METROLOGY

CROSS-REFERENCES TO RELATED APPLICATIONS

[001] This application claims the benefit of U.S. Provisional Application No.

62/770,364, filed on November 21, 2018, which is incorporated herein by reference.

STATEMENT AS TO FEDERALLY SPONSORED RESEARCH

[002] This invention -was made with Government support under Agreement No.

HR00111790014, awarded by DARPA. The Government has certain rights in the invention.

BACKGROUND

[003] This invention relates to high-speed pharmaceutical quality control metrology.

[004] Optical coherence tomography (OCT) has been demonstrated as a way to characterize the surfaces of objects, for example, to use height information for the inspection of pharmaceutical products where the surface characteristics information may be used in a quality inspection process.

SUMMARY

[005] There is a need to perform high-speed metrology, for example, at speeds commensurate with mass quality inspections of pharmaceutical products (e.g., pills) traveling at a high velocity, for example, on a conveyor belt.

[006] In one aspect, in general, a method is directed to characterizing a continuously moving pharmaceutical manufacture via interferometry-based scanning. The method includes forming a depth characterization of the pharmaceutical manufacture along a scan line on the surface of the pharmaceutical manufacture. During the scanning, the pharmaceutical manufacture undergoes continuous motion. The method further includes combining the determined depth characterization along the scan lines of the plurality of scan lines to form a depth map representing at least a depth of a portion associated with a location on the surface of the pharmaceutical manufacture in the third direction on a grid of locations arranged in the first and second directions. Forming the depth characterizations includes scanning a frequency dispersed pulsed optical signal in a first direction across the continuously moving pharmaceutical manufacture, the

pharmaceutical manufacture moving in a second direction substantially orthogonal to the first direction. The scanned optical signal forms a scan line on a surface of the

pharmaceutical manufacture in a third direction substantially orthogonal to the first direction and the second direction.

[007] Aspects can include one or more of the following features:

[008] Forming the depth map includes a density as a function of position in the third direction as part of forming the depth characterization.

[009] Multiple scan lines may be used to form the depth characterization, each scan line having its own respective depth characterization. The depth characterizations of each of the scan lines are combined to form a depth map in the first and second directions across the pharmaceutical manufacture, the map representing depth in the third direction.

[010] The method may further include reflecting the pulsed optical signal off a rotating multifaceted mirror, then proceeding to pass the pulsed optical signal through a lens after the pulsed optical signal has been reflected off the rotating multifaceted mirror.

[Oil] The method may further include concurrently scanning the signal forming multiple scan lines on the object. Each scan line may have a scan line offset in the first direction from the other scan lines. Scanning the signal forming multiple scan lines may include splitting said signal into multiple signals, and scanning each of the multiple signals across the pharmaceutical manufacture by reflecting it off a moving mirror surface. Each scan line may have a corresponding moving mirror surface off of which to be reflected. Scanning the signal forming multiple scan lines may include passing each of the multiple signals through a corresponding lens of multiple lenses disposed in the first direction. At least some lens of the multiple lenses may be offset in the second direction from another of the lenses.

[012] The method may further include scanning the pulsed signal including passing the signal through a cylindrical lens with an axis extending substantially in the first direction.

[013] The method may include interpolating the depth characterizations along the scan lines to points on the grid of locations. One or more synchronization signals may be monitored, and interpolation may be performed according to said synchronization signals. Monitoring one or more synchronization signals may include monitoring a synchronization signal associated with at least one of a trigger signal from an optical source, a moving mirror surface, a transport of the object in the second direction, and a time offset corresponding to an interval of time between a detection of a physical calibration marker and a start time of a first scan line of the plurality of scan lines.

[014] The method may be configured to characterize multiple continuously moving pharmaceutical manufactures.

[015] The method may include the pharmaceutical manufacture being a pharmaceutical dosage form. The pharmaceutical dosage form may have a container layer and a pharmaceutical dosage layer. The pharmaceutical dosage form may further include a pill such as a tablet or a capsule. Tire tablet may further include a caplet. The capsule may further include an outer container component and an inner medicine region. The imier container component encloses the inner medicine region. The inner medicine region may' hold a liquid medicine or a solid medicine. The solid medicine may further include a compacted powdered medicine. The capsule may include a single-piece gel

encapsulation. The capsule may also include a two-piece gel encapsulation. The two- piece gel encapsulation may include a first encapsulation half and a second encapsulation half. The preferred material of the two-piece gel encapsulation is a cellulose material.

[016] The method may further include the container layer including a pharmaceutical packaging. The pharmaceutical packaging may be a blister pack.

[017] The method may further include the container layer being opaque to visible light.

[018] The method may further include the container layer being translucent to visible light.

[019] The method may further include characterizing the continuously moving pharmaceutical manufacture in the following way: the method determines the locations of one or more material transitions from the depth characterization. Each of the material transitions can be one of the following: an air-container transition, an air medicine- transition, and a container-medicine transition.

[020] In another aspect, in general, an apparatus is used for characterizing a

continuously moving pharmaceutical manufacture via interferometry-based scanning.

The apparatus includes a scanner configured to generate a plurality of scan lines from the moving pharmaceutical manufacture. The scanner includes an optical source configured to generate a frequency dispersed pulsed optical signal, and a rotating polygon mirror. The rotating polygon mirror is configured to direct the frequency dispersed pulsed optical signal in a first direction across a surface of the continuously moving pharmaceutical manufacture. The pharmaceutical manufacture moves in a second direction substantially orthogonal to the first direction, and the scanned frequency dispersed pulsed optical signal forms the plurality of scan lines on the surface of the pharmaceutical manufacture. The rotating polygon mirror is further configured to direct a reflection of the frequency dispersed pulsed optical signal from the surface of the pharmaceutical manufacture to an optical signal processor. The apparatus further includes an optical signal processor configured to process the reflection of the pulsed optical signal from the pharmaceutical manufacture for each scan line of the plurality of scan lines. A depth characterization of the pharmaceutical manufacture in a third direction along the scan line on the

pharmaceutical manufacture is determined based on this processing of the reflection. The third direction is substantially orthogonal to the first direction and the second direction. The optical signal processor is further configured to combine the determined depth characterizations along the scan lines of the plurality' of scan lines to form a depth map representing at least a depth of a portion associated with a location on the surface of the pharmaceutical manufacture in the third direction on a grid of locations arranged in the first and second directions.

[021] The apparatus may include a lens disposed between the object and the mirror such that the frequency pulsed optical signal passes through the lens. The lens may be a cylindrical lens with an axis extending substantially in the first direction. The apparatus may include an optical splitter configured to split the frequency pulsed optical signal into several frequency pulsed optical signals. Several rotating polygon mirrors may also be included. Each of the rotating polygon mirrors may be configured to direct one corresponding frequency pulsed optical signals in a first direction across a corresponding part of the surface of the continuously moving pharmaceutical manufacture. A number of lenses may be disposed between one corresponding rotating polygon mirrors and a corresponding part of the surface of the continuously moving pharmaceutical

manufacture.

[022] The apparatus may further include a fiber coupled pulsed laser configured to generate a pulsed optical signal, and a dispersal component configured to receive a pulsed optical signal and to transform the pulsed optical signal into a time stretched optical signal. The apparatus may further include an optical processing portion configured to generate an interference signal representing the depth characterization of the

pharmaceutical manufacture. The apparatus may also include a digital processing portion configured to receive the interference signal and to process the interference signal to generate the depth map.

[023] The apparatus may further use an interpolation operation to generate a mapping of the depth characterizations along the scan lines to points on the grid of locations when combining the determined depth characterizations. The apparatus may further be configured to scan a plurality of continuously moving pharmaceutical manufacture.

[024] Other features and advantages of the invention are apparent from the following description, and from the claims.

DESCRIPTION OF DRAWINGS

[025] FIG. 1 is a block diagram of a scanning system:

[026] FIG. 2. is a plot of spectral energy of an optical pulse;

[027] FIG. 3 is a plot of spectral energy of the optical pulse of FIG. 2 after frequency dispersal;

[028] FIG. 4 is a diagram of an optical source;

[029] FIG. 5 is a block diagram of a scanner; and

[030] FIG. 6 is an illustration of two approaches for generating alignment data.

[031] FIG. 7 is a cross-section view of a pharmaceutical object during the scanning process.

DESCRIPTION

1 SYSTEM OVERVIEW

[032] Referring to FIG. 1, a scanning system 100 is used to characterize a plurality ⁷ of moving pharmaceutical objects 160 by scanning them as they move continuously without requiring the objects to stop while being scanned. With regard to a single moving pharmaceutical object 160, the pharmaceutical object is moved linearly along a conveyor belt 150 or other transport means such that it passes under the optical signal 113 of an optical scanner 118. The optical scanner 118 ultimately then uses the information gathered from scanning the pharmaceutical object 160 to produce a volume characterization of the object. In some embodiments, this may be characterized as surface information, such as a“depth map” 126 for the object (e.g., a fimction z(x, y) where z is the thickness or depth and (x,y) is a coordination of the object in the object’s frame of reference, and/or a density map p(x,y, _å ) for the points within the object). With regard to the optical scanner 118, the optical scanner is configured to form a series of scan lines 152 for example to scan many pharmaceutical objects on various places on the conveyor belt 150.

[033] Referring to FIG. 7, a cross-sectional view of an example pharmaceutical object 160 is a capsule, which includes an outer container part 161 enclosing an inner part, which includes air/gas space 162, a liquid 163 (e.g., a liquid medicine) and/or a solid medicine 164. It should be understood that by determining a density map p(x,y,z) for an object such as this example, the volume of the solid or liquid medicine may be determined by classifying which points are within the liquid or the solid, and

accumulating the volumes based on the classifications. Furthermore, surface

characteristics such as text or symbol markings on a tablet may be“read” from the output of the scanner. There are of course other forms of pharmaceutical object that may be scanned and analyzed using the techniques described in this document, including for example, coated or uncoated tablets, or glass or plastic containers holding liquid.

Furthermore, the objects may be packaged when they are scanned, for example, in blister packs.

[034] The sample optical signal 113 is formed using an optical source 102. The source includes a broadband laser 101, which emits optical pulses and a dispersal element 103, which processes the pulses emitted from the laser 101 such that different

wavelengths are delayed by different amounts of time yielding frequency-dispersed time- stretched optical pulses.

[035] The system includes an interferometry subsystem 112, which receives the frequency dispersed optical pulses from the optical source 102, and splits that signal into the sample optical signal 113, which is passed to the scanner 118, and a reference optical signal 105, which is delayed by a delay component 106, and combined with the received sensed signal 115 from the scanner to form an output signal 116 from the interferometry subsystem 112. This signal is passed to a signal processor 124.

[036] Tire signal processor 124 processes the optical signal 116 formed by the interferometry subsystem for each of the scan lines 152 to determine a depth characterization of each pharmaceutical object 150 along each scan line. For each pharmaceutical object scanned in this manner, the optical processor combines the determined depth characterizations along multiple scan lines to form the depth map 126 representing at least a surface height z on an area across x - v plane of the object, for example on a regular grid of locations arranged in the x and v directions, and/or a density map p(x,y,z) for the points within the object. Using the depth map or density map, the system is able to determine the material composition and/or structure of the pharmaceutical object.

[037] Continuing to refer to FIG. 1 , the conveyor belt 150 is illustrated as moving with a velocity v in the positive x' direction. If one scan along a scan line takes t time, then the object will have been displaced by vx r during the scan time. Therefore, although the scanning is along the fixed y’ direction (i.e., the beam 117 is scanned in the y’-z' plane) the object is sensed at locations with slightly varying x locations, exaggerated for illustration in FIG. 1.

[038] Tire signal processor 124 receives an indication of when each pulse is emitted from the optical source 102, an indication of when each scan caused by free-running rotation of the hexagonal mirror 114 begins to cross the object, and an indication of the x location of the object during its motion (or equivalently an indication of the start of the motion and the velocity, from which the x location may be derived). From this information, as described in more detail below, at any time during a scan, the optical signal processor 124 has sufficient information to determine the wavelength of the sample signal 113 (which is based on a time offset from the start of the pulse), from which is can determine the distance from the scanner to the point on the surface of the object from which the scanned emission occurs, and from that, the (x,y, _å ) coordinates of that point in the fixed frame of reference of the moving object.

[039] The signal processor 124 in this embodiment has two parts: a depth characterizer 120 and a depth mapper 122. The depth characterizer 120 serves to produce a depth characterization along a single scan line of the 3D object after receiving the reflected optical signal associated with that scan line. The depth mapper 122 serves to produce a depth map characterizing the entire scanned surface of the 3D object after receiving a set of depth characterizations, where each scan line has an associated depth characterization in the set of depth characterizations. The depth mapper, after receiving all depth characterizations for the scan lines associated with the 3D object, analyzes them with respect to the motion information to create a depth map characterizing the surface of the 3D object.

2 OPTICAL SOURCE

[040] As introduced above with reference to FIG. 1, the optical source 102 is configured to generate frequency-dispersed time-stretched optical pulses. In this embodiment, light from a fiber-coupled pulsed broadband laser 101 is directed to an optical component 103 with high group velocity dispersion such that wavelengths in the original pulse are mapped to time in a temporally broadened pulse. A variety of fiber- coupled pulsed lasers with a pulse length on the femto or picosecond time scale can be used as long as its bandwidth is wide enough to give the desired depth resolution and noise is sufficiently low. Examples of such choices are a Ti-Sapphire laser, a mode- locked fiber laser, a pulsed diode laser or a pulsed supercontinuum laser.

[041] In some embodiments, the optical component 103 includes an optical circulator that directs the laser pulse to a chirped fiber Bragg grating with a linear dispersion D in typical units of ps/nm. The time stretched pulse reflected from the grating is directed by the circulator into a fiber interferometer that is part of the interferometry subsystem 112 as described below. In an alternative embodiment, the optical component 103 may be implemented by transmitting the broadband laser pulse through a single mode fiber of sufficient length L in km with D = D _c {X)L where the group velocity dispersion coefficient D _c has typical units of ps/nm*km.

[042] FIG. 2 shows a typical laser pulse with power as a function of time and wavelength. FIG. 3 shows the same laser pulse after time stretching.

[043] Each laser pulse is stretched to a period which is shorter than the time between laser pulses. Therefore, for each laser pulse there is a period of time while an interference signal is generated at the output of the interferometry subsystem 112 and a period of time while there is no useful data. As discussed below the signal processor 124 selects the useful parts of the signal based on synchronization information provided to it by the optical source. In this embodiment, a synchronization signal is emitted from the pulsed laser, and in particular, an electronic pulse is emitted synchronously with the laser signal and passed to the signal processor 124.

[044] The time-stretched pulse may be optically amplified in the optical source 102 and/or after first entering the interferometry subsystem 112 if the dispersive component 103 causes significant attenuation such that the signal to noise at the detector is not high enough to make accurate measurements. For example, a Ytterbium-doped fiber amplifier may be used.

[045] After stretching, the wavenumber ( 2p / l ) as function of time is given by

where A _Q is the center wavelength of the pulse and t ₀ is its time position in the stretched pulse.

[046] Referring to FIG. 4, the dispersive component 103 may be implemente using a non-reciprocating, one-directional, three port device 410, with the input from the laser source 101 entering a first port and the frequency-dispersed pulse exiting a third port. The second port is coupled to a chirped fiber Bragg grating 420, which spreads out the pulses as a function of wavelength.

3 INTERFEROMETRY SUBSYSTEM

[047] Referring again to FIG. 1 , the optical signal produced by the optical source 102 is received as input for the interferometry subsystem 112. As introduced above, in operation, it is the function of the interferometry subsystem 112 to split the optical signal into two separate beams, then create an interference signal using those two beams after they have been passed through subsequent subsystems. Specifically, the interferometry subsystem 112 receives an optical signal in its optical fiber interferometer 104, which selves as a splitter component splitting the optical signal into a sample optical signal 113 to be received by the scanner 118 and a reference signal 105, which is combined with sensed signal 115 from the scanner. The sample signal 113 and reference signal 115 are split with a 90: 10 power ratio to account for attenuation in scanning the objet.

[048] More specifically, at the interferometry subsystem 112 each time-stretched pulse output from the optical source 102 is directed into a fiber interferometer 104 that splits light into reference and sample signals passing via a reference arm and a sample arm, respectively. The reference arm includes a variable delay component 106 and an optional variable attenuator (not illustrated in FIG. 1). As described further below, the light of the sample arm is directed via the rotating mirror 114 from the sample arm fiber to a lens 116 that focuses the light onto the pharmaceutical objects 160 at the point to be measured. Light is specularly reflected and/or scattered by the object back through the lens 116 to the mirror 114 and is coupled back into the sample arm fiber. The light from the sample and reference arms is recombined in the interferometer. If the optical path difference between the reference and sample arms is shorter than the coherence length of the laser, then an interference signal will be generated and emitted from the

interferometry subsystem.

4 SCANNER

[049] Further referring to FIG. 1, the scanner 118 receives the sample optical signal 113 from the interferometry subsystem 112 and uses it to scan the surface of the moving 3D object 150, then provides sensed signal 115 back to the interferometry subsystem 112. The seamier consists of two main parts: a rotating polygon mirror 114, and a lens 116. In operation, upon receiving the sample optical signal, the scanner reflects the signal off the rotating polygon mirror 114.

[050] Referring to FIG. 5, the scanner 118 includes a circulator 510, a collimator 520, and the rotating mirror 1 14 and lens 1 16, previously introduced. As shown in FIG. 5, the sample signal 113 and the sensed signal 115 travel in opposite directions on light path from the circulator 510 to the sample, and back to the circulator 520. The sample signal 113 enters port lof the circulator 520, and the sensed signal 115 exits port 3 of the circulator. The collimator 520 causes the sample signal 113 to become more aligned in the direction of the mirror (i.e., make collimated light or parallel rays) and/or to cause the spatial cross section of the beam to become smaller.

[051] Tire collimated light reflects off the rotating polygon mirror 114 and is swept across the back of a telecentric scan lens 116 such that a focused laser spot repeatedly and uni directionally moves along a scan line 152 (as shown in FIG. 1) oriented substantially orthogonal to the direction of motion of the object. The same face of the mirror reflects the sample signal toward the object and the sensed signal from the object, and both signals pass through the collimator 520.

[052] In this embodiment, the rotating polygon mirror 114 rotates in an open loop manner (i.e., without feedback positional control) such that there is no ability required to control where the reflected laser spot is located at any point in time. If a polygon mirror rotates without any feedback as to its position, then there is not enough information to know the location of the scanner laser spot on the sample relative to the angular position of the active mirror facet. Howe ver, the scan line data from each facet of the mirror must be aligned relative to each other in the signal processor 124 to generate the depth map 126 for the object 150. Two alternative ways to generate this alignment data may be used, both of which are illustrated together in FIG. 6

[053] In each of two alternative embodiments, a start of scan sensor 620 or 630, respectively, is used to generate a regular electronic pulse once per mirror facet passing at a regular time interval. In a first embodiment, the start of a scan signal is generated by a second laser 610 that reflects off a polygon facet that is a fixed number of facets from the facet currently directing the scan laser to the part being measured, in FIG. 6 two facets away. The second laser beam is directed to a light detector 620 once per facet passing whenever a facet is oriented at the correct angle to reflect the beam toward the detector.

In another embodiment, if a portion of a scan lens 116 does not overlap wdth the object being scanned then a detector 630 can be placed underneath the scan lens at a location that will not occlude the scan laser from reaching the object. This allows a start-of-scan signal to be generated by the sample signal 113 itself.

[054] Successive scan lines 152 shown in FIG. 1 result from reflections of successive faces of the polygonal mirror 114. That is, if there are N faces, then every N ^th scan line is from the same face of the mirror. In the ideal case where the full field of the lens exactly matches the full width of the part and the mirror has no dead zones between facets, each facet of the polygon sweeps the beam in the same direction across the entire part in the y' direction. For a mirror with N facets and speed s in revolutions per minutes, the single facet transit time is given by t = 60 / Ns seconds. As discussed above, the object moves a distance Ax = TV at velocity v during each such scan.

[055] In a third alternative embodiment, an encoder on the mirror measures the angular position of the mirror. For a mirror with N facets the laser pulses once per scan line corresponding to a change in angle of 360 / N degrees. As a calibration process, it is also necessary to compute the calibration time offset ST . This can be done manually (e.g., by viewing the resulting depth map 126 and selecting the appropriate shift in the y direction). Alternatively, it is possible to perform this by using a physical calibration marker, e.g., a depth change or strong reflectance change. For example, the calibration marker can be located at the beginning of the scan sweep on the platform. The signal from the calibration marker will be detected by processing the depth data. The number of laser pulses between the detection of the reference edge and the start of scan signal gives a temporal reference ST corresponding to the relative physical location of the scan laser at the time at which the start-of-scan signal is received. This allows the start-of-scan signal to be used to generate an absolute physical reference for the scan laser pulses. This procedure can be done once as a pre-calibration and can be manually tuned as necessary.

[056] In order to sense the surface of the object in range of the scanning optics, one approach is to keep the relative height of the scanning optics of the scanner 118 above the scamied object constant as the height of the part increases during fabrication. Either the platform holding the object can move down away from the scanning optics as the object is fabricated and more material is deposited or the scanning optics can move up away from the part as its thickness increases.

[057] Given that the object and scanner are moving in relationship to each other along the x axis, an additional synchronization signal is used by the signal processor 124 in order to correlate a given signal sample to a particular x position. In some

embodiments, this signal is generated by a high-resolution encoder associated with the x axis, for example, on a moving build platform for the moving object being fabricated. Equivalently, in embodiments in which the object is stationary and the scanner moves, this encoder would represent the motion of the scanner.

[058] For an object moving with velocity v and a sufficiently small spot, the scan resolution in the direction of motion is given by the distance between scan lines,

Ax = v x t . The distance between measured spots in the y direction is a function of the laser pulse repetition rate p per second, then the number of pulses and therefore spots across one scan line is r p , and the spacing of the spots is Ay = Y/ p , where Y is the length of the scan line in the y direction. Note that the temporal length of the stretched pulse must be less than the time between laser pulses, U p .

[059] Because the object moves past the scanner 118 in the x direction there is no limit imposed by the optics on the maximum length of a part in that dimension. However, the practical length of the scan lines 152 may be limited by the size of the lens 116.

Arrangements with multiple scan beams, which may increase the width limit are described later in this description.

5 SIGNAL PROCESSOR

[060] As introduced above, the output of the interferometry subsystem 112 is a combination of a fixed delayed version of the reference signal, and variable delayed version of the sample signal, where the variable delay is a function of the travel path from the scanner to the object and back and any fixed delay in the optical components of the scanner 118. Very generally, if the component of the output corresponding to the reference signal is in phase with the component corresponding to the sensor signal then a detector at the input to the signal processing subsystem 124 will sense a high intensity, while if they are out of phase, the detector will sense a low intensity. Because the wavelength of the si gnals varies during each pulse, the intensity will var during each pulse as well. This variation in intensity during each pulse is used by the signal processor 124 to determine the depth of each spot on each scan line on the object.

[061] More precisely, the temporally modulated photocurrent from a detector corresponding to a single reflective sample height can be written as follows:

where:

p is the detector responsivity

T _r and I _s are the absolute values of the optical power transmission of the reference arm and the sample arm between the input to the interferometer and the detector

P is the power spectrum of the stretched laser pulse after optional optical amplification re ^]t(> is the complex reflectance of the sample

2 _å is the round trip optical path difference between the reference and sample arms k is the wavenumber which varies in time as a function of the dispersion element

[062] The detector responsivity, spectral pow'er, system transmission and sample reflectivity are all originally functions of wavelength but can be converted to functions of time through the relationship between wavenumber and time.

[063] The signal from the detector is continuously streamed to an analog to digital converter (ADC). If the dispersion element is not linear with respect to wave number then the interference signal is renormalized to be linear with respect to wavenumber. The digitized interference signal is converted to depth information by taking the inverse discrete Fourier transform of the signal. A peak in the transform provides the depth information at the point. The synchronization signals received from the optical source 102, representing the start time of each pulse, from the scanner 118 representing the start of each scan line, and from the platform representing the x axis position of the object, are used to determine an (x,y, _å ) coordinate for each spot on a scan line. These locations are then interpolated onto a regular grid in the - y plane, to form the output depth map 126.

6 ALTERNATIVES

[064] A number of factors may limit the width (i.e., the y axis dimension) of the object being scanned. One limit may be the size of the lens 116. For example, because there is necessarily some physical distance between the rotating mirror and the lens, the lens must have sufficient usable width be able to scan the entire part. Another limit may be the maximum rate of pulsing, which limits the minimum separation of the spots that are sensed on the surface. One alternative is to use a cylindrical lens extending in the y ' direction to extend the usable range.

[065] As illustrated in FIG. 7, another general class of alternatives uses concurrent scamiing on multiple scan lines across the object. In this figure, each scan line 152A is scanned concurrently with one scan line 152B, for example, with both scans being in the same y '- z ' plane of the scanner frame of reference. Referring to FIG. 8, one approach to such concurrent scanning is to split the output of the optical source 102 in an equal splitter 702, and to pass that signal to a first interferometer subsystem 112A and corresponding scanner 118 A, and to a second interferometer subsystem 112B and corresponding scanner 118B. The outputs of the two interferometer subsystems 112A and 112B are passed to a signal processor 724, which independently processes each of the signals as in the signal processor 124, but then prior to interpolation maps each set of spots onto a common grid. As illustrated in the figure, the scan lines 152A and 152B overlap in the x direction, which permits a signal processor to“stitch” together the depth maps (or volumetric scans) from the different scanner subsystems to form a consistent depth map across the entire object. Of course, more than two seamier subsystems may be used to increase the scannable width of the object.

[066] The signal processor 724 takes into account synchronization signals from each of the scanner subsystems, whose mirrors are not necessari ly synchronized, and therefore the start of each scan line 152A and 152B may not be synchronized. The signal processor 724 determines a registration of the two sets of scan lines, for example, using overlapped regions of the scanned object, or using a calibration phase before the fabrication of the object begins.

[067] In the example above, each of the scanners 118A-B and their corresponding lens 116A-B are arranged in line along the fixed y' axis. Referring to FIG. 8, to increase the number of lenses that can fit along the y' the lenses may be offset in the x ' direction such that lenses 116A and 116C are on the same x' point, and lens 116B is offset in the x' direction. The operation of the system is identical to the multiple scanner approach described above, with the signal processor 724 taking into account this x' offset when interpolating the depth information at spots on the scan lines to form the depth map on a regular grid.

[068] In a similar manner to using multiple scanner subsystems 118 to increase scannable width, multiple scanner subsystems 118 can be used to extend the range in depth in the _å direction, for example, with different lenses focusing at different depths, for example, mounting of lenses 116A-C at different offsets in the z direction, but having their scan range overlap fully in the y direction. The signal processor 724 then combines the signals from the different scanner subsystems, for example, according to the signal to noise ratio of the different interferometer output signals.

[069] Although the discussion above focusses on computing a depth map of the object, other outputs may be derived. For example, density information as a function of three dimensions may be obtained by combining the information from scans in multiple passes of the object under the scanning system.

[070] It should be understood that although described in the context of scanning during three-dimensional fabrication, the same scanning approach may be used for other tasks that require continuous monitoring. For example, manufactured items passing on a conveyor belt under the seamier can be sensed, for example, for quality-control applications.

[071] Embodiments of the signal processor (e.g., 124, 724) may make use of hardware (digital and/or optical hardware) as well as software. The digital hardware may include application-specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) and the like, to process the output of a digital-to-analog converter that processes the output of a photo-detector at the input of the signal processor (or pair of

photodetectors configured in a differential input mode). The software can include instructions for causing a processor to implement data processing operations, for example, to implement the Fourier Transform calculation of the signal processor.

Additionally, embodiments of the system may make use of a controller, which also may be implemented in hardware, software, or a combination of hardware and software, which coordinates operation of the light source, scanner subsystem, signal processor, and/or built platform in order to sy nchronize the operation of the subsystems to yiel the depth map for the object in the object frame of reference.

[072] It is to be understood that the description is intended to illustrate and not to limit the scope of the invention, which includes the scope of the appended claims. Other embodiments than those described above are within the scope of the following claims.