WO2009078002A1

WO2009078002A1 - 3d camera and methods of gating thereof

Info

Publication number: WO2009078002A1
Application number: PCT/IL2007/001571
Authority: WO
Inventors: Giora Yahav; Gil Zigelman; Allan C. Entis
Original assignee: Microsoft International Holdings B.V.
Priority date: 2007-12-19
Filing date: 2007-12-19
Publication date: 2009-06-25
Also published as: EP2235563A1; CN102099703A

Abstract

A method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse having a pulse width at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a plurality of sets of gates, each set comprising at least one first, second and third gate having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start and stop times of the at least one first gate and the at least one second gate are between the start and stop times of the at least one third gate, wherein for at least one of the set of gates, the at least one third gate has a start time equal to about the stop time of the at least one third gate of another of the set of gates; and using amounts of reflected light imaged on the photosurface to determine distances to the scene.

Description

3D CAMERA AND METHODS OF GATING THEREOF

FIELD

The invention relates to methods and apparatus for acquiring 3D images of a scene. BACKGROUND

Three-dimensional (3D) optical imaging systems, hereinafter referred to as "3D cameras", that are capable of providing distance measurements to objects and points on objects that they image, are used for many different applications. Among these applications are profile inspections of manufactured goods, CAD verification, robot vision, geographic surveying and imaging objects selectively as a function of distance.

Some 3D cameras provide simultaneous measurements to substantially all points of objects in a scene they image. Generally, these 3D cameras comprise a light source, such as a laser, which is pulsed or shuttered so that it provides pulses of light for illuminating a scene being imaged and a gated imaging system for imaging light from the light pulses that is reflected from objects in the scene. The gated imaging system comprises a camera having a photosensitive surface, hereinafter referred to as a "photosurface", such as a CCD or CMOS photosurface and a gating means for gating the camera open and closed, such as an optical shutter or a gated image intensifϊer. The reflected light is registered on pixels of the photosurface of the camera only if it reaches the camera when the camera is gated open. To image a scene and determine distances from the camera to objects in the scene, the scene is generally illuminated with a train of light pulses radiated from the light source. For each radiated light pulse in the train, following an accurately determined delay from the time that the light pulse is radiated, the camera is gated open for a period of time hereinafter referred to as a "gate". Light from the light pulse that is reflected from an object in the scene is imaged on the photosurface of the camera if it reaches the camera during the gate. Since the time elapsed between radiating a light pulse and the gate that follows it is known, the time it took imaged light to travel from the light source to the reflecting object in the scene and back to the camera is known. The time elapsed is used to determine the distance to the object.

In some "gated" 3D cameras, only the timing between light pulses and gates is used to determine distance from the 3D camera to a point in the scene imaged on a pixel of the photosurface of the camera. In others, an amount of light registered by the pixel during the time that the camera is gated open is also used to determine the distance. The accuracy of measurements made with these 3D cameras is a function of the rise and fall times of the light pulses and their flatness and, how fast the cameras can be gated open and closed. Gated 3D cameras and examples of their uses are found in European Patent EP 1214609 and in US Patents 6,057,909, US 6,091,905, US 6,100,517 and US 6,445,884, the disclosures of which are incorporated herein by reference. The cameras described in these patents use amounts of light registered by pixels in the camera during times at which the camera is gated open to determine distances to features in contiguous slices of a scene. The slice locations and spatial widths are defined by length and timing of gates during which the cameras are gated open relative to pulse lengths and timing of light pulses that are transmitted to illuminate the scene.

Generally, at least two gates, at least one relatively long gate and at least one relatively short, "front" or "back" gate temporally corresponding to the front part or back part respectively of the at least one long gate, are used to acquire a 3D image of a slice of the scene. Relative to a time at which a light pulse is transmitted to illuminate the scene, a front gate starts at a same time that a long gate begins and a back gate ends at a same time as a long gate ends. The short gates optionally have a gate width equal to a pulse width of pulses of light used to illuminate the scene and a long gate has a pulse width equal to twice the light pulse width. Amounts of light registered during the at least one short gate by a pixel in the camera that images a feature of the scene are normalized to amounts of light registered by the pixel during the at least one long gate. The normalized amounts of registered light are used to determine a distance to the feature. A total acquisition time for acquiring data for a 3D image of a scene is substantially equal to a number of slices of the scene that are imaged, times a time, a "slice acquisition time", required to acquire 3D data for a single slice. A slice acquisition time is a function of a number of light pulses and gates required to register quantities of light for the various gates sufficient to provide data for determining distances to features of the scene located in the slice. A 3D camera using a pulsed source of illumination and a gated imaging system is described in "Design and Development of a Multi-detecting two Dimensional Ranging Sensor", Measurement Science and Technology 6 (September 1995), pages 1301-1308, by S. Christie, et al., and in "Range-gated Imaging for Near Field Target Identification", Yates et al, SPIE Vol. 2869, p374 - 385 which are herein incorporated by reference. Another 3D camera is described in U.S. patent 5,081,530 to Medina, which is incorporated herein by reference. A 3D camera described in this patent registers energy in a pulse of light reflected from a target that reaches the camera's imaging system during each gate of a pair of gates. Distance to a target is determined from the ratio of the difference between the amounts of energy registered during each of the two gates to the sum of the amounts of energy registered during each of the two gates.

SUMMARY

An aspect of some embodiments of the invention relates to providing a gating procedure for gating a gated 3D camera that provides a relatively short acquisition time for a 3D image of a scene.

An aspect of some embodiments of the invention relates to providing a configuration of gates for providing a 3D image of a scene for which, for a given width of a light pulse that illuminates the scene the gates provide 3D data for slices of the scene that have relatively large spatial widths.

There is therefore provided in accordance with an embodiment of the invention, a method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a plurality of sets of gates, each set comprising at least one first, second and third gate having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start and stop times of the at least one first gate and the at least one second gate are between the start and stop times of the at least one third gate, wherein for at least one of the set of gates, the at least one third gate has a start time equal to about the stop time of the at least one third gate of another of the set of gates; and using amounts of reflected light imaged on the photosurface to determine distances to the scene.

Optionally, the start time of the second gate is substantially equal to the stop time of the first gate. Optionally, the start time of the first gate is delayed relative to the start time of the third gate by a pulse width of the at least one light pulse. Optionally, the first and second gates have equal gate widths. Optionally, the gate width of the first and second gates is equal to half a pulse width of the at least one light pulse. Optionally, the gate width of the third gate is substantially equal to three pulse widths of the at least one light pulse.

In some embodiments of the invention, if amounts of reflected light from a feature imaged during the at least one first, second and third gates of a set of gates are denoted by Qj, Q2_> Q3 respectively and the start time of the at least one first gate by t\, then if Q\ ≠ 0 or Q2 ≠ 0, distance Df to the feature is determined in accordance with Df= ct!/2 + (CAc)KQ₁+ Q₂)/Q₃-l]/2, if Qi(n) < Q₂(Ii) and Df= ctχ/2 + (CAT)KI-(Q₁+ Q₂)/Q₃]/2 if Qi(n) > Q₂(Ii), where c is the speed of light and Δτ is the pulse width of the at least one light pulse. Additionally or alternatively, if amounts of reflected light from a feature imaged during the at least one first, second and third gates of a set of gates are denoted by Qi, Q₂, Q3 respectively, the start time of the at least one third gate by t^, and the amount of light from the feature imaged by the at least one third gate of the other set of gates by Q*3 then if Ql = Q2 = 0, Q3 ≠ 0 and Q 3 ≠ 0, distance Df to the feature is determined in accordance with Df= ct₃/2 - (cΔτp_W)(Q3(n)/(Q₃+Q*₃))/2 or Df= ct₃/2 - (cΔτ_/7W)(l-(Q*3)/(Q₃+Q*3))/2, where c is the speed of light and Δτ is the pulse width of the at least one light pulse. In some embodiments of the invention, the at least one third gate has a start time earlier than the stop time of the at least one third gate of the other set of gates by a period that is less than or equal to about one twentieth of the pulse width of the at least one light pulse.

There is further provided in accordance with an embodiment of the invention, a method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a plurality of equal length gates having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start time of at least one first gate is substantially a time half way between the start and stop times of at least one second gate; and using amounts of reflected light imaged on the photosurface during the at least one first gate and the at least one second gate to determine distance to a feature of the scene.

Optionally, the gates have a gate width substantially equal to twice a pulse width of the at least one light pulse.

Additionally or alternatively, if the start time and amounts of reflected light from the feature imaged during the at least one first gate are denoted respectively by X\, and Qj and for the at least one second gate by t2 and Q2 respectively, distance Df to the feature is determined in accordance with

Df= (ct_n/2) + cΔτ(Q2/Qi)/2 if Q2 < Qi and Df= (ct₂/2) + cΔτ(l-(Qi/Q₂))/2 if Qi < Q₂, where c is the speed of light and Δτ is the pulse width of the at least one light pulse.

There is further provided in accordance with an embodiment of the invention, a method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse having a pulse width at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a set of gates comprising at least one first, second and third gates having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start and stop times of each first gate and each second gate are between the start and stop times of a same third gate; and using amounts of reflected light imaged on the photosurface to determine distances to the scene.

Optionally, the first gate has a start time delayed relative to a start time of the third gate by about the pulse width. Additionally or alternatively, the second gate has a stop time that precedes a stop time of the at least one third gate by about a pulse width. In some embodiments of the invention, the first gate has a gate width equal to about half a pulse width. In some embodiments of the invention, the second gate has a gate width equal to about half a pulse width. In some embodiments of the invention, the third gate has a gate width equal to about three pulse widths.

Optionally, the first and second gates have gate widths equal to half a pulse width and the third gate has a gate width equal to about three pulse widths wherein, and if amounts of reflected light from a feature imaged during the at least one first, second and third gates are denoted by Qi, Q2, Q3 respectively, and the start time of the first gate by ti, distance Df to the feature is determined in accordance with Df= cti/2 + (cΔτ)(Q/Q3-l)/2 if Q₂ < Qi; and Df = cti/2 + (cΔτ)(l-Q/Q₃)/2 If Q₁ > Q₂, where c is the speed of light, Δτ is the pulse width of the at least one light pulse and

Q = (Qi + Q2).

There is further provided in accordance with an embodiment of the invention, a camera useable to acquire a 3D image of a scene comprising: a gateable photosurface; and a controller that gates the photosurface in accordance with an embodiment of the invention. Optionally, the camera comprises a light source controllable to illuminate the scene with a pulse of light.

BRIEF DESCRIPTION OF FIGURES

Non-limiting examples of embodiments of the invention are described belowi with reference to figures attached hereto that are listed following this paragraph. Identical structures, elements or parts that appear in more than one figure are generally labeled with a same numeral in all the figures in which they appear. Dimensions of components and features shown in the figures are chosen for convenience and clarity of presentation and are not necessarily shown to scale. Fig. 1 schematically illustrates a gated 3D camera being used to acquire a 3D image of a scene;

Fig. 2 shows a time-distance graph that illustrates a temporal configuration of gates of the camera and at least one light pulse that illuminates the scene shown in Fig. 1 used to acquire a 3D image of the scene, in accordance with prior art;

Fig. 3 shows a time-distance graph that illustrates an ambiguity in determining distance to a feature in a scene, in accordance with prior art;

Fig 4 shows a time-distance graph that illustrates a temporal configuration of gates for removing the ambiguity illustrated in Fig. 3, in accordance with an embodiment of the invention;

Fig. 5 schematically shows a scene and slices of the scene for which a gated camera provides 3D data;

Fig. 6 shows a time-distance graph that graphically illustrates timing of gates used to acquire 3D data for the plurality of slices of the scene shown in Fig. 5 using the gating configuration illustrated in Fig. 4, in accordance with an embodiment of the invention;

Fig. 7 shows a time-distance graph that graphically illustrates timing of gates used to acquire 3D data for a plurality of slices of a scene, in accordance with another embodiment of the invention;

Fig. 8 shows a time-distance graph that graphically illustrates another timing configuration of gates used to acquire 3D data for a scene, in accordance with prior art;

Fig. 9 shows a time-distance graph that graphically illustrates another timing configuration of gates used to acquire 3D data for a scene, in accordance with prior art;

Fig. 10 shows a time-distance graph that graphically illustrates timing of gates configured as shown Fig. 9 used to acquire 3D data for a plurality of slices of a scene, in accordance with prior art; and

Fig. 11 shows a time-distance graph that graphically illustrates timing of gates used to acquire 3D data for a plurality of slices of a scene, in accordance with another embodiment of the invention.

DETAILED DESCRIPTION

Fig. 1 schematically illustrates a gated 3D camera 20 being used to acquire a 3D image of a scene 30 having objects schematically represented by objects 31 and 32.

Camera 20, which is represented very schematically, comprises a lens system, represented by a lens 21, and a photosurface 22 having pixels 23 on which the lens system images the scene. Photosurface 22 is "gateable" so that it may be selectively gated on or off to be made sensitive or insensitive respectively to light for desired periods. Optionally, pixels in the photosurface are independently gateable. Photosurfaces that are gated by shutters are described in US Patents 6,057,909, 6,327,073, 6,331,911 and 6,794,628, the disclosures of which are incorporated herein by reference. Photosurfaces having independently gateable pixels are described in PCT Publication WO 00/36372 describes photosurfaces having pixels that are independently gateable. A "gate" refers to a period during which a photosurface or pixel in the photosurface is gated on and made sensitive to light. For convenience of presentation, it is assumed that all the pixels 23 in photosurface 22 are gated on or off simultaneously. Camera 20 optionally comprises a suitable light source 26, such as for example, a laser or a LED or an array of lasers and/or LEDs, controllable to illuminate scene 30 with pulses of light. A controller 24 controls pulsing of light source 26 and gating of photosurface 22.

To acquire a 3D image of scene 30, controller 24 controls light source 26 to emit at least one light pulse, schematically represented by a wavy arrow 40, to illuminate scene 30. Light from each at least one light pulse 40 is reflected by features in scene 30 and some of the reflected light is incident on camera 20 and collected by lens 21. Reflected light from at least one light pulse 40 that reaches camera 20 is schematically represented by wavy arrows 45. Following emission of each at least one light pulse 40, controller 24 gates photosurface 22 on at a suitable time relative to a time at which the light pulse is emitted to receive and image reflected light 45 collected by lens 21 on photosurface 22. Amounts of light 45 imaged on pixels 23 of the photosurface are used to determine distances to features of scene 30 that are imaged on the pixels and provide thereby a 3D image of the scene.

Fig. 2 shows a time-distance graph 60 that illustrates relationships between timing of a light pulse 40 that illuminates scene 30, a gate of photosurface 22 and amounts of light 45 collected by lens 21 and registered by a pixel 23 in the photosurface that images a feature of scene 30 located at a distance D from camera 20. In graph 60, distance D from camera 20 is indicated along an abscissa 61 and time is indicated along right and left ordinates 62 and 63 respectively. For convenience, the ordinates are scaled in units proportional to "ct" - time "t" multiplied by the speed of light "c".

A time-distance line 70 in graph 60 shows a relationship between distance D of a feature in scene 30 from camera 20 and a time, in units of ct, at which a photon emitted by light source 26 at an arbitrary time t₀ = 0 that is reflected by the feature reaches the camera. Since a round trip time, in units of ct, of a photon from light source 26 to the feature and back is given by the equation ct = 2D, line 70 in graph 60 has a slope equal to 2. If the feature is illuminated by photons in a pulse of light, such as light pulse 40 having a pulse width "Axp_W", camera 20 will receive light reflected from the light pulse by the feature for a period having duration Δxp_W from a time at which a first photon reflected by the feature from the light pulse reaches the camera.

For example, assume a feature, represented by a circle 71, of object 31 in scene 30 that is illuminated by a pulse of light 40 emitted at time t₀ = 0 is located at a distance Df from camera 20. In graph 60, a circle, also labeled with the numeral 71, located on time-distance line 70 having a D coordinate Df schematically represents the feature. Reflected light from feature 71 first reaches camera 20 at a time tj(Df), which satisfies an equation, t₁(Df) = 2Df/c. (1)

Light from light pulse 40 reflected by feature 71 continues to be incident on the camera for a period equal to the pulse width Aτp_W of the light pulse 40 until a time t2(Df), t₂(D_f) = 2D_f/c + Δτ_pw, (2) at which time a last reflected photon from light pulse 40 reaches the camera. Block arrows 72 extending from time tj(Df) to time t2(Df) and having a length Axp_W shown along left and right hand time ordinates 62 and 63 schematically represent light reflected by feature 71 that reaches camera 20.

Assume further that controller 24 gates photosurface 22 (Fig. 1) on for a relatively short gate represented by a rectangle 80 that extends from a start time Wj to a stop time W₂, along left hand time ordinate 62 and has a gate width Δτ_Sgw ⁼ ^sg2 " Wl • Optionally, Aw₁₁, = Aτp_W.

If short gate 80 is on, as shown in graph 60, during the period from ti(Df) to t2(Df) for which light 72 reflected by feature 71 from light pulse 40 is incident on camera 20, a pixel 23

(Fig. 1) imaging the feature will register light from the feature. The amount of light 72 registered by the pixel imaging feature 71 is useable to determine distance Df and is graphically represented by a shaded area 81 in rectangle 80. Let the amount of registered light 81 be represented in symbols by "Q". Then distance Df of feature 71 from camera 20 is given by, Df= ctø/2 + (cΔ_τ/w)(Q/Q₀-l)/2 if t_sgχ < t₂(D_f) < W₂. and (3) Df= cWi/2 + (cAxp_W)(l-Q/Q₀)/2 if tø < I₁(Df) < W₂. (4)

In equations (3) and (4), Q₀ is an amount of light that would be registered by pixel 23 imaging feature 71, were all the light reflected from pulse 40 by the feature and collected by lens 21 registered by the pixel, irrespective of whether or not the reflected light reached the camera during short gate 80. Q₀ is a function of reflectance of feature 71 and its distance Df from the camera. Generally, Q₀ is determined by controlling light source 26 to emit at least one light pulse 40 and for each emitted light pulse 40, gating camera 20 with a long gate so that all light reflected by feature 71 from the light pulse that reaches the camera is imaged on photosurface 22.

Camera 20 receives and registers light reflected from light pulse 40 on its photosurface 22 during short gate 80 for any feature of scene 30 located at a distance Df between lower and upper bound distances, DSL ³^ DSy respectively, from the camera. The upper and lower bound distances define a "slice", schematically indicated in graph 60 by a shaded rectangle 83, of scene 30. Shaded rectangle 83 has diagonal corners that lie at points along time-distance line 70 having distance coordinates DSL ^and DSu- DSL, DSU and Df satisfy a relationship, DS_L = <t_sgl-Aτ_pw)/2 < Df < DSu = ct_Jg2/2. (5)

Slice 83 has a spatial, "slice width" SW = (DSu - DS_L) = (<%2 - ct^l + cΔw,,)/2 = cAτ_pw, (6) where the last equality holds under the assumption that short gate 80 has a gate width Aty,_™, equal to pulse width Δτp_W. Because time-distance line 70 has a slope 2, rectangle 83 has an extent along time ordinate 62 equal to 2SW.

A long gate corresponding to short gate 80 for determining Q₀ for any feature located in slice 83 (i.e. any feature imaged with light reflected from a light pulse 40 by photosurface 22 during gate 80) is schematically represented by a rectangle 90 along right hand time ordinate 63. Light registered by pixel 23 that images feature 71 during long gate 90 is graphically represented by a shaded area 91 in gate 90. To assure that Q₀ may be determined for any feature of scene 30 for which light is registered on a pixel 23 during short gate 80, long gate 90 optionally has a gate width Δτ/,™, at least equal to a sum of short gate width Aw₁₄, and twice the pulse width Aw₁, of light pulse 40, i. e.

^Δτ/gw ⁼ Δτ_sgw + 2Ax_pw = 3A-Cp₁₁,. (7)

The last equality in equation (7) is true under the assumption that short gate 80 has a gate width Δ^χsgw equal to pulse width Aτ_pw of at least one pulse 40. Gate 90 optionally has a start time tjgi relative to time t₀ at which light pulse 40 is emitted that is equal to a time that a first photon from a feature in scene 30 having a distance DSL f^rom camera 20 reaches the camera. Gate 90 optionally has a stop time t/g2 equal to a time at which a last photon from pulse 40 reaches camera 20 from a feature of the scene having a distance DSu fr°^m camera 20. Time t/p-i is earlier than a start time of short gate 80 by a period equal to Aτp_W and t/g2 *^s l^{ater m}an ^a stop time of short gate 80 by Δ%_w so that t/gl ⁼ Gsgl " ^Δτpw) and t/g2 ⁼ βsg2 + Δ^tpw)- (8)

Whereas long gate 90 provides information sufficient to determine Q₀ for any feature in slice 83, it is noted that neither a quantity Q of light 81 registered on pixel 23 that images feature 71 during short gate 80 or a quantity Q₀ of light 91 registered by the pixel during long gate 90, provides information as to which of equations (3) or (4) should be to used to determine

Df. For any feature located in slice 83, and any set of values Q and Q₀ for the feature, there are two different distances for the feature in the slice that will result in a same set of values Q and Q₀.

For example, for any Df for feature 71 there is another complimentary distance Df* for the feature that is located in slice 83 for which pixel 23 that images feature 71 would register same amounts of light during short and long gates 80 and 90 as are registered respectively for the gates for distance Df. hi particular, Df and Df * satisfy a relationship: |D_f- ct_sgl/2| = |Df* - ct_Sg2/2|. (9)

Fig. 3 schematically shows distance Df and complimentary distance Df* for feature 71 in a time-distance graph 75. Amounts of light registered by pixel 23 for distance Df are schematically shown by shaded areas 81 and 91 for short and long gates 80 and 90 respectively, both of which gates are schematically shown along left time ordinate 62. Position of feature 71 for distance Df* is indicated along time-distance line 70 by a circle 71*. Amounts of light registered by pixel 23 for Df* are schematically represented by shaded areas 81* and 91* for gates 80 and 90, which for distance Df* are shown along right hand time ordinate 63. A block arrow labeled 72* schematically represents light from light pulse 40 reflected from feature 71 that reaches camera 20 were the feature to be located at Df*. It is noted that whereas light is registered for distances Df and Df* by pixel 23 at different times during gates 80 and 90, the amounts of light registered during gate 80 are the same in each case, as are the amounts registered during gate 90.

In accordance with an embodiment of the invention, the ambiguity of whether to use equation (3) or equation (4) in determining Df is removed by gating photosurface 22 on following each of at least one light pulse 40 for two short gates, a first, "front" short gate and a second, "back" short gate, optionally having equal gate widths. For example, to determine distance to features of scene 30 in slice 83 having lower and upper bound distances, DSL ^m& DSTJ, photosurface 22 is gated on for a front short gate having a gate width Δτp^g^ at a front gate start time, tpg, following a time at which the light pulse is emitted. The time tpg is given by the equation, tPg ^{= l}sgl = 2DS_L/c + Aτ_pw. (10)

Similarly, the photosurface is gated on for the back short gate at a time following each of at least one light pulse 40 at a time tβg defined by an equation, tBg ⁼ tFsg ^{+ Δτ}Rsrgw C^{1 1}) where, as noted above Δτps,™, is the width of the front gate.

Optionally, both short gates have a gate width, Δτ_Sg_M,, equal to half the light pulse width, so that Aτ_sgw = Aτ_pw/2. (12)

Fig 4 schematically illustrates how two short gates, in accordance with an embodiment of the invention, having gate width Ax₅^ = Δτp_W/2 remove the ambiguity with respect to equations (3) and (4) and thereby whether a feature, such as feature 71 imaged on a pixel 23, is located at a distance Df or Df . A long gate 90 and two short gates, a front short gate 101 and a back short gate 102, are shown in Fig. 4 along left hand time ordinate 62 of a time-distance graph 85. A shaded area 91 represents' light registered by pixel^' 23 during long gate 90 for distance Df. Shaded areas 103 and 104 in front and back short gates 101 and 102 respectively, graphically represent amounts of light registered by pixel 23 for feature 71, located at Df, during the short gates. The same front and back short gates 101 and 102 and long gate 90 are shown along right hand time ordinate 63. Shaded areas 105 and 106 in front and back short gates 101 and 102 shown along right hand time ordinate 63 represent amounts of light registered by pixel 23 during the short gates respectively, assuming that feature 71 were to be located at Df . A shaded area 91 represents light registered by pixel 23 during long gate 90 for Df*. From Fig. 4 it is readily seen that if feature 71 is located at Df, as shown for short gates 101 and 102 along left hand time ordinate 62, front gate 101 registers a greater amount of light than back gate 102. On the other hand, if feature 71 is located at Df*, as shown for front and back gates 101 and 102 shown along right hand time ordinate 63, the front gate registers a smaller amount of light than the back gate. IfQp and Qg represent in symbols amounts of light registered by pixel 23 during front and back short gates 101 and 102 respectively, then Df may be determined in accordance with equation (3) if QQ < Qp and in accordance with equation (4) ifQ_F > QB, /•<?•: Df= ct_Fg/2 + (cΔτ^XQ/Qo-l)^ if Q_B < Q_F; and (13) Df= ct_Fg/2 + (cΔτ_pvi;)(l-Q/Q₀)/2 if Q_F > Q_B. (14)

In equations (13) and (14), optionally, Q = (Qp + Qg).

Typically, to determine Qp₅ Qβ and Q₀ for each pixel 23 in photosurface 22 at least one light pulse 40 comprises a plurality of light pulses and each quantity of registered charge Qp₅ QB and Q₀ is determined from a train of light pulses 40. For example, to determine Qp, controller 24 optionally controls light source 26 to emit a train of light pulses 40. For each light pulse 40, following a time tpg (equation (10)) after the light pulse is emitted, controller 24 gates photosurface 22 on for a short gate 101. A total amount of light registered by each pixel 23 for all short gates 101 is used to provide Qp for the pixel. Similarly, to provide Qg, another train of light pulses 40 is emitted and following a delay of tβg (equation (11)) after an emission time of each light pulse, photosurface 22 is gated on for a back gate 102. A total amount of light registered by each pixel for all the light pulses in the pulse train is used to determine Qg. Q₀ is similarly determined from a pulse train of light pulses 40 and a long gate 90 for each light pulse in the light pulse train, which gate 90 follows the light pulse by a delay of t/gj (equation (8)).

Let a "slice acquisition time", "Ts", be a time during which a scene, such as scene 30, is illuminated by light source 26 with light pulses 40 and camera 20 gated to acquire values for Qp, QB and Q₀ for pixels 23 in photosurface 22 to provide distances to features in a slice of the scene imaged on the pixels. Often, slices imaged by a gated camera such as camera 20 are relatively narrow, and a scene for which distances are to be determined using the camera typically has features located in a range of distances substantially greater than a range defined by lower and upper bound distances DSL ^anc^ DSy (Figs. 2-4) of the slice. Therefore, in general, to determine distances to features in a scene, a plurality of substantially contiguous slices of the scene are imaged using camera 20. For each slice, the scene is illuminated by at least one light pulse 40 and camera 20 is gated on for front, back and long gates to acquire distances to features in the slice. If the range over which distances to features in the scene extends from a distance R\ to a distance R₂, and a slice has a slice width SW, a plurality of N slices, where N = (R₂-Ri)/SW, (15) are required to provide distances to substantially all features in the scene. And a total 3D scene acquisition time Tf, required to image N slices is

T_x = NTs- O⁶)

By way of example, Fig. 5 schematically shows scene 30 shown in Fig. 1 and a plurality "N" of slices S\, S₂ ... Sj^ that are used by camera 20 to provide a 3D image of the scene, which extends from a range R\ to a range R.2- A time-distance graph 95 in Fig. 6 shows timing of the front, back and long gates for the slices relative to time t₀ = 0 at which a light pulse 40 is emitted by light source 26 to illuminate scene 30. The slices shown in Fig. 5 are represented in graph 95 by shaded rectangles labeled Sj, S2 ... Sjq- along time-distance line 70. For convenience, slices represented by S_n having a larger subscript n are farther from camera 20 than slices represented by S_n having a smaller index, and slices whose indices differ by 1 are contiguous.

The short front and back gates and the long gate that define a given slice S_n are graphically represented by rectangles labeled respectively FG_n, BG_n and LG_n respectively. Gates for adjacent slices S_n are shown on opposite sides of time ordinate 62. By way of example, a first slice Sj in range R1-R2 defined by gates FGj, BGj and LG^ is slice 83 shown in Fig. 4 and shaded regions 103, 104 and 91 (Fig. 4) graphically representing light registered by a pixel 23 for feature 71 are shown for gates FGj , BGj and LGj. To provide the contiguous slices, the front gate of any given slice S_n and the front gate of a next further contiguous slice S_n+i are delayed relative to each other by a time equal to 2Atp_W = 2SW/c, the spatial slice width of the slices multiplied by 2 and divided by the speed of light. .

For some applications, a total 3D acquisition time Tj (equation (16)) required to acquire a 3D image of a scene by 3D imaging contiguous slices as described above can be too long to provide satisfactory 3D imaging of the scene. For example, an acquisition time Tj may be too long for scenes having moving features that displace by distances on the order of a slice width during time Tj.

The inventors have determined that a total time Tf for acquiring a 3D image of a scene can be reduced by modifying timing of gates used to acquire values for Qp, QQ and Q₀ for slices of the scene, optionally, without changing the lengths of the gates. In particular, the inventors have noted that, a time delay equal to 2Axp_W temporally separates start times of front gates for contiguous slices, and long gates for the slices overlap (relative to t₀) as shown in Fig. 4. On the other hand, a 3D image of a scene can be acquired, in accordance with an embodiment of the invention, in a reduced total acquisition time Tγ if long gates of adjacent slices are timed so that they do not overlap and when one long gate of adjacent slices ends, the other begins.

However, by timing the long gates so that they do not overlap, front gates for adjacent slices are separated by a time 3Δτp_W rather than a time 2Axp_W. As a result, adjacent slices defined by front and back short gates are no longer contiguous, but are separated by

"interstitial" slices. For the interstitial slices, the front and short gates do not provide 3D information. In accordance with an embodiment of the invention, the non-overlapping long gates provide additional information for determining distance to features located in the interstitial slices. For convenience of presentation, slices of a scene for which short front and back gates provide information are referred to as "regular slices".

5 By configuring gating of a 3D gated camera, such as camera 20 so that long gates provide 3D data for interstitial slices, each long gate and its associated front and back short gates provide 3D data for a regular slice and in addition 3D data for an interstitial slice. If the front, back and non-overlapping long, gates have same gate widths "SW" as corresponding gates having overlapping long gates shown in Fig. 6 that define a slice of a scene, a regular

10 slice of the scene in accordance with an embodiment of the invention has a same spatial width as a slice in Fig. 6. Since a set of front, back and non-overlapping long gates provide 3D data for an interstitial slice in addition to a regular slice, each set of gates that define a regular slice of the scene provides 3D data for features in larger range of distances than a corresponding set of gates in Fig. 6. Therefore, a smaller number of sets of gates having non-overlapping long

,15 gates are required to acquire a 3D image of a scene than is required when long gates overlap. However, in accordance with an embodiment of the invention, a slice acquisition time for a regular slice is about the same as a slice acquisition time for a slice defined using overlapping gates. A total 3D acquisition time Tj for a scene using non-overlapping long gates in accordance with an embodiment of the invention, is therefore generally shorter than a total

-20 acquisition time for the scene using overlapping long gates.

Fig. 7 schematically shows a timing and distance graph 115 that graphically illustrates timing of short front and back, and non-overlapping long gates, in accordance with an embodiment of the invention. Regular slices of scene 30 that are imaged by camera 20 are, graphically represented by shaded rectangles that are labeled RSj RS2 ... RS^ along time-

25 distance line 70. For a given slice RS_n, corresponding front back and long gates of camera 20 that define the slice are graphically represented by rectangles labeled RFG_n, RBG_n and RLG_n respectively. The gates are graphically shown along left hand time ordinate 62, with gates for adjacent regular slices shown on opposite sides of the time ordinate. Relative to a start time for any given long gate RLG_n, the corresponding short front and back gates RFG_n, RBG_n have a

30 same start time as front and back gates FG_n, and BG_n (Fig. 6) have relative to a start time for a corresponding lpng gate LG_n.

First regular slice RS_j in graph 115 is, by way of example, identical to slice S\ shown in graph 95 of Fig. 6. And relative to an emission time t₀ of a first light pulse 40, the start, stop and gate lengths for gates RFG_j, RBGi ³^ RLGi ³^ ^me same respectively as for gates FGj, BG^ and LGμn Fig. 6. Distance to a feature located in a regular slice RS_n is determined from amounts of light Qp, Qg and Q₀ registered by a pixel 23 that images the feature during front, back and long gates RFG_n, RBG_n and RLG_n that define the slice, using equations (13) and (14). Amounts of light registered by a pixel that images feature 71 during gates RFGi, RBGl and RLGj are graphically represented by shaded areas 103, 104 and 91 respectively in the gates.

Because, in accordance with an embodiment of the invention, a long gate RLG_n begins when a preceding long gate RLG_n. \ ends, rather than overlapping the preceding long gate as shown in Fig. 6, regular slices RS_n are not contiguous but are separated by interstitial slices. Interstitial slices are graphically represented in Fig. 7 by shaded rectangles labeled IS_{n n}+i, where the subscripts n, n+1 indicate the regular slices RS_n and RS_n+ \ that bracket the interstitial slice. Assuming that short gates RFG_n, RBG_n have gate widths equal to Axp_W/2 and that long gates RLG_n have gate width equal to 3Δ%_W, each interstitial slice has a slice width "ISW" equal to one half the slice width of a regular slice so that ISW= SW/2 = cΔxpy/1. ^■ (17)

For a feature located in an interstitial slice IS_{n n}+\, a pixel 23- that images the feature does not register light reflected by the feature from a light pulse 40 during short and front gates of regular slices that bracket the interstitial slice. As a result, 3D information for the feature is not acquired by camera 20 during the front and back gates and distance to the feature cannot be determined using conventional equations such as equations (13) and (14). However, whereas the pixel does not register any light reflected from a light pulse 40 by the feature during the short gates, the pixel does register light reflected from a light pulse 40 during long gates of the regular slices that bracket the interstitial slice. By way of example, a feature of scene 30 located in interstitial slice IS 1 2 ^{at a} distance Df(121) is schematically represented by a circle labeled 121 along time-distance line 70. Reflected light registered by a pixel 23 that images the feature during long gates RLGj and RLG2 is graphically indicated by shaded regions 122 and 124 respectively. In accordance with an embodiment of the invention, light registered by the long gates is used to determine distance to the feature.

Let the light reflected from a light pulse 40 by a feature in an interstitial slice IS_{n n}+i of scene 30 and registered during long gates RLG_n and RLG_n+ \ be represented in symbols by QL(Π) and QL(n+l) respectively. If a time at which long gates RLG_n and RLG_n+ 1 begin are represented by and tLg(n) and tLg(n+l) respectively, distance Df to the feature is determined, in accordance with an embodiment of the invention, in accordance with equations

Df= ct_Lg(n+l)/2 - (cΔτp_W)(Q_L(n)/Q_o(n))/2; or (18) Df= ct_Lg(n+l)/2 - (cΔτ^Xl- Q_L(n+l)/Q₀(n))/2. (19)

In equations (18) and (19), Q₀(n) = (QL(Π) + Qjjn+l)).

For a camera, such as camera 20, used to determine distances to features of a scene and gated in accordance with an embodiment of the invention, let amounts of light registered by a pixel that images a feature in the scene during gates RFG_n, RBG_n be represented by Q_F(n) and Q_B(n) respectively. Let times at which the gates RFG_n and RBG_n are turned on be represented by tpg(n) and t_Bg(n) respectively. Then distance Df to the feature may be determined in accordance with an embodiment of the invention, using the following general set of conditions and equations. If QF(II) ≠ 0 or Q_B(n) ≠ 0 (20)

Df= ct_Fg(n)/2 + (cΔτp_W)[(Q_F(n)+ Q_B(n))/Q_L(n) -l]/2, if Q_B(n) < Q_F(n). and (21)

Df= ct_Fg(n)/2 + (cΔ_τi7W)[l-(Q_F(n)+ Q_B(n))/Q_L(n)]/2 if Q_B(n) > Q_F(n) (22)

If Q_F(n) = Q_B(n) = 0 and Q_L(n) ≠ 0 and Qi/n+1) ≠ 0 (23)

Df= ct_Lg(n)/2 - (cΔτ_pw)(Q_L(n)/Q₀(n))/2. or (24) Df= ct_Lg(n)/2 - (cΔτ^_wχi-Q_L(n+l)/Q_o(n))/2. (25)

It is noted that for features located at a distance exactly at the border between a regular slice and an adjacent interstitial slice, light is registered by a pixel 23 imaging the feature only during the long gate that defines the regular slice. Furthermore, there is an ambiguity as to whether the amount of light registered by the pixel corresponds to the feature being located at the border between the regular slice and the following adjacent interstitial slice or the preceding adjacent interstitial slice. Generally, such a situation arises only for a very few features in a scene and in some embodiments of the invention features for which this occurs are ignored.

In some embodiments of the invention, adjacent long gates are timed, relative to time t₀, to "overlap" slightly to remove the ambiguity. For example, a long gate may have a start time earlier than the stop time of an immediately previous long gate by a period equal to or less than about a tenth of a pulse width. Optionally, to provide the overlap long gates are lengthened by the period of the overlap.

From the above discussion and Fig. 7 it is seen that by gating a 3D camera in accordance with an embodiment of the invention, for each set of gates RFG_n, RBG_n and RLG_n 3D data is acquired for a regular slice RS_n of scene 30 having a same width SW as a slice S_n shown in Fig. 6. However, in addition to acquiring 3D data for a regular slice, a set of gates RFG_n, RBG_n and RLG_n and a long gate RLG_n+ j provide 3D information for an interstitial slice ISn,n+l of the scene. For a sufficiently large number N of sets of gates RFG_n, RBG_n and RLG_n, each set of gates may be considered to acquire 3D data for a regular slice and an interstitial slice. Since each regular slice RS_n has a slice width SW and each interstitial slice ISn,n+l has a width ISW = SW/2, a set of gates RFG_n, RBG_n and RLG_n may be considered to acquire 3D data for an "enlarged slice" having an enlarged slice width,

ESW = SW + ISW = (3/2)SW. (26)

To acquire a 3D image of scene 30 in accordance with an embodiment of the invention camera 30 is gated on for N sets of gates RFG_n, RBG_n and RLG_n where

(R₂-Ri)/[3/2)SW], (27) By comparing N to a number of slices, N, defined by equation (15) above, it is seen that iV= (2/3)N. (28)

Assume that a slice acquisition time T^ for a regular slice, in accordance with an embodiment of the invention, is the same as a slice acquisition time for a slice acquired in accordance with an embodiment of the invention illustrated in Fig. 6. A total time, Tj/, to acquire data for a 3D image of scene 30 in accordance with an embodiment of the invention is therefore 2/3 the time required to 3D image the scene in accordance with gating shown in Fig. 6.

In some prior art 3D gating algorithms, distances to features in a slice of scene, such as scene 30, are determined by a short, "extended", front gate having a gate width equal to the pulse width A%p_W of light pulses 40 (Fig. 1) that illuminate the scene. The extended front gate and corresponding long gate begin at a same time relative to an emission time t₀ of respective light pulses 40 that illuminate the scene and provide light that is reflected from the scene and imaged by camera 20 during the gates.

By way of example, Fig. 8 shows a time-distance graph 200 that illustrates timing of an extended front gate 202 and its corresponding long gate 204 relative to an emission time t₀ of a light pulse 40 that illuminates scene 30. Front gate 202 is shown along left hand time ordinate 62 and long gate 204 is shown along right hand time ordinate 63. Gates 202 and 204 begin at a same time relative to time t₀. Whereas, long gate 90 shown in Figs. 2-4 optionally has a gate width equal to 3Axp_W, long gate 204 used with extended front gate 202 optionally has a shorter gate width 2A%p_W. Extended front gate 202 and its associated long gate 204 define a slice, of scene 30 having a spatial width cAip_W/2. The slice defined by the gates is graphically represented by a shaded rectangle 206 along time-distance line 70 in Fig. 8.

Let a time at which extended front gate 202 and its corresponding long gate begin following a light pulse emission time t₀ be represented by t_Sg. Then light acquired by pixels 23 in photosurface 22 during the gates may be used to determine distances Df to features in a slice of the scene having lower and upper distance bounds DSL, DSU where DSL, DSy and Df satisfy a relationship,

DSL ^{= ct}sg^/2 ≤ Df ≤ DSu ⁼ [ct_sg+cAτ_pw]/2. (29)

Slice 206 has a spatial, "slice width" SW = (DSu - DS_L) = [cΔτp_W]/2 (30)

Light registered by a pixel 23 that images a feature 71 of scene 30 located in slice 206 during extended front gate 202 and long gate 204 is graphically represented by shaded areas 208 and 210 respectively in the gates. From Fig. 8 it may be seen that for any feature located in slice 206, an amount of light 210 registered by pixel 23 is substantially all the light reflected by feature 71 that reaches camera 20. However, for such a feature, an amount of light 208 registered by the pixel during extended front gate 202 is dependent on the location of the feature in slice 206. Let the amounts of registered light 208 and 210 be represented in symbols by "Q" and "Q₀" respectively. Then distance Df of feature 71 from camera 20 is given by, Df= ct_5g/2 + (cΔτp^Xl - Q/Q_o)/2. (31) It is noted that unlike for the configuration of gates shown in Figs. 2 and 3, for the configuration of gates shown in Fig. 8 there is no ambiguity as to the location of a feature, such as feature 71, in slice 206. However, pixel 23 that images the feature will register light during both extended front gate 202 and long gate 204 for features that are not in slice 206 but are located just before slice 206 at distances Df' that satisfy an equation, ctyg/2 - (cΔτpw) < Df < ct^/2. (32)

For all such features, an amount of light registered during extended front gate 202 is equal to an amount of light registered during long gate 204, the ratio Q/Q_o is equal to 1 and equation (31) provides a same distance Df = cX_sJ2. In some prior art methods, to prevent erroneous distances measurements for features outside of slice 206, distances Df are not determined for pixels for which Q/Q_o =1. In general, since there are relatively few features for which Q/Q_o =1, ignoring such features does not substantially, adversely affect providing a complete 3D image of the slice.

Distances to features in slice 206 may also be determined in accordance with prior art using an extended back gate instead of an extended front gate. The extended back gate has a gate width equal to that of an extended front gate but ends at a same time relative to a light pulse emission time t_o as a corresponding long gate, instead of beginning at a same time as the long gate. An extended back gate may be considered a mirror image of an extended front gate. Fig. 9 shows a time-distance graph 220 that illustrates timing of an extended back gate

222 and its corresponding long gate 224 relative to an emission time t₀ of a light pulse 40 that illuminates scene 30. Extended back gate 222 is shown along left hand time ordinate 62 and long gate 224 is shown along right hand time ordinate 63. Gates 222 and 224 end at a same time tgp- relative to time t₀. Extended back gate 222 and its associated long gate 224 define a same slice 206 of scene 30 that extended front gate and long gate 202 shown in Fig. 7 define if t_eg = t_Sg+2Axp_W. Light reflected from feature 71 registered by pixel 23 that images the feature during extended back and long gates 222 and 224 is represented by shaded areas 226 and 228 respectively. Distance Df to a feature in slice 206 is given by

Df= ct_eg/2 + (cΔτp_W)(Q/Q₀-2)/2, (33) where "Q" and "Q₀", represent amounts of registered light 226 and 228 respectively. As in the case for gate configurations similar to that shown in Fig. 6, configurations of an extended front gate 202 (Fig. 8) or an extended back gate 222 (Fig. 9) together with a long gate 204 and 224 respectively may be repeated to acquire distances to features in a plurality of optionally contiguous slices of scene 30. If a range over which distances to features in the scene extends from a distance R\ to a distance R₂, and a slice has a slice width SW, a plurality of N slices, where

are needed to acquire a 3D image of the scene. Let a "gate acquisition time" ΔTg be required for a pixel 23 to register an amount of light for an extended front gate, extended back gate or a long gate suitable for use in determining a distance Df to a feature of a scene imaged on the pixel. A 3D acquisition time for a slice of the scene may then be written

TS = 2ATg, (35) and a total acquisition time for the scene be written,

Fig. 10 shows a time-distance graph 240 that illustrates temporal relationships of a plurality of extended back gates and associated long gates that are used to acquire 3D images of slices of a scene, such as scene 30.

A first extended back gate 242 and its associated long gate 244 are shown along left time ordinate 62 and define a slice 246. Amounts of light reflected from a light pulse 40 (Fig. 1) from a feature 248 in slice 246 that are registered by a pixel imaging the feature during extended back gate 242 and long gate 244 are shown as shaded areas 249 and 250 respectively in the gates. A second extended back gate 252 and its associated long gate 254 are shown along right time ordinate 63 and define a slice 256. By way of example, slice 248 in scene 30 are assumed contiguous with slice 256 in the scene and as a result in Fig. 10 touch at a corner. KiπumHwm

Amounts of light reflected from a light pulse 40 (Fig. 1) from a feature 258 in slice 256 that are registered by a pixel 23 imaging the feature during extended back gate 252 and long gate 254 are shown as shaded areas 259 and 260 respectively in the gates.

The inventors have realized that if only long gates are used to acquire a plurality of slices of a scene a total acquisition time, Tp, to acquire data for a 3D image of scene 30 may be reduced relative to a total acquisition time required using an extended front or back gate and an associated long gate.

Fig. 11 shows a time-distance graph 280 that illustrates temporal relationships of a plurality of long gates used to acquire a 3D image of a scene, in accordance with an embodiment of the invention. A sequence of long gates LGj , LG2, LG3 ... used to acquire a 3D image of scene 30 are shown along left time ordinate 62. For clarity of presentation, gates, hereinafter "odd gates", labeled with an odd subscript and gates, hereinafter "even gates", labeled with an even number subscript are shown on opposite sides of left time ordinate 62. Gates LGj and LG2 are also repeated along right time ordinate 63. Each pair of sequential odd and even gates defines a spatial slice of scene 30. For example, gates LGj and LG2 define a slice located along time-distance line 70 labeled SI4 2 in Fig. 11. Light registered by pixels 23 of photosurface 22 during gates LG \ and LG2 may be used in accordance with an embodiment of the invention to determine distance to all features in slice SLj 2 located in the slice. Similarly, gate pairs (LG3.LG4) define slice SL3 4 shown in Fig. 11. In an embodiment of the invention, even and odd gates have a same gate width,

^AτLGW^{= 2}Δτp_W, (37) and each even and odd gate begins at a time (relative to t₀) that is half a gate width, i.e. Aτp_W later than a time at which a preceding odd and even gate respectively begins. As a result, a slice SL_{n n}+i defined by two long gates has a spatial width, SW = cΔτp_W. (38)

Features that are located in slices SLj 2_» ^aΩ& SL3 4, which are imaged by camera 20 are schematically indicated by circles 281, 282 and 283 along time-distance line 70. Amounts of light registered by pixels 23 that respectively image features 281, 282 and 283 during the gates are graphically indicated by shaded regions AB respectively. Amounts of light registered for feature 282 are shown in gates LGj and LG2 along right hand time ordinate 63.

In accordance with an embodiment of the invention, if a pixel 23 registers amounts of light Q_n and Q_n+ 1 during gates LG_n and LG_n+ 1 respectively, distance Df to the feature imaged by the pixel is determined in accordance with the following equations,

D_f= (ct_n/2) + cΔτp_W(Q_n+i/Q_n)/2 if Q_n+1 < Q_n, (39) Df= (ctn+l/2) + cΔτ^OCQn/Qn+O)/! if Q_n < Q_n+I- (40)

From the above discussion and Fig. 11, it is seen that every two long gates provides 3D information for a slice of a scene having spatial width equal to cΔτp_W which is and that η long gates LG_n provides 3D information for features having distances Df in a range from R\ to R2 for which (R2-R1) = ηcΔτp_W/2 so that η = 2(R₂-Ri)/cΔτ_Jw. (41)

Assuming a gate acquisition time ΔTg, a total acquisition time for a scene bounded by distances R^ and R₂ may therefore be written,

A total acquisition time for the scene, in accordance with an embodiment of the invention is therefore one half the prior art acquisition time given by equation (36).

It is further noted for the "interstitial slice" gating method in accordance with an embodiment of the invention schematically illustrated in Fig. 7, as the number of slices used to provide a 3D image of a scene bounded by distances R\ and R₂ increase, a total number of gates required to image the scene approaches that given by equation 41. As a result, for a same scene, a total 3D acquisition time for the scene using the gating configuration schematically illustrated in Fig. 7 approaches that using the gating configuration schematically illustrated in Fig. 11.

In the description and claims of the present application, each of the verbs, "comprise" "include" and "have", and conjugates thereof, are used to indicate that the object or objects of the verb are not necessarily an exhaustive listing of members, components, elements or parts of the subject or subjects of the verb.

The invention has been described with reference to embodiments thereof that are provided by way of example and are not intended to limit the scope of the invention. The described embodiments comprise different features, not all of which are required in all embodiments of the invention. Some embodiments of the invention utilize only some of the features or possible combinations of the features. Variations of embodiments of the described invention and embodiments of the invention comprising different combinations of features than those noted in the described embodiments will occur to persons of the art. The scope of the invention is limited only by the following claims.

Claims

1. A method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse having a pulse width at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a plurality of sets of gates, each set comprising at least one first, second and third gate having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start and stop times of the at least one first gate and the at least one second gate are between the start and stop times of the at least one third gate, wherein for at least one of the set of gates, the at least one third gate has a start time equal to about the stop time of the at least one third gate of another of the set of gates; and using amounts of reflected light imaged on the photosurface to determine distances to the scene.

2. A method according to claim 1 wherein the start time of the second gate is substantially equal the stop time of the first gate.

3. A method according to claim 2 wherein the start time of the first gate is delayed relative to the start time of the third gate by the pulse width of the at least one light pulse.

4. A method according to claim 3 wherein the first and second gates have equal gate widths.

5. A method according to claim 4 wherein the gate width of the first and second gates is equal to half a pulse width of the at least one light pulse.

6. A method according to claim 5 wherein the gate width of the third gate is substantially equal to three pulse widths of the at least one light pulse.

7. A method according to claim 6 wherein if amounts of reflected light from a feature imaged during the at least one first, second and third gates of a set of gates are denoted by Q\, Q2, Q3 respectively and the start time of the at least one first gate by t\, then if Qi ≠ 0 or Q2 ≠

0, distance Df to the feature is determined in accordance with Df= cti/2 + (cΔτ)[(Qi+ Q₂VQs-IR if Qi(n) < Q₂(n) and

Df= Ct₁Il + (CAT)[I-(Q₁+ Q₂)/Q₃]/2 if Q₁(Ii) > Q₂(Ii), where c is the speed of light and Δτ is the pulse width of the at least one light pulse.

8. A method according to claim 6 or claim 7 wherein, if amounts of reflected light from a feature imaged during the at least one first, second and third gates of a set of gates are denoted by Q₁, Q₂, Q3 respectively, the start time of the at least one third gate by t3, and the amount of light from the feature imaged by the at least one third gate of the other set of gates by Q*3 then if Q₁ = Q₂ = 0, Q3 ≠ 0 and Q 3 ≠ 0, distance Df to the feature is determined in accordance with

Df= ct₃/2 - (cΔτ_pw)(Q3(n)/(Q3+Q*3))/2 or

Df= ct₃/2 - (cΔτ_iWχi-Q*3/(Q3+Q*3))/2, where c is the speed of light and Δτ is the pulse width of the at least one light pulse.

9. A method according to any of the preceding claims wherein the at least one third gate has a start time earlier than the stop time of the at least one third gate of the other set of gates by a period that is less than or equal to about one twentieth of the pulse width of the at least one light pulse.

10. A method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a plurality of equal length gates having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start time of at least one first gate is substantially a time half way between the start and stop times of at least one second gate; and using amounts of reflected light imaged on the photosurface during the at least one first gate and the at least one second gate to determine distance to a feature of the scene.

11. A method according to claim 10 wherein the gates have a gate width substantially equal to twice a pulse width of the at least one light pulse.

12. A method according to claim 10 or claim 11 wherein if the start time and amounts of reflected light from the feature imaged during the at least one first gate are denoted respectively by t\, and Qi and for the at least one second gate by t2 and Q2 respectively, distance Df to the feature is determined in accordance with Df= (cti/2) + cΔτ(Q2/Qi)/2 if Q2 ≤ Qi and Df= (ct₂/2) + cΔτ(l-(Q!/Q₂))/2 if Qi ≤ Q₂, where c is the speed of light and Δτ is the pulse width of the at least one light pulse.

13. A method of acquiring a 3D image of a scene, the method comprising: transmitting at least one light pulse having a pulse width at a transmission time to illuminate the scene; imaging light reflected by the scene from the at least one light pulse on a gateable photosurface during a set of gates comprising at least one first, second and third gates having start and stop times and for which, relative to a transmission time of a light pulse of the at least one light pulse, the start and stop times of each first gate and each second gate are between the start and stop times of a same third gate; and using amounts of reflected light imaged on the photosurface to determine distances to the scene.

14. A method according to claim 13 wherein the first gate has a start time delayed relative to a start time of the third gate by about the pulse width.

15. A method according to claim 13 or claim 14 wherein the second gate has a stop time that precedes a stop time of the at least one third gate by about a pulse width.

16. A method according to any of claims 13-14 wherein the first gate has a gate width equal to about half a pulse width.

17. A method according to any of claims 13-16 wherein the second gate has a gate width equal to about half a pulse width.

18. A method according to any of claims 13-17 wherein the third gate has a gate width equal to about three pulse widths.

19. A method according to claim 14 wherein the first and second gates have gate widths equal to half a pulse width and the third gate has a gate width equal to about three pulse widths wherein, if amounts of reflected light from a feature imaged during the at least one first, second and third gates are denoted by Qj, Q2, Q3 respectively, and the start time of the first gate by t\, distance Df to the feature is determined in accordance with Df= ctj/2 + (cΔτ)(Q/Q3-l)/2 if Q2 < Qi; and

Df= cti/2 + (cΔτ)(l-Q/Q₃)/2 IfQ₁ > Q₂, where c is the speed of light, Δτ is the pulse width of the at least one light pulse and

Q - (Qi + Q2)-

20. A camera useable to acquire a 3D image of a scene comprising: a gateable photosurface; and a controller that gates the photosurface in accordance with any of claims 1-19.

21. A camera according to claim 20 and comprising a light source controllable to illuminate the scene with a pulse of light.