the serial board 100 is a half-slot board which
`provides means for directly controlling the scanner
`activity by sending low level commands from a terminal
`emulator. This direct control of the scanner is useful for
`debugging. The presence of this board is preferable, it is
`not required for the scanner 20 to operate under normal
`conditions.
`
`40
`
`45
`
`50
`
`55
`
`SCSI board 102 preferably includes a NCR 53C94
`chip and a floppy disk interface. The SCSI board 102 is
`the communications link between the scanner 20 and
`the host computer 130, as to be more fully described.
`The DSP board 104 preferably includes a Motorola
`56001 DSP digital signal processor (image processor)
`76. This processor performs a variety of operations in
`real-time on the pixel data coming from the sensor
`board 74,
`including gamma correction,
`light source
`compensation, and image resizing. The RAM is used as
`a buffer for the pixel data and for downloaded lookup
`
`65
`
`5,303,064
`
`10
`
`15
`
`20
`
`25
`
`4
`tables. The board 104 preferably includes circuitry
`which controls exposure timing.
`The motor board 106, more fully depicted in relation
`to FIG. 4, includes DC servo control circuitry for the
`carriage motor 202 (for film carriage 32), and control
`circuitry for the three stepper motors: filter motor 208
`(for filter assembly 70), focus motor 212 (for focus as-
`sembly 68), and zoom motor 216 (for zoom assembly
`66). The board 106 also has interrupt circuits for each of
`the four home position flag sensors (carriage 206, filter
`208, focus 212, and zoom 218). The alphanumeric front
`panel liquid crystal display 150 is also connected to
`circuitry on the motor board 106.
`The CPU board 108 preferably includes a 16 MHz,
`80386 processor, 8 MB of RAM, interrupt controller,
`DMA controller, and bus interface circuitry. The sys-
`tem software running on the CPU initiates and/or con-
`trols all activity in the scanner.
`Power supply 144 preferably provides a number of
`DC voltages (+ 5 vo1t,+/-12 volts, and +/-15 volts)
`used by the PC boards and motors. Lamp supply 144, as
`seen in FIG. 4, provides power for the fluorescent lamp
`(light source) 30. Preferably, a photodetector at
`the
`lamp 30 measures the light intensity and feeds a signal
`proportional to the intensity back to the lamp power
`supply 144. This feedback circuit is used to adjust the
`drive signal from the supply so that a constant light
`intensity is maintained.
`FIG. 4 provides, in block form, a depiction of the
`scanner electronics. Many of the elements have been
`previously discussed, for instance, the DSP board 104.
`The DSP board 104 is connected to sensor board 74.
`Preferably, the sensor board 74 includes a CCD sensor
`72 which includes 2592 sensors (photosites); a Fairchild
`CCD 181M-4 is sufficient for this purpose.
`DSP board 104 is also coupled to backplane 120
`which is coupled to CPU board 108. Backplane 120 may
`be any PC/AT compatible passive backplane, prefera-
`bly with six slots. Backplane 120 is also coupled to a
`host computer 130 through output module (SCSI
`board) 102. Preferably, the host computer 130 is manu-
`factured by Apple Computer, Cupertino, Calif., and the
`output module 102 is a SCSI interface, such as a NCR
`53C94 chip. Backplane 120 also interacts with serial
`board 100.
`
`Lamp supply 110 and power supply 144 are con-
`nected to input line power 146. Power supply 110 drives
`fan 140. Lamp power supply 144 is coupled to lamp
`(light source) 30.
`Backplane 120 is coupled to disk controller 124
`which is coupled to hard disk 112. Preferably, hard disk
`112 is a Quantum ProDrive 40 AT and disk controller
`124 is a Quantum ProDrive AT-Bus Adapter.
`Finally, backplane 120 is coupled to motor driver
`(motor board) 106 which controls a number of motor
`operations in the scanner 20. Motor board 106 controls
`carriage motor 202 which drives film carriage 32. Lin-
`ear encoder, or carriage encoder, 114 is also coupled to
`motor board 106. Carriage home flag 206 generates an
`interrupt signal when the film carriage 32 reaches a
`designated home position.
`More specifically, carriage motor 202 operates in
`conjunction with film carriage 32 in the following man-
`ner. The carriage 32 moves the transparency 33, posi-
`tioned in a film holder, past the light source 30 during a
`scan. The carriage 32 preferably moves along a thread-
`less leadscrew. A pulley belt connects the leadscrew to
`
`

`
`20
`
`5
`carriage motor 202 which is mounted below the film
`carriage 32. Preferably, linear encoder 114, in accor-
`dance with the prior art, produces two sine wave output
`signals which are 90° out of phase. The circuitry on the
`motor board 106 converts these sinusoidal signals to
`digital square waves with a 50% duty cycle. The resul-
`tant square waves go into a quadrature decoder circuit
`which serves as an 8-bit relative position counter for the
`carriage 32.
`Filter motor 208 drives filter assembly 70 which is
`capable of moving clear, red, green, and blue filters into
`the light path. The filter home flag 210 generates an
`interrupt signal when the filter reaches a designated
`home position.
`Focus motor 212 moves focus assembly 68 such that
`the distance from the lens assembly 64 to the sensor 72
`is adjusted. Focus motor 212 drives a pulley belt that is
`attached to a leadscrew which operates in conjunction
`with focus assembly 68. The focus home flag is acti-
`vated when the focus assembly 68 is positioned closest
`to the sensor 72.
`Zoom motor 216 controls the movement of zoom
`assembly 66. Specifically, the zoom motor 216 moves
`both the lens assembly 64 and sensor 72 in relation to
`film 33. Zoom motor 216 drives a pulley belt attached to
`a leadscrew.
`‘
`Motor board 106 is also coupled to front panel 150
`which may be an LCD which provides status informa-
`tion regarding the scanner 20.
`Turning now to FIG. 5, a block diagram is provided
`to demonstrate how the scanner 20 of the present inven-
`tion is utilized. Each of the three color planes of the
`image from the film 33 is conveyed through the optical
`assembly 60 and impinges upon sensor board 74. Sensor
`board 74 processes the data, as to be more fully de-
`scribed herein, and conveys the data, in digital form, to
`data signal processing board 104. The DSP board 104
`further processes the signal and conveys it to CPU
`board 108 which includes a RAM buffer capable of 40
`storing one color plane. Through disc controller board
`124, each color plane is stored on hard disk 112. Hard
`disk 112 also includes a number of programs which are
`called by the CPU board 108 and control the sensor
`board 74 and the data processing. After each color
`plane has been stored on hard disk 112, it is again con-
`veyed through disk controller board 124 to the CPU
`board 108 where the individual color planes are inter-
`leaved, pixel by pixel, to form full color image data.
`This data is then conveyed to SCSI board 102 which
`conveys it to host computer 130. Host computer 130
`may then manipulate the image data and integrate it
`with other data.
`Turning now to FIG. 6, disclosed therein is a block
`diagram of sensor board 74, also referred to herein as
`analog electronics. Preferably, Charge Coupled Device
`(CCD) 72 includes a clock input 302 and a power input
`304. One clock input 302 is coupled to clock 303 while
`another clock input is coupled to header 342 which
`serves as an interface between the sensor board 74 and
`the DSP board 104. The DSP board 104 generates an
`integration pulse which controls the exposure time.
`CCD 72 drives first data channel output (video A)
`signal 306 and second data channel output (video B)
`signal 308. The video A 306 signal represents the pixel
`value, or pixel response to light of one half of the pixels.
`That is, film 33 is divided into a plurality of pixel values,
`bytes of data, which are individually interpreted by
`
`6
`sensor board 74. The video B signal 308 represents the
`pixel value of the other half of the pixels.
`The remainder of the circuit 74 serves to equalize the
`two channels (video A 306, video B 308) in order to
`obtain corrected pixel valves. That is, by equalizing the
`two channels, a pixel value which is not affected by
`variables such as temperature drift and noise may be
`obtained. Thus, the resultant pixel (read at analog to
`digital converter 340) has an enhanced dynamic range
`in the image.
`Preferably, preamplifiers 310 boost and filter the re-
`spective signals. The signals are then conveyed to sam-
`ple-and-hold devices 312 which are controlled by con-
`trol means 313. The signals are then fed to differential
`amplifiers 314. Preferably, the inputs to the differential
`amplifiers 314 are modified by a coarse offset 316 and a
`fine offset 318. The coarse offset 316 is used to get a
`rough zero DC bias on both channels so that there are
`not any signals below ground. The fine offset 318 only
`feeds into the video B signal 308; it is used to balance the
`dark levels of the two channels. Both offsets preferably
`utilize an 8-bit digital to analog converter coupled to the
`DSP board 104. This feature enhances the digital con-
`trol of the analog signal. If the coarse offset 316 and fine
`offset 318 are correctly adjusted, then an approximately
`fiat signal at a zero DC offset will result when the sen-
`sors 72 are exposed to black.
`Preferably, the outputs from the differential amplifi-
`ers 314 are conveyed to A gain element 320 and B gain
`element 322. The gain elements 320 and 322 include
`8-bit digital to analog converters, again providing digi-
`tal control of the analog signal. The video signals serve
`as the reference voltages. The gain elements (320, 322)
`serve to balance the light end of thetwo channels. In
`other words, if the sensors 72 were exposed to the bare
`lightsource 30, there should be no perceptible differ-
`ence between the two signals (306, 308).
`From the gain devices, the signals are preferably fed
`to multiplexer 330. The single signal output then passes
`through amplifier 332 and into an analog-to-digital con-
`verter 340. The resultant digital signal is conveyed to
`header 342 and eventually to DSP board 104. Thus, the
`sensor board 74 includes digital control of the analog
`data. In addition, analog manipulation of the signals is
`minimized as the analog signals are quickly transformed
`to a digital signal.
`A preferable embodiment of a scanner 20 to be used
`in accordance with the present invention has now been
`provided. The calibration of this scanner for optimal
`performance will now be disclosed.
`invention is
`When the scanner 20 of the present
`turned on it proceeds through the same power-on boot
`sequence as any DOS computer. Initially, memory and
`attached devices are diagnostically checked; after-
`wards, the operating system is loaded into memory,
`finally the system software including calibration in-
`structions, is loaded into memory and begins execution.
`Preferably, the system software first initializes the out-
`put module 102 and the image processor 104. The soft-
`ware then directs the carriage motor 202, filter motor
`208, focus motor 212, and zoom motor 216 to their
`home positions and assumes a ready state, waiting for
`commands from the host computer 130 via output mod-
`ule 102.
`
`Typically, calibration of the scanner 20 begins at this
`point. The calibration of the scanner 20 may involve all
`of the steps described herein or only a few of the steps.
`
`5,303,064
`
`10
`
`15
`
`25
`
`30
`
`35
`
`.
`
`45
`
`50
`
`55
`
`65
`
`

`
`5,303,064
`
`7
`Preferably, the calibration of the scanner 20 is repeated
`for each resolution value utilized on the scanner 20.
`The steps of the calibration method 350 of the present
`invention are generally disclosed in relation to FIG. 7.
`Preferably,
`the optical assembly is initially focused
`(block 360). Afterwards, a nominal exposure time for
`white light is defined for preliminary calibration scans
`(block 362). Using this nominal exposure, offsets (coarse
`offset 316; fine offset 318) for the sensor board 74 are
`calculated (block 364) and then gains (A gain 320; B
`gain 322) for the sensor board 74 are calculated (block
`366). A saturation exposure level of the CCD 72 is then
`calculated (block 368). Using this saturation informa-
`tion, offsets and gains are recalculated (blocks 364 and
`366). The offsets and gains may be refined through
`iteration, if desired (block 370).
`After any desired iteration, an analogous method is
`followed using the different color filters of filter assem-
`bly 70. First, a color is set (block 371). Afterwards, a
`nominal offset (block 373) is established; this nominal
`offset may be the same as the calculated offset from step
`(364). Analogously, a nominal gain is established (block
`375). The saturation exposure for the given color is then
`established (block 368A). This method is identical to the
`method described in relation to FIG. 11, except the
`saturation of analog-to-digital converter 340 is being
`measured, not the saturation of the CCD 72. Calcula-
`tion of the offsets (block 364) and calculation of the
`gains (block 366) is then undertaken in light of the satu-
`ration exposure information. This process may be re-
`fined through iteration (decision block 370).
`After this iterative procedure, the next color on the
`filter assembly 70 (decision block 381) is invoked and
`the procedure is reinitiated by setting nominal offsets
`(block 373) for the new color. After all colors have been
`tested (decision block 371), a correction is calculated
`(block 372).
`In relation to the first step (block 360) of the calibra-
`tion sequence 350, the optical assembly 60 is focused in
`the following manner. A slide 33, preferably with hori-
`zontal black and white lines, is placed in carriage 32.
`Optical assembly 60 receives the signal from the slide
`33, then sensor board 74 and DSP board 104 interpret
`the signals. The sharpness of the pixel data is measured
`using prior art methods.
`—
`After preliminary readings, focus assembly 68 is repo-
`sitioned and new data is received by sensor board 74.
`This process is repeated until the largest difference or
`greatest sharpness between elements is obtained, indi-
`cating that the slide 33 is in focus.
`The next step in the calibration sequence 350 is to
`define a nominal exposure (block 362). The exposure
`time—the time the sensors 73 are exposed to light
`source 30 before CCD 300 is read—varies roughly be-
`tween 40 and 300 milliseconds. The host computer 130
`uses an arbitrary scale from 100 (minimum) to 500 (max-
`imum) to define the exposure range. A nominal expo-
`sure of 300 may be selected at this juncture.
`The procedure for calculating the offsets (block 364)
`is disclosed in relation to FIGS. 8 and 9. FIG. 8 dis-
`closes the method 364A for setting the coarse offset 316.
`The purpose of this procedure is to insure that the sen-
`sor board 74 is not clipping signals (video A 306; video
`B 308). FIG. 9 discloses the method 364B for setting the
`fine offset 318. The goal of this method is to find the
`closest balance between the video A 306 and video B
`308 channels.
`
`'
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`45
`
`50
`
`55
`
`8
`Turning to FIG. 8, light source 30 is initially blocked
`(block 401). Thereafter, scanning takes place (block
`404). Preferably, an averaging algorithm is utilized in
`accordance with all scanning described in relation to
`the calibration method of the present invention. Specifi-
`cally, a certain number of scans will be taken: 20, 100, or
`1()()O. Through CPU board 108,
`these scans will be
`accumulated and then divided by the number of scan
`lines to obtain an average pixel value. Through this
`averaging, random noise in the data is reduced.
`The scanned values are measured (block 406) to de-
`termine whether each pixel value is above a defined
`threshold. The threshold value is generally assigned to
`be zero. If clipping is present, the coarse offset 316 is
`incremented, as directed by CPU board 108, by approx-
`imately 50 mV (block 408). On the other hand, if no
`clipping is present, then the offset value is stored, via
`CPU board 108 (block 410).
`After setting the coarse offsets 316, the fine offsets
`318 are set, in accordance with FIG. 9. Again, the light
`is blocked (block 401). Scanning then takes place (block
`424) and a pixel balance value is computed and stored
`(block 426). A decision is then made (decision block
`428) to determine whether a predetermined range of
`offsets have been tested. If the range is not complete,
`the fine offset 318 is then incremented by approximately
`4 mV (block 430). If the desired range has been tested,
`then all the stored pixel balance values within the range
`are searched to determine the closest balance between
`signals (block 432). This value is then stored (block
`434).
`The method of calculating gains 366 is disclosed in
`reference to FIG. 10. Gain 320 and gain 322 are set
`independently by setting the other gain value to zero,
`thereby effectively turning off the other channel.
`Thereafter, light source 30 is unblocked (block 467) and
`scanning begins (block 470). The scanned data is ana-
`lyzed to determine whether every pixel is below a pre-
`defined saturation level (decision block 472). If every
`pixel is below saturation, the gain (video A gain 320 or
`video B gain 322) is incremented (block 474). If satura-
`tion is present, then the present value, or the previous
`non-saturated value is stored (block 476).
`The procedure for calculating saturation exposure
`368 is disclosed in relation to FIG. 11. This procedure
`serves to define a more accurate exposure time than that
`nominal value previously assigned (block 362). Sean-
`ning is initiated (block 502) and data is collected from
`sensor board 74 and interpreted by DSP board 104 to
`determine whether the CCD 72 is saturated (decision
`block 504). This saturation level is determined by the
`fiat response on the sensor board 74: conventional mea-
`sures must be taken to insure that one is not measuring
`the saturation of the analog to digital converter 340. If
`saturation has not been reached, then the exposure time
`is incremented (block 506) and scanning is repeated
`(block 502). This procedure is repeated until a satura-
`tion level is reached and the resultant saturation value,
`or previous value, is stored (block 508).
`The exposure time obtained in this method is used to
`calculate an actual exposure time to be used in further
`scans. However, if one were to use the actual number
`store