GB2472242A - Coded aperture imaging - Google Patents

Abstract

A coded aperture imaging (CAI) system incorporates a detector array 60 adjacent to a coded aperture mask 62 comprising a zone plate array. The mask 62 has light transmissive regions and opaque regions. The opaque regions are sufficiently absorbing that they prevent any significant radiation intensity being reflected at their interfaces and reaching the detector array 60 as detrimental background radiation. The coded aperture mask 62 is much thicker than a prior art equivalent, and radiation only passes through the mask 62 if it has an angle of incidence in a narrow range between rays 64. Consequently, the CAI system field of view is narrowed by making the mask 62 thicker; this causes the CAI system's point spread function (PSF) to have radiation intensity which reduces with change in angle of incidence away from the normal incidence direction and to reach a negligible level when truncated by a detector array edge. This allows optimal image reconstruction.

Description

Coded Aperture Imaging This invention relates to coded aperture imaging.

Coded aperture imaging is a known imaging technique originally developed for use in high energy imaging, e.g. X-ray or 7-ray imaging where suitable lens materials do not generally exist: see for instance E. Fenimore and T.M. Cannon, "Coded aperture imaging with uniformly redundant arrays", Applied Optics, Vol. 17, No. 3, pages 337 -347, 1 February 1978. It has also been proposed for three dimensional imaging, see e.g. "Tomographical imaging using uniformly redundant arrays", Cannon I M, Fenimore E E, Applied Optics 18, no.7, p. 1052-1 057 (1979) Coded aperture imaging (CAl) exploits pinhole camera principles, but instead of using a single small aperture it employs an array of apertures defined by a coded aperture mask.

Each aperture passes an image of a scene to a greyscale detector comprising an array of pixels, which consequently receives a pattern comprising an overlapping series of images not recognisable as the scene and measures an intensity sum. The detector array provides an output which is processed to reconstruct an image of the scene.

A coded aperture mask may be defined by apparatus displaying a pattern which is the mask, and the mask may be partly or wholly a coded aperture array; i.e. either all or only part of the mask pattern is used as a coded aperture array to provide an image of a scene at a detector. Mask apertures may be physical holes in screening material or may be translucent regions of such material through which radiation may pass.

It is an object of the present invention to provide an alternative coded aperture imaging technique.

The present invention provides a coded aperture imaging (CA!) system having a detector array with a normal incidence direction and a coded aperture mask which is sufficiently thick to narrow the CAl system's field of view and to provide for the CAl system's point spread function (PSF) to have radiation intensity which reduces with change in angle of incidence away from the normal incidence direction and reaches a negligible level when truncated by a detector array edge.

The invention provides the advantage that it overcomes the problem of a PSF being partly detected and partly not detected due to the PSF overlapping the detector array's edge, and consequently being truncated to simulate a change when angle of incidence moves away from normal incidence.

The mask and detector array may be separated by a distance D, the mask having thickness h and apertures of width a the detector array having a light sensitive area of width or diameter W, and: a W The mask may have focussing properties or be immediately adjacent to a focusing element, the mask or focusing element having focal length f equal to D. It may be an array of zone plates or one or more photon sieve structures. It may be a plurality of photon sieve structures each with light transmissive regions comprising square section pinholes. It may have has tapering apertures which reduce or increase in cross-section towards the detector array.

In another aspect, the present invention provides a method of coded aperture imaging comprising arranging for radiation from a scene to pass to a detector array via a coded aperture mask, the detector array having a normal incidence direction and the coded aperture mask being sufficiently thick to provide a point spread function with radiation intensity which reduces with change in angle of incidence away from the normal incidence direction and reaches a negligible level when truncated by a detector array edge.

The method aspect of the invention has preferred features equivalent mutatis mutandis to those of the CAl system aspect.

In order that the invention might be more fully understood, embodiments thereof will now be described by way of example only, with reference to the accompanying drawings, in which:-Figure 1 is a perspective view of a prior art coded aperture imaging system; Figure 2 shows two plan views of a CAl system detector array with the system's point spread function in different positions; Figure 3a shows a sectional side view of a detector and associated coded aperture mask illustrating the field of view of a prior art CAl system; Figure 3b is a sectional side view of a detector and associated coded aperture mask illustrating the field of view of a CAl system of the invention; Figure 3c is also a sectional side view of a detector and associated coded aperture mask, and corresponds to part of Figure 3b shown on an expanded scale; Figure 4 shows a photon sieve structure for use in a CAl system; Figure 5 shows a photon sieve structure for use in a CAl system and incorporating square cross section pinholes; Figure 6 is a sectional view of an angular field of view restrictor illustrating the principle on which the invention is based; Figure 7 is a sectional view of an angular field of view restrictor which is much larger than previous embodiments; Figure 8 is a sectional view of an angular field of view restrictor which again is much larger than previous embodiments other than that shown in Figure 7, and which uses one or more thin restrictors; and Figure 9 shows a mask or angular field of view restrictor which uses a tapering aperture.

There is a need to limit a CAl system's field of view to prevent unwanted light entering the system at high angles and impinging on the detector. Limiting the field of view enables the following: * the system field of view to be set without having to change detector dimensions and mask-to-detector distance: this allows use of a large area detector even with a short mask-to-detector distance, which is advantageous for making a thin imager; * off-axis aberrations to be avoided, as light incident at large angles may be poorly focused; and * control of the amount of overlap between images formed by different focusing elements: image reconstruction tends to be better when overlap is less.

In addition, there is a criterion for the CAl system's point spread function (PSF), this being the pattern produced on the system's detector by light coming from a point in the scene: for optimum image reconstruction it is desirable that the PSF varies as little as possible with angle of incidence of light upon the detector.

Figure 1 shows a prior art coded aperture imaging system indicated generally by 10 and taken from WO 2007/091045. Rays of light indicated by arrows such as 12 extend from points in a scene 14 and fall on to a coded aperture mask 16. The coded aperture mask 16 acts as a shadow mask and causes a series of overlapping coded images to be produced on a detector 18, which consists of an array of pixels shown as small squares such as 20. Intensities from the overlapping, coded images falling on each detector pixel are summed at that pixel. The detector array 18 produces an output which is passed to a processor 22: the processor 22 decodes an image of the scene from the detector output using a variety of digital signal processing techniques.

Referring to Figure 2, on the left hand side of the drawing a detector array indicated by a rectangle 30 has a centrally located PSF 32A. The PSF 32A is generated by light passing along a CAl system's optical axis (not shown) from a point in a scene at the centre of the system's field of view (on axis): here the centre of the PSF 32A is at normal incidence and all its light is captured by the detector 30. On the right hand side of the drawing, the detector array 30 is shown receiving light from a point in the scene near the edge of the field of view (off axis): this gives rise to a PSF 32B with a centre which is not at normal incidence; the PSF 32B overlaps the edge of the detector 30, which therefore only partly intercepts it. Consequently, the PSF 32B is effectively truncated by the edge of the detector. In terms of light actually detected by the detector 30, the PSF 32B has changed with the change in angle of incidence away from the normal: even if the PSF 32B as a whole is identical to the PSF 32A, the detected part of the PSF 32B is not equal to the detected part of the PSF 32A.

Referring now to Figure 3a, a detector array indicated by a rectangle 40 has a prior art coded aperture mask 42 above it. As is conventional in the prior art, the coded aperture mask 42 is thin, a typical mask design comprising a layer of chromium 44 deposited on glass 46. Light rays from a scene (not shown) having angles of incidence which are not more than those of light rays 48 pass through the mask 42 to the detector 40. Light rays having angles of incidence which are greater than those of light rays 48 are intercepted by opaque regions of the mask 42 and do not pass to the detector 40. Because the mask is thin, it allows light to pass through with a wide range of angles of incidence between rays 48 giving rise to the possibility of a PSF being truncated by the detector's edge.

In Figure 3b, a detector array indicated by a rectangle 60 has a coded aperture mask 62 above it, the detector array 60 and mask 62 being for use in a CAl system of the invention. The mask 62 incorporates an array of zone plates. It has light transmissive unshaded regions and opaque shaded regions: the opaque regions are absorbing, and light intercepted by them is absorbed and not transmitted to the detector array 60. The opaque regions are sufficiently absorbing that they prevent any significant light intensity being reflected at their interfaces and reaching the detector array 60 as detrimental

background light.

The coded aperture mask 62 is much thicker than the prior art coded aperture mask 42 in Figure 3a. For reasons given in relation to Figure 3a, light rays from a scene (not shown) only pass through the mask 62 to the detector 60 if these rays have angles of incidence which are less than or equa' to those of light rays 64: other light rays are absorbed by the mask's opaque regions. Because the mask 62 is much thicker than a prior art mask, it allows light to pass through only if the light has an angle of incidence in a narrow range between rays 64. Consequently, the CAl system field of view is narrowed by making the mask 62 thicker; this causes a PSF's light intensity to reduce with change in angle of incidence away from the normal and to reach a negligible level when truncation occurs at the edge of the detector 60. This allows optimal image reconstruction.

The geometry of the optical system defined by the mask 62 and detector 60 is shown in Figure 3c, in which parts described earlier are like-referenced. The detector 60 is separated from the zone plate array in the mask 62 by a zone plate focal distance f. The mask 62 has thickness h and apertures of width a; with an image of a scene centred on the detector 60 then the criterion for the image not overlapping the boundary of the detector's light sensitive area of width or diameter W is: a W If a mask is used which does not have focussing properties or an immediately adjacent focusing element (lens or zone plate), then replacing f by a mask -detector separation 0 gives: a W The boundary of the detector's light sensitive area limits the maximum extent of the detector region to be illuminated by a light cone, i.e. the diameter of the light cone at the detector should not be any greater than that of this area. However, it may be desirable for the diameter of the light cone to be smaller than the detector's light sensitive area, as in the case where the useful field of view of the zone-plate is limited by its off-axis aberrations. Alternatively, a smaller field of view may be required and you are using the larger detector area to increase the amount of light collected and thus boost the signal to noise.

The mask 62 has the advantage of potential ease of manufacture: it could be made from a sheet of photographic film, for example. However, there is one disadvantage in that the width of the zone-plate rings varies: the central zone is widest, and the width of the higher-order zones decreases as their radii increase. So as the angle of incidence increases, light that goes through the higher-order zones is cut-off first. This will introduce some undesirable angle-dependent variation in the PSF: the PSF's focussed spot will get larger and the positions of its side-lobes will change. One way to avoid this is to use a "photon sieve" structure in which a zone consists of an array of pinholes, as opposed to a continuous region.

A photon sieve structure 70 is shown in Figure 4: it consists of a central near-circular array of pinholes surrounded by three rings of pinholes, all pinholes being shown unshaded. It acts similarly to a zone plate but the rings are not continuous -instead they are arrays of unconnected apertures. The structure 70 is not a conventional photon sieve, because the pinholes all have the same size and they all cut off light rays at the same incidence angle (uniform field of view restriction). A conventional photon sieve has apertures which decrease in size with distance from the sieve centre. Less light is transmitted by a photon sieve compared to a zone plate. Photon sieves are known for use with X-rays: see L. Kipp, M. Skibowski, R. L. Johnson, R. Berndt, R. Adelung, S. Harm & R. Seemann, "Sharper images by focusing soft X-rays with photon sieves" Nature, Vol. 414, p184 (2001).

Referring now to Figure 5, another embodiment of the invention is shown: it incorporates a detector array 80 and a coded aperture mask 82 seen in sectional side view on the right of the drawing. A plan view 82p of the coded aperture mask 82 is also shown, looking in the direction of an arrow 84. As shown in the plan view 82p, the coded aperture mask 82 consists of twenty-eight partially transmissive regions 84 in a two dimensional array. Each partially transmissive region 84 is shown on an expanded scale at 86, and consists of a central near-circular array of pinholes surrounded by two rings of pinholes, all pinholes being of square cross-section as shown at 88 giving the region 84 a box-section structure. The square pinholes 88 are closely packed into a grid structure, which improves light transmission.

It is desirable for the partially transmissive regions 84 to be as thin as possible (i.e. pinholes as short as possible) to increase light transmission.

The field of view restriction technique of the invention implements the principle illustrated in Figure 6. In this drawing an opaque block 100 (shown in section) with thickness h has an aperture 102 through it with diameter a. A transmitted ray indicated by an arrow 104 has a maximum permitted angle of inclination 8 to the block's thickness dimension. 0 is given by:-a tan0=-h For 8 �= 10, i.e. a 10 or less angular field of view restrictor, and a 10pm diameter aperture 102, the mask thickness h would be �= 56.7pm. A typical metal mask on a glass substrate with less than 1 pm thickness of metal (h value) and hence aperture length

would not restrict the field of view at all.

Referring now to Figure 7, an angular field of view restrictor 120 is shown in section which is much larger than those previously described. As illustrated, the restrictor 120 has an aperture 122 which allows light to be incident upon a zone plate 124 at angles of incidence up to a maximum which is that of a transmitted ray indicated by an arrow 126.

In a practical version of the restrictor 120, there are multiple apertures 122 and zone plates 124 (which form an array), but there is a single aperture per zone plate. The apertures 122 are much larger than in previous embodiments i.e. lOOs of pm, in order to allow light to reach the whole of a zone plate, and they extend through 1mm of opaque material. The restrictor 120 is a layer which is separate from the array of zone plates, and the centres of the apertures 122 require alignment with those of respective zone plates 124 in the array.

Referring now to Figure 8, an angular field of view restrictor 140 is shown which again is much larger than embodiments other than restrictor 120. The restrictor 140 is much thinner than restrictor 120, and it is held 1mm away from a zone plate 142 by means of an intervening sheet of glass (not shown) 1mm thick. Here again, in a practical version of the restrictor 140, there are multiple apertures 144 aligned with respective zone plates 142 in an array, but there is a single aperture per zone plate.

This single thin layer restrictor approach 140 may give rise to rays with high angles of incidence entering the aperture leading to one zone plate but intended for another zone plate. This is not a problem for sparse arrays, i.e. arrays in which not au possibie zone plate locations are actually populated with zone plates, but it may contribute to noise in those more densely arranged.

A laminated approach may optionally be employed in which one or more additional thin restrictors 146 and intervening sheets of glass (not shown) simulate a thicker restrictor.

This laminated approach is potentially cheaper than a single thick-layer restrictor but it requires additional relative alignment of apertures and zone plates.

Referring now to Figure 9, an angular field of view restrictor 160 is shown which uses a tapering aperture 162 reducing in cross-section towards a detector array 164. This is sometimes referred to as a horn or funnel structure. The aperture 162 allows light to be incident upon a zone plate 124 at angles of incidence up to a maximum which is that of a transmitted ray indicated by an arrow 166. Here once more, in a practical version of the restrictor 160, there are multiple apertures 162 aligned with respective zone plates (not shown) in an array, but there is a single aperture per zone plate. In an alternative embodiment, in an angular field of view restrictor with tapering apertures, the taper could increase in cross-section towards a detector array.