GB2186149A - Image differencing using masked CCD - Google Patents

Abstract

A charge coupled device has some of its "pixels" masked by a material which is opaque to the radiation to which the device is to be exposed, each masked region being employed as a storage zone into which the charge pattern from the unmasked pixels can be transferred to enable a subsequent charge pattern to be established on further exposure of the unmasked pixels. The components of the resulting video signal corresponding to the respective charge patterns read-out from the CCD are subtracted to produce a video signal corresponding to the difference between the two images which formed the respective charge patterns. Alternate rows of pixels may be masked, Fig. 5a, or chequer-board pattern masking may be employed, Fig. 5b. In an X-ray imaging system the CCD is coupled to image intensifying and converting means. <IMAGE>

Description

SPECIFICATION Improvements in and relating to image capture Field of the invention This invention concerns image capture devices and is concerned with devices and systems which are equally applicable in scientific, medical and industrial applications. The invention also concerns X-ray imaging systems involving image capture and is again concerned with improvements which are equally applicable in medical and industrial X-ray applications.

Background to the invention In medical applications the majority of X-ray examinations still depend upon transmission X-radiology using film. The method has the advantage of good resolution and simplicity but the storage of patient records and the delays in processing film to provide X-rays for consultant use lack compatability with other information handling systems being introduced into hospitals and clinics. Thus patient records are now increasingly stored on computer data bases and can be called up instantly by a consultant for review and update. However by its nature, X-ray film, often an essential ingredient of the patients records, cannot be simi larly recalled.

It is one object of the present invention to provide an alternative X-ray analysis system which enables at least one X-ray record of a patient to be stored for recall using a computer data base.

It a further object of the present invention to provide an X-ray analysis system which will enable a consultant to interface with the radiologist during the production of X-ray photographs of a patient in a more convenient manner than has hitherto been possible so as to enable the radiologist to obtain precisely those X-ray photographs which are particularly applicable to the patient and the consultant requirements.

It is a related object to speed up the process of producing an X-ray image of the patient and recording same.

It is a still further object of the invention to reduce patient exposure to X-rays.

As a result of the invention it should be possible to interface more easily X-ray analysis data relating to a patient together with other data arising from imaging devices such as NMR, CT and ultra-sound scanners and from nuclear medicine.

In many inspection systems more generally, it is necessary to be able to detect change such as changes in shape or position or colour.

Typically this is done by capturing a succession of images of the subject under study and comparing electrical signals derived from the capture of each of the images to obtain a difference signal. Where no detectable change has occurred the difference signal will have a steady state value and any changes will appear as measurable departures in the difference signal from the value associated with the steady state condition.

In other inspection procedures the object of interest may be only barely discernable from its surroundings with the result that its precise shape or size is difficult to evaluate in a screen display. If the object under study has a variable component such as a changing shape or changing position relative to its surroundings, then by using a differencing technique as described above, a screen display can be limited to the partial outline of the object itself making observation and study and measurement of a screen display of the object much easier.

Where the component is moving relatively slowly with time the image capturing process is relatively straightforward and conventional TV scanning processes can be employed for producing the succession of images in combination with a strobe light source, flash illumination or the like (where such procedures are necessary to arrest the motion or the changing parameter).

However where very high speed image capture is needed due to very short duration events of interest or very high speed of movement or change in the object of interest, conventional imaging techniques such as described are inadequate and it is an object of the present invention to provide an image capture device capable of forming two images separated by as little as 1/10 of a microsecond. The invention lies in the modification of a known form of charge coupled device as defined in the following brief description, and hereinafter referred to as a device of the type described.

Such a device comprises a charge coupled device having a plurality of discrete regions in the surface thereof each of which is separately responsive to incident radiation and is referred to as a pixel. The pixels are arranged in rows and columns and the electrical charge resident in any one of the pixels is determined by the amount of radiation (such as light radiation) incident thereon. An electrical signal corresponding to the charge pattern is obtained by shifting the charges from one pixel to the next along one of the rows of pixels so that the charges along that row (the output row) appear in succession at an output of the device as a serial data stream.The charges along the other rows can be read out in a similar serial manner from the same output by parallel transfer of the charge pattern from one row to the next so as to successively displace the charge pattern in the output row with that from an adjoining row, each charge pattern so resident in the output row being read out seri ally via the output before the next parallel transfer occurs.

Summary of the invention.

An image capture device according to one aspect of the invention comprises a charge coupled device of the type described in which some of the pixels are masked by a material which is largely opaque to the radiation to which the device is to be exposed wherein the masked region comprises a storage zone into which the charge pattern from the unmasked pixels can be transferred by fast parallel shift to enable a subsequent charge pattern to be established on further exposure of the unmasked pixels.

Thus after the device has been exposed to produce a first pattern of charge on the unmasked pixels the pattern can be shifted to the protected (masked) pixels to enable a further charge pattern to be formed on the unmasked pixels.

Where the number of masked pixels is N times as great as the unmasked pixels it is possible to shift N successive charge patterns from the unmasked pixels into the masked pixels and with the last remaining charge pattern in the unmasked pixels store N+1 charge patterns on the device for subsequent serial readout via the output of the device.

In one embodiment of the invention there comprises a charge coupled device of the type described in which alternate rows of pixels are masked by the opaque material.

In another embodiment of the invention there comprises a charge coupled device of the type described in which alternate pixels along each row are masked by the opaque material and the masking along adjoining rows of pixels is staggered so as to produce a socalled chequer-board pattern of masked areas.

Where a chequer-board pattern of masking is used, the result will be that the pixels will alternately contain information from the first and the second of two images supplied to the device and on reading out the array in the normal manner pixel by pixel, a difference signal can be obtained in real time by simply subtracting the first and second analogue signals, separated for example by high speed switching as each pixel is read in turn or by a delay stage to enable the two signals to be synchronised for presentation to the subtraction stage.

In an alternative arrangement the digitised output from the CCD may be supplied both directly and via a one-pixel read-out period delay to the input of a digital subtraction stage so as to produce the aforementioned difference signal.

As each pixel is read out, so a difference signal corresponding to the difference between the current pixel information and the previous pixel information will arise at the output and can be displayed using a television type display or the like.

The production of a difference signal of this nature can be used to a low inspection of a moving part relative to stationary surroundings even though normally it would be virtually impossible to distinguish between the moving part and the surrounding material. This arises from the fact that image content in the output signal relating to the stationary material or surroundings will remain constant over the length of time that the analaysis is made whilst the moving part will have moved a small distance and will produce a disturbance in the charge pattern on the CCD camera. It is this change in the charge pattern which will be seen when the two signals are subtracted and which will constitute the information content of the difference signal.

Thus for example it should be possible to produce the outline of a piston moving in a cylinder in an engine when inspected by an Xray probe since the casing of the engine will remain stationary throughout the inspection whilst the piston will have moved through a very short distance between the first image and the second image.

The first and second images may be produced by a strobing of the illumination of the object in synchronism with the row transfer of the charges in the charge coupled device.

In another application, the movement of an internal organ in the human body may allow the outline of that organ to be "seen" by exposing the body to X-rays, establishing a first charge pattern on the CCD camera, transferring it to the masked pixels and forming a second charge pattern and thereafter reading out the two interlaced charge patterns and forming a difference signal. Those parts of the body which have remained stationary during the two exposures (which need only be separated by a microsecond or less) will not produce any information content in the difference signal whereas any organ which is moving will produce a difference signal which will correspond to the outline of the organ in so far as that outline has moved in the time taken to transfer the charge pattern below the opaque regions and form the second charge pattern.

Since it is possible to switch the CCD camera on a continual basis and thereby shift the two charge patterns alternately produced, so it is possible to continually inspect an object over a period of time, the display being thereby limited to the outline of any part of the original object which is moving in the time for which it is observed. Thus for example during an X-ray examination the outline of a beating heart or lung or artery maybe inspected in real time not necessarily with the need for Catheters and lodine and the like dyes which have characterised previous radiographic analyses of this type.

The modified charge coupled device of this aspect of the invention may to advantage be incorporated in an X-ray inspection system.

Thus according to another aspect of the invention an improved X-ray analysis system comprises: (1) An X-ray source, (2) An imaging device including image intensifying means for producing a reduced size optical image of a subject illuminated by the X-ray source, (3) An image converter including a photo sensitive element having a plurality of regions therein which are separately addressable and for which digital electrical signals can be generated corresponding to the level of light incident thereon, and (4) Signal processing means by which the digitised signals can be stored distributed and or displayed using a television or the like display device, wherein the image converter preferably comprises a charge coupled device in which some of the pixels are masked to enable a charge pattern to be stored therein during subsequent exposure of the device to a second image.

In a medical situation, such a system will enable a radiologist to display for a consultant a television type picture of the X-ray image immediately it is been taken to enable the consultant to advise whether any further views are required. Since the system can operate in real time, the consultant may even interface with the radiologist and have displayed a sequence of shots with the ability to store any or all of the shots as the views are displayed on the screen.

An advantage of the digitization of the electrical signals corresponding to the light pattern produced by the X-ray image is that the signals are immediately suited for processing by computer and storage in conventional digital storage devices such as discs and on tape.

Advantageously the image converter has a wide dynamic range. By this is meant that the device can handle images with a wide range of brightness and simultaneously enable detail to be seen in regions of low local contrast at all levels of brightness within the overall brightness range to which the system can respond. This has a distinct advantage over Xray systems associated with film in that using a charge coupled device camera as the image converter, a dynamic range can be obtained which is much greater with that associated with ordinary film.

The wide range can be fully exploited by digitizing the signals obtained by addressing the different regions of the camera to an appropriate number of levels and, if sixteen bit digitization is utilised, a grey level scale of the order of 1 to 50,000 can theoretically be obtained. Whilst the eye may not be able to distinguish the different grey levels, digital processing circuitry can be used to detect and enhance the signals from such an image converter by using digital thresholding techniques.

Normally all of the separately addressable pixels within a CCD camera will be addressed during read out so as to produce the highest possible resolution. Clearly the greater the number of pixels consulted, the greater will be the information content of the resulting information signal.

On the other hand the larger the number of pixels to be addressed, the slower will be the operation of the system and sometimes it is more important to obtain a less highly resolved image in less time or reduce the amount of information which must be transmitted to reconstitute at least an approximation to the original image to enable the data to be transmitted over a transmission system of reduced bandwidth.

Typically an inspection system such as one within which a modified charge coupled device as described herein has been incorporated will include a mode selection device for selecting a high resolution mode of operation or one or more lower resolution modes of operation. In the high resolution mode, all of the pixels will be interrogated during read out whilst in a lower resolution mode, only some of the pixels are interrogated.

If the device is modified with a chequerboard mask so that alternate pixels when read out relate to two different images, a difference signal can be produced very simply and easily using a simple analogue circuit so that a difference signal can be obtained in real time.

Depending on the time required to read out a large CCD array, so the difference signal can be obtained almost instantaneously from the presentation of the two images to the system.

If a large CCD array is used and the blanking of the photoreceptor elements or pixels covers segments of 9 adjacent pixels within any column, then up to 10 separate images can be stored in the array for subsequent read out and using averaging techniques, noise and related interference can be substantially reduced so as to provide a reliable image for subsequent analaysis. By suitably phased shifting of an image up and down to re-expose and remask it many times, 10 (say) average images can be simultaneously accumulated on the CCD.

By altering the phase of the switching so a series of images can relate to different phases of the movement of a machine part or engine member and by altering the synchronisation position, for example each of a large number of turbine blades rotating past a given position can be inspected in turn by synchronously inspecting each blade as it rotates through that given position by producing a series of images of that blade.

According to another aspect of the present invention in an inspection or image analysis system fitted with a modified charge coupled device as aforesaid, the charge coupled device is adapted to operate as a large analogue shift register to enable image information captured thereon to be recycled to thereby "Freeze" the image for subsequent display by means of a monitor immediately after capture.

By using an appropiate charge coupled device having charge output stages at two opposite corners of the array, designed for bidirectional read out, charge can be introduced into the CCD from one corner (input) corresponding to the signal currently being put out from the opposite corner (the output) and in this way the charge pattern relating to the entire image can be continually recycled and simultaneously read out for continuous and simultaneous video display.

Each element of the charge pattern that is read out corresponds to an element of an overall analogue signal and clearly if this signal were simply reinjected at the input without correction for charge lost during the passage through the CCD and for electrical noise in the output stage, the recycled image information would become progressively degraded with each passage through the device. This problem can be avoided by dividing the continuous range of output amplitudes into N discreet levels or steps where the amount of each charge per pixel lost at each cycle through the device is significantly less than Qmax/N, where Qmax is the maximum charge which can be stored at any one of the regions in the display.If an analogue to digital converter (ADC) is provided the output of the display and a charge Q is measured by the ADC for any given region, then the charge which must be reinjected at the input stage is taken to be the nearest multiple of Qmax/N to the analogue value Ct. This technique of quantisation of intensity levels into discrete steps prevents the progressive degradation of the image.

It has been proposed to use charge coupled It has been proposed to use charge coupled devices (CCD's) to capture successive images typically separated by about 1 ms, although resolution could be traded in exchange for a proportional increase in capture rate. In order to achieve this, an optical mask has been employed which covers completely a section of the surface of the device to thereby constitute a distinct storage zone separate from the imaging area. The capture rate of such a device is limited by the time necessary to shift a newly acquired image by the many rows (typically several hundred) needed to bring it under the mask of the storage zone. During the readout phase, the images exposed during the imaging phase are read out separately, one after another.

The present invention employs an optical mask which is interleaved evenly throughout the whole CCD-imaging area, so that an exposed image may be brought to lie in the storage area of the present invention by the parallel shift of only one row (or N rows if (N+1) images are to be captured successively). Thus the "dead time" between successive captures is reduced to the time needed to shift by only one row, and is independant of the resolution of the images. In an application where it is desired to exchange an image between the imaging and storage areas many times, so as (for example) to re-expose an aircraft turbine in the various phases of its rotation over a long period of time, this feature of the present invention becomes of crucial importance because the fraction of charge lost during the exchange cycle is reduced by the power of the number of row shifts.During the readout phase, the present invention allows corresponding regions of the sequence of images to be read out together, so that arithmetic combination of the images may be performed simultaneously with the operation of readout, without the necessity of prior digitisation and recording in computer storage media.

It has also been previously proposed to use a television camera in conjunction with an Xray system but the present proposals should not be confused with such previous arrangements. The inherent signal to noise ratio of a television based system is several hundred times worse than the signal to noise ratio obtainable from a charge coupled device camera and the latter is quite adequate to match the full quality of brightness information obtainable from an image intensifier. Excessive X-ray doses are required to obtain sufficient brightness levels for television based analysis systems to function correctly and the combination of a charge coupled device camera with an image intensifier has allowed for the first time a significant reduction in X-ray dosage to be obtained whilst still maintaining full sensitivity over the complete dynamic range of the image intensifier.

According to another aspect of the present invention, in an X-ray analysis system, means is provided for monitoring the quantity of light that has entered the image intensifying means which is thus proportional to the quantity of X-rays passing through the subject and which form the X-ray image. Since the quantity of light processed by the image intensifier is directly proportional to th electrical signal pattern which will develop in the first image intensifier, and since the sensitivity of the latte ie. current per photon will be known, the Xray source can be switched off as soon as sufficient current has been passed in the first image intensifier.

According to a further aspect of the present invention an X-ray analysis system of the type described herein can be adapted to capture two or more X-ray images separated by periods of time as short as one microsecond or less by employing a modified CCD image capture device as described in our co-pending patent application reference C151R.

By using the modified CCD device, two im ages can be captured almost simultaneously on the same CCD device and then read out in the normal way as one image for subsequent processing, particularly digital subtraction.

The incorporation of the second aspect of the invention in an X-ray analysis system makes possible the realisation of clinical techniques based on image subtraction, free from effects of blurring due to patient or organ motion, but previously excluded by the long switching time between images of typically a few tenths of a second. A primary application is to digital subtraction angiography involving rapid switching between monoenergetic X-rays having energies above and below lodine Kabsorption edges or involving the subtraction of images of coronary arteries or arteries of the brain obtained at cardiac systole and diastole.

The modified apparatus also makes possible the X-ray analysis of moving objects such as rotating turbine blades, oscillating pistons and meshing gears to enable inspection of component parts of machines, engines and the like whilst under load and in operation. The high speed switching as between one image and another can be synchronised with the movement of the part which is to be inspected and a strobe effect obtained.

The invention will now be described by way of example with reference to the accompanying drawings in which: Figure 1 is a schematic circuit diagram of an X-ray analysis system primarily intended for medical application.

Figure 2 is a schematic circuit diagram of an inspection system, Figure 3 is a similar schematic circuit diagram of the image hold arrangement envisaged for use with the modified device of the invention, Figure 4 illustrates diagrammatically the sensitive regions on a CCD camera chip and the masking of alternate regions.

Figure 5 illustrates in two views the two different ways in which alternate sensitive regions in a CCD camera chip can be masked to allow for successive image capture.

Figure 6 is a schematic circuit diagram showing one way of obtaining a difference signal in real time during read-out of the CCD array of Fig 4b.

Detailed description of drawings In Figure 1 X-rays from a source 10 permeate a subject shown as a dotted line at 12 and impinge on the input of a first image intensifier 14. The input diameter of unit 14 is typically 350 millimeters and occupies the position which would normally be occupied by film or a photographic plate.

Image intensifier 14 has the following properties and characteristics: (a) A fluorescent screen of a type (for example ZnCds) chosen to yield absorption of typically 15% for X-rays and to match the range of X-ray energies from the source 10; (b) Electrostatic focusing with a demagnifying characteristic such that the output diameter is typically 55 millimetres; (c) A photon gain in the range 10 to 50 for visible (green) photons; (d) Fibreoptic input and output windows; (e) An S2Q photocathode and a P20 or some other suitable phosphor such as P46 or P47 if higher speed is required.

(f) A resolution typically of the order of 5 to 10 line pairs per millimeter at 50% MTF referred to the input.

The input window and the output windows of intensifier 14 are denoted by reference numerals 16 and 18 respectively.

The second image intensifier 20 having an input window 22 and output window 24 is supplied with the output from the output window 18 of intensifier 14. The second intensifier input and output windows have similar diameters of 55 millimeters and the following characterisitics and properties: (a) A photon gain chosen so that when combined with the photon gain of the image intensifier 14, the complete dynamic range of a charge coupled device camera 26 can be realised when the latter is fitted to an output window 24 of the intensifier 20.

(Typically a photon gain with a range of 500 to 1000 would be appropriate for a first generation two-stage electrostatically focused image intensifier tube having an S20/P20 combination.) (b) A resolution typically of the order of 25 line pairs per miilimeter at 50% MTF.

(c) Fibre-optic input and output windows.

Mention has been made of the CCD camera device or image converter 26. The device selected has the following properties: (a) A format of 2048 > < x 2048 2048 pixels (b) A pixel size of approximately 27 x 27 microns (c) An imaging area of approximately 55 x 55 square millimeters (d) A dynamic range of the order of 50,000: 1 (ie the photon level required to completely saturate a pixel with charge is approximately 50,000 times greater than the photon level corresponding to the device noise level).

(e) Charge saturation per pixel is of the order of 700,000 electrons.

(f) Fibre optic input coupling device or lens for receiving an image from the image intensifier.

A series of image intensifiers as described will have a resolution which is signifcantly better than that of a CCD device having a 2048 x 2048 pixel format so that the resolution will be dictated by the latter and not by the image intensifying apparatus. Effects of distortion and vignetting can be corrected digitally during subsequent processing of the signals pro duced from the CCD camera.

For convenience the two image intensifier units and the CCD camera device are assembled in a light tight electrically insulated container to form a compact portable unit to replace the film unit of a conventional X-ray imaging system and these components together with the image intensifier power supplies generally designated 28 are shown contained within a dotted outline 30 in Figure 1 corresponding to the light tight container. Optionally the CCD control, digitiser and image hold circuits to be described may also be included within the container if appropriate.

Associated with the X-ray source 10 is an automatic "Dose" control 32. The latter serves to monitor the quantity of light that has entered the first image intensifier 14 and since this is proportional to the X-rays incident thereon, this gives a measure of the Xrays which have passed through the subject 12. The circuit 32 generates a short response-time (typically less than one microsecond) switching signal to shut off the X-ray source 10 after a given quantity of light has been processed by the image intensifier so as to minimise the X-ray dose delivered to the subject whilst still ensuring good picture quality.

As shown in Figure 1, the CCD camera 26 is associated with a control circuit 34 a digitiser 36 and an image hold feedback loop 38.

CCD device has charge output stages at two opposit corners designed for bi-directional read out. Charge can be introduced into the CCD device from one corner (the input) and by simultaneously reading the charge pattern from the opposite output corner, via a digitiser 36, and feeding this information back via the image hold circuit 38, so the entire charge pattern can be continually recycled to enable continuous video display of the charge pattern to be obtained by display of the recycling signal.

If the analogue output signal were simply reinjected at the input corner without correction for charge lost during passage through the CCD and noise arising at the output stage, the charge pattern would become progressively degraded with each passage. This problem is avoided by dividing the continous range of output possibilities from each region in the display into N discrete steps (where the amount of charge per pixel lost at each cycle through the device is significantly less than Qmax/N (where Qmax is the maximum charge which can be stored at any pixel).If a charge Ct is measured by the digitiser 36 at the output stage then the charge which should be reinjected at the input stage is selected as the nearest multiple of Qmax/N to the analogue measurement Ct. By quantising the intensity levels into discreet steps the progessive degradation of the image can be prevented.

In Figure 2 light from a source 200 permeates a subject shown as a dotted line at 180 to be focussed by a lens 240 on a modified charge coupled device of the present invention again referenced 26. Alternatively the subject may be illuminated from the front by source 220 and the reflected light therefrom is focussed by the lens 240 as aforesaid.

For high resolution work and if a large dynamic response range is required then the device 26 should have the properties previously detailed in relation to Figure 1.

For convenience the lens 240 and the CCD camera device 26 are assembled in a lighttight electrically insulated container to form a compact portable imaging unit to replace the film unit.

These components together with the CCD power supply 280 are shown contained within a dotted outline 300 in Figure 2 corresponding to the light-tight container.

As shown in Figures 1 and 2, the CCD camera 26 is associated with a control circuit 34 a digitiser 36 and an image hold feedback loop 38.

The preferred CCD device has charge output stages at two opposite corners designed for bi-directional read out. Charge can be introduced into the CCD device from one corner (the input) and by simultaneously reading the charge pattern from the opposite output corner, via a digitiser 36, and feeding this information back via the image hold circuit 38, so the entire charge pattern can be continually recycled to enable continuous video display of the charge pattern by display of the recycling signal. If the analogue output signal were simply reinjected at the input corner without correction for charge lost during passage through the CCD and for noise arising at the output stage, the charge pattern would become progressively degraded with each passage.This problem is avoided by dividing the continuous range of output possibilities from each region in the display into N discrete steps (where the amount of charge per pixel lost at each cycle through the device is significantly less than Qmax/N (where Qmax is the maximum charge which can be stored at any pixel). If a change Ct is measured by the digitiser 36 at the output stage then the charge which should be reinjected at the input stage is selected as the nearest multiple of Qmax/N to the analogue measurement dCt. By quantising the intensity levels into discreet steps the progressive degradation of the image can be prevented.

To this end the digitiser 36 comprises an analogue to digital converter and the image hold circuit 38 is a synchronised switch and a digital to analogue converter for reconstituting the charge level on a quantised basis for each digital value obtained during read-out. By applying the electrical signal so produced to the input of the CCD 24, so the charge pattern is reconstituted.

Referring again to Figures 1 and 2, the CCD control circuit 34 additionally controls the rate at which the CCD is addressed and the number of charged regions which are seen during read out. It is thereby possible to operate the system in a high resolution mode addressing each of the separate pixels during read out so as to obtain maximum information about the image presented to the CCD camera or in one or more lower resolution modes in which less than all of the pixels are addressed during read out. Thus for example in a CCD array having approximately 2000 x 2000 pixels a reduction in resolution in the ratio of 4 to 1 is obtained by simply combining four adjacent pixels during read-out thus creating an effective array of 500 x 500 pixels on a coarser spacing with a consequent decrease in the time required to read out the information from the display.

By limiting the number of pixels addressed the time required to read out the image information from the CCD camera can be reduced sufficiently to allow operation of a 40 HZ or similar rate television type display with approximately 500 x 500 pixel resolution thereby allowing real time display of the image presented to the CCD camera. Thus by adjusting the CCD control 34 to read out the CCD display at the appropriate lower resolution rate, or by using suitable buffering to enable the full resolution to be obtained, so signals are available for display on a television monitor 42 or like display device and are also available via a distribution device 44 for display via a signal path 46 on similar monitors.

Hard copy can also be produced using a dotmatrix printer, laser printer or the like such as 48 connected to the distribution network 46 via signal path 50.

The digital signal produced by the ADC digitiser 36 can be transferred via a digital link 52 for display, processing, storing and distribution. The appropriate network control and processing circuitry is denoted by reference numeral 54 and typically this will allow for certain digital processing of the digital signal should this be required and for the display of the signal via a signal path 56 on a television monitor screen such as 58 or in a high resolution display device capable of displaying the full high resolution signal obtainable from the CCD camera 26 albeit at a lower frame repetition rate of typically 4 Hz. A high resolution display is shown at 60.

As with the lower resolution image, hard copy using a dot-matrix or laser printer or the like 62 can be obtained by transferring the high resolution signal along signal path 64.

The network control stage 54 also produces an output along signal path 64 by which information can be supplied to a network shown diagrammatically at 66 to enable remote display of information at either high or low resolution and for archiving purposes where the archiving process is to be remote from the place of image capture.

The detail of the image hold cycle is contained in Figure 3. Here the CCD array is shown diagrammatically at 68 with the input "Corner" being designated at 70 and the output "Corner" at 72.

The direction of charge flow during read out is denoted by the arrows such as 74 and 76.

Thus the bottom-most line of the array of charge coupled elements is read out as if it were a shift register and after the complete charge pattern of that line has been read out the charge pattern on each of the lines in the array is shifted in parallel from one line to the next adjoining line in the direction of the arrows 74 so that the preceeding line of charge pattern now appears in the last line ready for serial read out to follow the previous line of charge pattern already read out.

The signal produced by the serial read out of the charge pattern from the bottom-most line of the array appears as an analogue signal of varying current (or voltage) and this is converted using a high speed analogue to digital converter 78 into a digital signal which is then recirculated via a high speed digital to analogue converter 80 to be impressed at the input 70 over the charge pattern remaining in the top-most line in the array which like the bottom line can be addressed like a shift register. Thus information applied serially at the input is stepped along the line but can then be shifted in parallel from that line to the next line in the array as each line shift occurs.

By using an analogue to digital conversion followed by digital to analogue conversion, signal to noise ratio can be improved and degradation of the signal which would otherwise occur if the analogue of the output signal from 72 were to be recycled and inserted at 70, can be prevented.

For the purposes of display on a video monitor such as 42, a digital to analogue convertor 82 is also provided for converting the digital output from the ADC 78 to a signal more suited as the video input signal to a monitor 42.

The digitiser is partly under the control of the CCD control circuit 34 of Figures 1 and 2 which determines the manner of read out of the display 68. Thus if fast read-out and low resolution are required, the parallel transfer from line to line which would normally operate on a one to one basis i.e. one line shifting to the next and then waiting whilst the last line of the array is read out, can be changed to operate so that two, three, four or more parallel line shifts are effected before the bottommost line is read out again.

Read out of the last line can itself be speeded up if the analogue to digital converter has only to operate every second, third or fourth etc. pixel.

It will be noted that if only every (for example) third or fourth line is read out pro perly and if (for example) only every third or fourth pixel is digitised, the signal which will be available for insertion at 70 will be a degraded version of the signal which would have been available at 72 if the array had been read at high resolution. The result is that after a first read-out in the low resolution mode, the information which will be fed back and restored in the array will be of the low resolution picture and not the original high resolution information. If a subsequent high resolution signal is required, it will be necessary to reexpose the camera to the original image before read out at the high resolution rate.

In accordance with the invention a high speed comparison between one image and another can be effected by masking some of the light-sensitive pixels in the CCD array and exposing the array to the first image, to then transfer the charge pattern into at least some of the masked pixels to leave the unmasked pixels available for the second image to be presented thereto.

It will be seen that by transferring information from one line of exposed pixels into a line of masked pixels so the information in the whole of that line will be preserved and the exposed pixels will then be available for exposure to the changing light pattern (assuming that the image is altering) to enable a second electrical charge pattern to be built up corresponding to the second image for subsequent read out.

Where a chequer-board pattern of masking is used, the result will be that the pixels in the entire display will alternately contain information from the first and then the second of the two images supplied to the CCD array and on reading out the array in the normal manner pixel by pixel, a difference signal can be obtained in real time by simply subtracting the first and second analogue signals, separated for example by high speed switching as each pixel is read in turn or by a delay stage to enable the two signals to be synchronised for presentation to the subtraction stage.

Figure 4 of the drawings illustrates how the different regions in the surface of a charge coupled device are alternately masked as by an opaque covering 84. The intermediate regions are left exposed to incident light.

The opaque regions 84 may extend along the rows of pixels leaving alternate rows uncoated as shown in Figure 5a or alternatively alternate regions along each row may be covered with an opaque layer such as 84 and the regions in adjoining rows are likewise alternated so as to produce a so called chequerboard pattern of opaque areas such as shown in Figure 5b. Here the opaque regions are defined by reference numerals 86 in one row and 88 in the next row.

In Figure 5a the left hand view illustrates the first charge pattern I on the exposed rows of pixels. In the right hand view the rows have been shifted in the direction of the arrow 90 so that charge pattern I is now held in the covered pixels and the exposed regions now receive light from the second image.

In Figure 5B, it will be seen that with a similar shift in the direction of arrow 90, the information in the alternate exposed regions along the rows and columns of the array shown in the left hand view of Figure 5B is transferred with the one line shift in a downward direction to the covered regions leaving the exposed alternate pixel regions capable of receiving information relating to the second image.

Figure 6 demonstrates how a CCD array such as that obtained after a double exposure as shown in the right hand view of Figure 5B can be read to produce a real time difference signal. The output from the CCD is supplied both directly and via a one-pixel read-out-period delay 92 and an inverting amplifier 94 to an adding stage 96. The inverting amplifier converts for example a +5 signal to a -5 signal and the one-pixel read-out delay device 92 ensures that the two signals supplied to the inputs of the adding stage 96 correspond to the signal arising from the current pixel which is being read out and the inverted value of the information from the preceeding pixel.

By adding the real time signal to the delayed inverted signal so a difference signal is obtained immediately which relates to the difference between the information contained in adjoining pixels.

As the information arising at point 98 varies as each pixel is read out, so a difference signal corresponding to the difference between the current pixel information and the previous pixel information will arise at the output 100 and can be displayed using a television type display or the like.

The production of a difference signal of this nature can be used to allow inspection of a moving part relative to stationary surroundings even though normally it would be virtually impossible to distinguish the one form the other.

This arises from the fact that image content in the image supplied to the CCD camera relating to the stationary material or surroundings will remain constant over the length of time that the analaysis is made whilst the moving part will have moved a small distance and will produce a disturbance in the charge pattern on the CCD camera. It is this change in the charge pattern which will be seen when the two signals are added in the adding stage 96 and which will constitute the information content of the difference signal at output 100.

Thus for example it should be possible to produce the outline of a piston moving in a cylinder in an engine when inspected by an Xray probe and image intensifier for producing the light image for application to the CCD array, since the casing of the engine will remain stationary throughout the inspection whilst the piston will have moved through a very short distance between the first image and the second image.

The first and second images may be produced by strobing the X-ray source in sychronism with the row transfer of the charges in the CCD camera.

In another application, the movement of an internal organ in the human body may allow the outline of that organ to be "seen" by exposing the body to X-rays, establishing a first charge pattern on the CCD camera, transferring it as in Figure 5B and forming a second charge pattern and thereafter reading out the two interlaced charge patterns and forming a difference signal. Those parts of the body which have remained stationary during the two exposures (which need only be separated by a microsecond or less) will not produce any information content in the output signal at junction 100 whereas any organ which is moving will produce a difference signal which will correspond to the outline of the organ in so far as that outline has moved in the time taken to transfer the charge pattern below the opaque regions and form the second charge pattern.

Since it is possible to switch the CCD camera on a continual basis and thereby shift the two charge patterns alternately produced, so it is possible continually to inspect an object over a period of time, the display being thereby limited to the outline of any part of the original object which is moving in the time for which it is observed. Thus for example the outline of a beating heart or lung or artery maybe inspected in real time not necessarily the need for Catheters and iodine and the like dyes which have characterised previous radiographic analyses of this type.

Although the examples given have involved the use of an X-ray probe and possibly the use of image intensifiers the invention is not limited to the use of such devices and is equally applicable to being usd with a normal optical image obtained as shown in Fig. 2.

Thus the invention provides an ultra high speed time lapse camera which depending on the masking pattern provided on the pixels can receive 2, 3 ,4 or more successive images and store each even though the time lapse between exposure is less than 1 microsecond. In the case of a CCD array having a capability of shifting one line to the next in 1/10 th of a microsecond the time lapse will be of that order.

In a modification of the invention, the CCD camera element 26 may be incorporated into the image intensifier 20 to remove the need for a final phosphor in the second image intensifier.

Claims (17)

1. An image capture device comprising a charge coupled device of the type described in which material which is largely opaque to the radiation to which the device is to be exposed wherein the masked region comprises a storage zone into which the charge pattern from the unmasked pixels can be transferred by fast parallel shift to enable a subsequent charge pattern to be established on further exposure of the unmasked pixels.

2. Device according to claim 1, including means to expose and thereby produce a first pattern of change on the unmasked pictures and thereafter to shift the pattern to the protected (masked) pixels to enable a further charge pattern to be formed on the unmasked pixels.

3. A device according to claim 2 wherein the number of masked pixels is N times as great as the unmasked pixels and to means are provided to shift N successive charge patterns from the unmasked pixels onto the masked pixels and, with the last remaining charge pattern in the unmasked pixels, to store N + 1 charge patterns for subsequent serial read out via the output of the device.

4. A device according to claim 2 in which alternate rows of pixels are masked by the opaque material.

5. A device according to claim 3 in which alternate pixels are masked by the opaque material and the masking along adjoining rows of pixels is staggard so as to produce a chequer-board pattern of masked areas.

6. A device according to claim 5, having a chequer-board pattern of masking, including means for obtaining a different signal in real time by subtracting the first and second analogue signals separated by high speed switching as each pixel is read either in turn or by a delay stage to enable the two signals to be synchronised for presentation to the subtraction stage.

7. A device according to any of claims 1 to 5, in which the digitised output from the device is supplied both directly and via a one pixel read out period delay to the input of a digital subtraction stage so as to produce the aforementioned different signal.

8. A device according to any of claims 1 to 7, wherein the first and second images are produced by strobing of the illumination of the object in synchronism of with the row transfer of the charges in the device.

9. An X-ray analysis system comprising: (1) An X-ray source, (2) An imaging device including image intensifying means for producing a reduced size optical image of a subject illuminated by the X-ray source, (3) An image converter including a photo sensitive element having a plurality of regions therein which are separately addressable and for which digital electrical signals can be generated corresponding to the level of light incident thereon, and (4) Signal processing means by which the digitised signals can be stored distributed and or displayed using a television or the like display device.

10. A system according to claim 9 wherein the image converter comprises a charge coupled device in which some of the pixels are masked to enable a charge pattern to be stored therein during subsequent exposure of the device to a second image.

11. A device or system according to any one of the preceeding claims, wherein means are provided for addressing the CCD camera during read out so as to produce the highest possible resolution.

12. A device or system according to any preceeding claim, incorporating a mode selection device for selecting a high resolution mode of operation or one or more lower resolution modes of operation.

13. A device or system according to any one of the preceeding claims, wherein the charge coupled device is adapted to operate as a large analogue shift register to enable image information capture thereon to be recycled, thereby to "freeze" the image for subsequent display by means of a monitor immediately after capture.

14. A device or system according to any one of the preceeding claims wherein the charge coupled device has charge output stages at two opposite corners of the array, designed for bi-directional, whereby charge can be introduce into the device from one corner corresponding to the signal currently being fed out from the opposite corner.

15. A device or system according to any one of the preceeding claims, including means for dividing a continous range of output amplitudes into N discreet levels or steps wherein the amount of each charge per pixel lost at each cycle through the device is of significantly less than Qmax/N, where Ctmax is the maximum charge which can be stored at any one of the regions in the display.

16. A charge coupled device of the type described substantially as herein before exemplified with reference to the accompanying drawings.

17. An X-ray imaging system substantially as herein before described with reference to the accompanying drawings.