PATENT SPECIFICATION an 1 564 324

(21) Application No. 49448/76 (22) Filed 26 Nov. 1976 (31) Convention Application No. 636 212 (32) Filed 28 Nov. 1975 in CO (33) United States of America (US) (44) Complete Specification published 10 April 1980 3 (51) INT. CL. G01T 1/17 (52) Index at acceptance G1A A4 C13 CI C3 C4 C6 D10D11 D12D3 G16 G17 G1 G2 G7 P5 R2 Sll S3 SL T15 T20 T3 H4F D18X D27E D27K D27M D30K D56X D61D83B D85L (72) Inventor CARL JOSEPH BRUNNETT

(54) SCANNING APPARATUS AND METHOD

(71) We, PICKER CORPORATION, a are unobtainable by conventional radio- corporation of the State of New York, of graphical procedures. In Tomography, an 595 Miner Road, Cleveland, Ohio 44143, image viewed from a cross-sectional plane United States of America, do hereby de- of interest extending through a subject is ' 5 clare the invention, for which we pray that developed by sequentially directing X-rays 50 a patent may be granted to us and the through the subject from a plurality of method by which it is to be performed, to directions. The resulting image clearly re- be particularly described in and by the veals relative spatial relationships of inter- following statement: — nal structures of the subject in the plane 10 This invention relates generally to sys- of interest, and is relatively easily inter- 55 tems for nondestructively examining ob- preted. jects using penetrative radiation and, more Early tomographic systems utilized a particularly, relates to an improved method radiation detector whose movement was co- and apparatus for measuring the intensity ordinated with movement of a radiation 15 of X-radiation emerging from the subject source which directed a radiation beam to 60 of a clinical examination. the detector. The source-detector pair The following United States patents and moved about an axis passing through the patent applications are referred to here- subject, and the system produced a cross- in:— sectional image of the subject in a selected 20 United States Patent Application Serial plane of interest which extended trans- 65 No. 559.411, (Patent No. 3,976,885) (here- verse to the axis of the X-ray beam. The in referred to as the SINGLE MOTION movement of the source-detector pair was patent application); such that internal structure in the plane United States Patent Application Serial of interest was continuously scanned by 25 No. 559,412, (Patent No. 4,052,620) (here- the beam. This scanning technique resulted 70 in referred to as the DUAL MOTION in substantially continuous changes in the patent application); spatial relationship between the detector United States Patent Application Serial and the source and the internal structure No. 635,952, (Patent No. 4,008,400) (here- of the subject. These changes blurred images 30 in referred to as the BACKSIDE SCAN- of the structures out of the plane of in- 75 NING patent application). terest with the result that an image of the A conventional radiograph is a two- structures in the plane of interest was pro- dimensional shadow image of a three- duced. dimensional subject. The depth dimension Other tomographic procedures have 35 is not apparent since all interior struc- been proposed which develop a cross-sec- go tures of the subject appear to be in a single tional image of the subject viewed from a plane. As a consequence, a conventional plane which includes the axis of the X-ray radiograph sometimes fails to provide beam. Tomography which produces such necessary detail concerning relative spatial images is known as transverse section tomo- 40 locations of interior structures, is difficult graphy. This type of tomography has re- 85 to interpret, and may not reveal the exist- suited in production of a reconstructed ence of a condition of interest in the sub- image, or representation, of a transverse ject. section through the subject being examined. Tomographic procedures have been de- Transverse section scanning has evolved 45 veloped to fulfill some objectives which into two general types of systems. In one 90 2 1 564324 2

such system a radiation source-detector measured for all such paths during all the pair scans a subject with a beam of radia- scans are back projected to produce a scan- tion emitted as the source-detector pair are by-scan build-up, or reconstruction, of the translated in a plane containing the sec- image. 5 tion of the subject to be examined. A num- More specifically, each value of the 70 ber of such scans are completed during radiation transmission as it is back pro- each examination with the angular orien- jected for a given path is kept constant, tation of the beam relative to the subject and the respective values of each back being changed from one scan to another. projection at points of intersection of the 10 Each scan is divided into individual scan respective paths are added together. Each 75 segments. The radiation passing through point on the reconstructed image is there- the subject during each scan segment con- fore representative of the sum of the back stitutes, in effect, a single beam passing projected intensities of the paths passing through the subject along a narrow path. through that point. This technique is de- 15 The detected intensity of the beam for each scribed in Kuhl, "A Clinical Radioisotope 80 scan segment throughout each scan is Scanner for Cylindrical and Section Scan- recorded for computing X-ray transmission ning," PROC. SYMP., Athens 1964, Medi- (or X-ray absorption) characteristics cal Radioisotope Scanning, I.A.E.A., through the scanned section. The charac- Vienna, 1, 273, 1964. 20 teristics are appropriately processed to The back projection technique has been 85 provide a reconstructed image of the in- improved with the introduction of filtered ternal structure of the subject in the scan- back projections and data processing using ned plane. Fourier analysis. A discussion of Fourier In another transverse section scanning reconstruction using filtered back projec- 25 system, a radiation source-detector pair tions is set forth in Chesler, "Positron 90 scans the subject with the beam while or- Tomography and Three Dimensional Tech- biting about the subject in a plane. After nique," PROC. SYMP. on Radionuclei each orbit the source-detector pair is in- Tomography, New York, N.Y., 1972. An crementally pivoted about an axis passing Algorithm for processing the data using 30 through the source, and another orbit in convolutions on a digital computer is given 95 the same plane is completed. Each orbital in Shepp, et al., "Some Insights into the scan is formed by a continuous succession Fourier Reconstruction of a Head Section," of individual scan segments, and the in- Bell Laboratories, Murray Hill, N.J., 1974. tensity of the beam for each scan segment Prior Art 35 is detected and recorded for computing the Fundamental to the success of the tomo- 100 X-ray transmission or adsorption charac- graphic scanning systems utilizing recon- teristics of the subject. The accumulated struction tomography procedures is the data from the scans are processed to pro- ability to accurately determine the inten- duce a reconstructed image viewed from sity of an X-ray beam as it impinges upon a 40 the plane. detector after having emerged from the 1°5 In a modification of the noted "orbital" subject. In these scanning systems the de- system, multiple closely spaced detectors tector includes a scintillator coupled with have been used with a common X-ray a for generating an source. Use of multiple detectors enables, analog data signal whose level is represen- 45 in some circumstances, production of good tative of the intensity of the detected radia- 110 image resolution after a single orbit of the tion. source and detectors about the subject. In The fundamental approach to measur- effect the single detector-multiple scan ap- ing the intensity of the detected radiation proach is traded off, at least in some cir- has been to integrate the data signal over 50 cumstances, for a multiple detector-single the time period of the scan segment, called 115 scan approach. This latter approach is de- an integration interval, to produce a sig- scribed in the DUAL MOTION, the nal representative of the average detected SINGLE MOTION, and the BACKSIDE intensity over that time period. Early pro- SCANNING applications. posals for doing this employed a conven- 55 Transverse section tomography systems tional integrator for integrating the data 120 of both general types have commonly util- signal throughout the integration interval ized a computational technique known as and an analog-to-digital converter for con- "back projection" for processing the radia- verting the integrated data signal to a digi- tion intensity data to reconstruct the im- tal value. This digital value, when com- 60 age. The detected intensity of the X-ray pared with the duration of the time period, 125 beam passing through the subject along a represented the average detected intensity given narrow path (defined by a scan seg- of the beam over the period. ment) is back projected, or attributed, to Because the integrated data signal was all points on the path of the beam. The an analog signal, evaluating it with extreme 65 values of radiation transmission intensity precision was difficult. Accordingly the 130 3 1,562.3 3

statistical accuracy of the radiation inten- charge to the integrator circuit each time sity measurements produced by this ap- the integrator output level exceeds a proach was limited. The integrators them- threshold. This coaction allows substan- selves introduced an unavoidable delay in tially all input current to be utilized by 5 processing the data signals with the extent the integrator while providing a low delay, 70 of the delay being proportional to the wide bandwidth circuit. bandwidth of the integrator. When wide One such voltage-to-frequency converter bandwidth integrators were employed for is commercially available from Teledyne accommodating a wide range of data sig- Filbrick Corporation under Model No. 10 nal values, the processing delay produced by 4707, and is described by that company's 75 the integrators tended to be maximized. data sheet of October 15, 1974. One sug- The delays were inherent in the construc- gested application for this converter is in tion of the integrators and to the extent the nuclear data acquisition wherein the input delays were not completely compensated current may vary between 0 and 850 micro- 15 for, the statistical accuracy of the radia- amps. While such input current levels are 80 tion intensity measurements suffered. acceptable for nuclear data acquisition, A subsequent approach which improved tomographic X-ray systems typically gener- statistical accuracy somewhat used a data ate 10 microamps maximum input current, signal amplitude-to-frequency converter. therefore rendering the referenced volt- 20 In one prior art system the data signal was age-to-frequency converter unsuitable for 85 converted to a variable frequency signal X-ray tomographic applications. ranging from zero to ten megahertz in Another limitation encountered with the direct proportion to the data signal ampli- variable frequency data pulse approach was tude. In this approach, a conventional in- that, if the data pulse frequency is low 25 tegrator circuit was coupled with a thresh- compared to the extent of the integration 90 hold detector. The integrator was operated interval, errors of substantial magnitude so that its output level was driven below could be encountered in determining the the threshold of the threshold detector im- average frequency for any given integra- mediately after the integrator output level tion interval. More specifically, the inte- 30 exceeded the threshold. This caused the gration interval may begin and end at any 95 threshold detector to produce a train of point between the occurrences of sequen- data representing pulses whose frequency tial data pulses. Assuming for the purpose varied in direct relation to variations in of discussion, that the data pulse frequency the data signal level. is constant over a period somewhat longer 35 The variable frequency data pulse ap- than an integration interval, if the inte- 100 proach improved statistical accuracy since gration interval commences immediately the number of pulses occurring during a after the conclusion of one pulse and con- given integration interval could be accu- cludes immediately before a pulse, it is rately counted. An average frequency was apparent that the number of intervening 40 determined by, in effect, dividing the pulse pulses counted during the interval is one 105 count by the duration of the integration less than the count would be if the interval interval, and the resulting value repre- had happened to commence just prior to sented the average detected radiation in- the first mentioned pulse and included the tensity for the corresponding path tra- last mentioned pulse. Accordingly, the 45 versed by the X-rays. The bandwidth re- pulse count for a given integration inter- 110 lated processing delay inherent in the con- val could differ by a count of one pulse struction of the integrator adversely af- depending upon the occurrence of data fected the operation of these systems be- pulses, in. time, relative to the beginning cause, during the time the integrator out- and end of the integration interval. When 50 put was driven below the threshold level the data pulse frequency is high relative 115 of the threshold detector, the integrator to the duration of the integration inter- circuit was disabled from responding to in- val a one-pulse error may be substantially put data signals. As a result, input data negligible (e.g., on the order of 0.1% was irretrievably lost. Minimizing the loss of where the pulse frequency is 1000 Hertz 55 data in these systems required elaborate high per integration interval). However, at a 120 speed, relatively expensive electronics if ac- low count rate, for example 20 pulses per ceptable operation at high frequencies (e.g., integration interval, a one-pulse error up to 10 megahertz) were to be expected. causes the precision of the measurement to An improved voltage-to-frequency con- deteriorate to a 5% error. 60 verter utilizing a different integrator cir- Still another limitation on the statistical 125 cuit, known as a charge pump integrator, accuracy of the systems referred to resides was recently developed. Charge pump in- in variations in extent of the integration tegrators include a charge generator as- interval. Assuming that a constant radia- : sociated with an integrator circuit. The tion level is incident on a detector, a con- 65 charge generator selectively dispenses stant data signal level is therefore input 130 4 1564324 4

to the voltage-to-frequency converter. As through a subject during a predetermined a result, a constant frequency signal is out- period includes detecting the intensity of put from the converter. However, if the radiation emerging from the subject; pro- integration intervals indicated to the con- ducing a train of signal pulses the rate of 5 verier are not of precisely equal length, the occurrence of which varies according to 70 resultant digital representation of the aver- the detected radiation intensity; counting age detected radiation level will vary from pulses which occur during the predeter- one integration interval to another in ac- mined period and producing a pulse count cordance with variations in the indicated representative of one less than the num- 10 integration interval. ber of pulses occurring during said prede- 75 The extent of the integration intervals termined period; timing an interval defined has conventionally been determined by de- by the time elapsing between the first and tecting incremental changes in position of last pulses occurring during the predeter- a source-detector pair relative to a fixed mined time period and producing a time 15 reference point. Typically the position de- signal the value of which represents said 80 tection scheme has utilized a series of interval; and dividing the number of pulses equally spaced marks which are scanned counted by the time signal value to pro- by a photosensitive device as the source- duce an output signal representative of the detector pair moves. The output from the average radiation intensity detected during 20 photosensitive device is a series of pulses, said predetermined period. 85 each pulse corresponding in time to a mark The invention is particularly applicable disposed at a predetermined location rela- to a transverse section tomographic radia- tive to the photosensitive device. The oc- tion system including a radiation scanning currence of each pulse terminates one in- system for accomplishing a succession of 25 tegration interval and initiates a succeed- scans of a subject being studied; a data 90 ing interval. processing unit for processing signals from The integration intervals must be of the scanning unit which are indicative of brief duration (e.g., no more than about 5 (1) detected radiation intensity, (2) scan milliseconds) to obtain acceptable resolu- segment intervals, and (3) radiation path 30 tion of the tomographic image. This has location relative to the subject; and an 95 normally required the marks to be very imager which receives reconstructed image accurately located, but unavoidable varia- signals output from the data processing tions in relative distance between marks unit and produces a cross-sectional image resulted in inconsistent integration of a desired planar section of the subject. 35 intervals and consequent loss of accuracy. The cross-sectional image of the subject 100 The foregoing example assumes that the is reconstructed by use of detected radia- source-detector pair speed is constant. It tion intensity data accumulated from small should be apparent through, that if the segments of a number of scans of the radia- rate of speed of the source-detector pair tion in the plane of the cross-section of the 40 varies even slightly, the extent of the inte- subject. In a preferred embodiment of the 105 gration interval is altered with the same invention each of the scan segments is de- resulting effect on system accuracy as that fined in time by a primary time period dur- caused by inconsistent mark spacing. ing which the detected radiation passes According to one aspect of the present through the subject along a narrow path in 45 invention, a radiation detection system in- the plane (i.e., the radiation is effectively no cludes means to produce a series of pulses a narrow beam). having an instantaneous frequency which Radiation intensity detected during each varies in relation to detected radiation in- primary time period is represented by an tensity; and circuitry for determining the analog DATA signal which is converted 50 rate of occurrence of said pulses during a to a pulsating signal whose frequency 115 predetermined period, the circuitry com- varies as a function of the amplitude of the prising first counting means to count pulses DATA signal. Pulsations of the converted which occur during the predetermined DATA signal occurring during the prim- period, timing means to provide a time ary time period are counted and the occur- signal having a value representative of the rence of the pulsations is used to define a 120 time elapsing between the first and last secondary time period of variable duration pulses occurring during the predetermined within the primary time period. The pulse time period, and means responsive to the count for a given secondary period is output of the first counting means and the compared to the actual duration of the 60 time signal to produce an output indicative secondary period to provide a relatively 125 of the average radiation intensity detected precise indication of the average radiation during the period. intensity detected during the secondary According to another aspect of the pre- period. sent invention, a method of detecting the This average intensity is attributed to 65 intensity of penetrative radiation passing the entire primary time period, and thus 130 4 1 562 768 5

to the beam path of the corresponding scan period. Accordingly, the duration of the segment, in reconstructing an image of the secondary time period, during which the subject. As a result, statistical error result- pulses actually occur, is precisely deter- ing from small unavoidable variations in minable and is accurately represented by 5 the extent of the primary time periods the first output signal. The pulses, apart 70 from scan segment to scan segment are ob- from the first pulse, occurring during the viated. Moreover, statistical errors which secondary time period are counted by the could be encountered as a result of varia- data signal and the number of tions in relative positions (in time) of the pulses counted is accurately represented by 10 beginning and end of the primary time the second output signal. 75 period with respect to the occurrence of The reconstruction processor, in a pre- the converted DATA signal pulsations are ferred embodiment of the invention, in- avoided. cludes a comparator which compares the In the preferred embodiment of the in- first and second data signal processor out- 15 vention, an X-ray tomographic scanning put signals and produces the AVERAGE 80 system having an X-ray source and an as- INTENSITY signal. The reconstruction sociated X-ray detector is employed. The processor also receives POSITION signals detector produces analog DATA signals output from the scanning system so that representative of the detected intensity of each AVERAGE INTENSITY signal can 20 X-rays which have passed through a sub- be correlated with the position of its re- 85 ject under study. The scanning system is spective path relative to the subject. constructed and arranged to produce a The reconstruction processor accumu- sequence of READ signals which signal lates the data represented by the AVER- the end of each primary time period and AGE INTENSITY signals and processes 25 the beginning of the next succeeding period. the data to, in effect, reconstruct an im- 90 The scanning system also produces POSI- age from the data and produce the RE- TION signals which serve to identify, re- CONSTRUCTED IMAGE signal for oper- lative to the subject, the path traversed ating the imager. by the X-rays during each scan segment. Another feature of the apparatus resides 30 The data processing unit processes the in the use of, in an X-ray tomographic sys- 95 DATA, READ and POSITION signals to tem, a charge pump integrator designed produce an AVERAGE INTENSITY sig- for accommodating analog DATA signals nal representing the average detected X- output from an X-ray detector and con- ray intensity for each X-ray path through verting the DATA signals to a variable 35 the subject. The AVERAGE INTENSITY frequency pulsating signal without incur- 100 signals are, in turn, processed and com- ring appreciable loss of signal information. bined by the data processing unit to pro- The charge pump integrator includes an duce RECONSTRUCTED IMAGE sig- integrator circuit, a threshold level detec- nals which are output to the imager. The tor, an output circuit, and a feedback cir- 40 data processing unit includes a data signal cuit. The integrator circuit responds to 105 processor and a reconstruction processor. the DATA signal and tends to produce an The data signal processor receives READ integrated data signal characteristic of the and DATA signals from the scanning sys- integral of the DATA signal. The threshold tem and produces a first output signal re- level detector is coupled to the output of 45 presentative of the duration of a second- the integrator and produces a threshold de- 110 ary time period (occurring within the prim- tector output whenever the integrator out- ary time period) during which X-rays are put signal exceeds the threshold of the indicated as being detected, and a second threshold level detector. The threshold de- output signal representative of an amount tector output is received by the output 50 of radiation detected in one path during circuit which is coupled to the integrator H3 the secondary time period. The reconstruc- input via the feedback circuit. When- tion processor combines these signals to ever a threshold detector output is pro- produce the AVERAGE INTENSITY sig- duced the feedback circuit momentarily nal which represents the average radiation and abruptly causes the integrator output 55 intensity detected in the path throughout signal to drop below the threshold level of 120 the primary time period. the threshold detector. Accordingly, the The data signal processor preferably threshold detector output is abruptly cut converts the analog DATA signals to vari- off and the output waveforms of both the able frequency pulsating signals. The data integrator circuit and the threshold detec- 60 signal processor also initiates the second- tor are variable frequency pulse trains. The 125 ary time period coincident with the first feedback circuit preferably operates by dis- data pulse occurring within the primary charging a feedback associated time period, and terminates the second- with the integrator circuitry by a predeter- ary time period coincident with the last mined quantum each time a threshold de- 65 pulse occurring within the primary time tector output is detected. This is accom- 130 6 1 564 324 6

plished quite rapidly so that the integrator ably supported on the assembly 16, and a circuit remains substantially continuously subject supporting table 20. responsive to the DATA signal input to it. The supporting assembly 16 includes a A general object of the present inven- main frame unit 22 (see FIGURES 2-4) 5 tion is the provision of a new and improved which is positioned on a floor of a building 70 method and apparatus for processing data in which the system 10 is located, a in a tomographic scanning system. chassis-like housing assembly 24 (FIGURE Other objects and a fuller understand- 1) supported by the main frame for enclos- ing of the invention will be apparent by ing the assemblies 16, 18, and a drive unit 10 referring to the following detailed descrip- 26 for moving the scanning assembly 18 75 tion of a preferred embodiment in con- relative to the supporting assembly 16. junction with the accompanying drawings, The frame unit 22 includes an upright in which plate-like rectangular body 28 having a FIGURE 1 is a partly perspective, partly peripheral stiffening flange portion 30 pro- 15 schematic drawing of a radiation scanning jecting from its circumference transverse 80 system employing the invention; to the plane of the body 28. A mounting FIGURES 2-4 are views of a scanner aperture 32 for the scanning assembly 18 used in the scqpning system of FIGURE 1; is defined centrally in the body plate 28 FIGURE 5 is a circuit schematic of a and the drive unit 26 is supported on the 20 pulse generator in the form of a charge frame adjacent the aperture 32. The drive 85 pump integrator used in the scanning sys- unit 26 preferably includes an electric motor tem of FIGURE 1; 34 drivingly connected to the scanning as- FIGURE 6 is a schematic diagram of a sembly 18 via an associated drive transmis- time counter, a time store, and a data sion 36 (schematically illustrated) of suit- 25 counter and store used in the scanning sys- able construction. 90 tem of FIGURE 1; The scanning assembly 18 is connected FIGURES 7 and 8 are schematic dia- to the frame 22 for rotation about an axis grams of timing generators used in the 40 extending through the aperture 32. The scanning system of FIGURE 1; scanning assembly 18 includes a rotatable 30 FIGURES 9a-9i illustrate examplary support unit 42 journaled to the frame 22, 95 timing signals used in the system of FIG- a framework 44 movably connected to the URE 1; and, support unit 42, and X-ray source and de- FIGURE 10 is a schematic diagram illus- tection assemblies 46, 48, respectively, trating the relationship between primary which are mounted at spaced apart loca- 35 and secondary time periods. tions on the framework 44. The X-ray 100 A transverse section X-ray tomography source and detection assemblies 46, 48 are system embodying the present invention is rotatable with the support unit 42 and illustrated schematically by FIGURE 1. orbit about the axis 40 relative to the frame The tomography system includes an X-ray 22 and are adjustably movable, with the 40 scanning unit 10 for scanning a subject framework 44, relative to the support unit iq5 with X-rays in a given plane and producing 42. electrical output information in the form The support unit 42 includes a tubular of signals descriptive of the scanning pro- cylindrical body 50 extending through the cedure; a processing unit 12 for receiving aperture 32 coaxially with the axis 40 and 45 and processing the scanning unit output supported on the frame 22 by bearings (not 110 signals and producing electrical output sig- shown) for rotation about the axis 40. A nals (RECONSTRUCTED IMAGE sig- radially outwardly extending flange struc- nals) representing a transverse cross-sec- ture 54 is formed at the end of the body tional view of the subject in the plane; and 50 adjacent the frame 22 and the project- 50 an imager 14 which responds to the RE- ing end of the body 50 carries a surround- 115 CONSTRUCTED IMAGE signals to pro- ing flange-like mounting plate 56. The duce an actual image of the transverse mounting plate 56 includes a projecting cross-sectional view. arm 57 which carries a trunnion 58 to The imager 14 can be a display device which the framework 44 is connected. 55 using a cathode ray tube or an image print- The framework 44 is preferably con- 120 ing device, either of which can be oper- structed from lengths of angle irons which ated from the processing unit output. Suit- are welded together at their ends to de- able imager devices are already known and fine a generally flat, open frame, and a therefore the imager 14 is not further illus- bearing arm structure 60 extending from 60 trated or described in detail. the frame to the trunnion 58 so that the 125 A scanning system 10 constructed ac- framework, along with the supported cording to a preferred embodiment of the source and detection assemblies, can pivot invention is illustrated schematically by about the axis 62 of the trunnion relative FIGURES 1-4 and includes a supporting to the supporting unit 42. The rotation 65 assembly 16, a scanning assembly 18 mov- axis 40 extends through the open center of 130 10 10 1,555,872

the framework so that the framework is opening 82 is sufficiently large to surround rotatable about the axis 40 with the unit a human torso aligned with the axis 40. 42. The subject supporting table 20 is A positioning drive motor 64 supported schematically illustrated (FIGURE 1) as 3 by the plate 56 controls the position of the including a wheeled supporting base 83, a 70 framework 44 relative to the supporting pedestal 84, and a subject supporting table plate 56 via a suitable transmission schem- top 85 projecting from the pedestal. A atically illustrated by the reference charac- supine human subject on the table top 85 ter 66. The motor and transmission are is advanced into the opening 82 until a de- 10 constructed and arranged so that when the sired section of the subject's body is dis- 75 framework is positioned as desired with posed in the plane of the X-ray beam 74 respect to the axis 62 it is positively locked and with the rotation axis 40 extending in the adjusted position. through the subject. The X-ray source assembly 46 is schem- The support unit 42 is then rotated about 15 atically illustrated and since it may be of the axis 40 with the X-ray source assembly 80 any suitable or conventional construction 46 operated to scan the subject with the is not described in detail. As illustrated the beam 74. Since the beam 74 is simultane- X-ray source assembly includes an X-ray ously directed to multiple X-ray detector tube head 70 defining an X-ray focal spot units, a single orbit of the source and de- 20 71 on the pivot axis 62 and a collimator 72 tection assemblies about the subject effec- 85 associated with the tube head for directing lively produces a number of scans corres- X-radiation toward the detection assembly ponding to the number of detector units 48 in a fan-shaped beam configuration 74 employed in the detection assembly 48. If which is preferably only about two milli- the reconstructed image resolution pro- 25 meters deep. vided by a single orbit of the source and 90 The X-ray detection assembly 48 is sup- detection assemblies about the subject is ported by the framework 44 opposite the not sufficient, the framework 44 is pivoted source assembly 46 and includes a colli- relative to the unit 42 about the trunnion mator 76, a plurality of X-ray detector axis 62 and another orbit is completed. 30 units 78a-78n, and a supporting apron 79 This procedure is followed until a desired 95 for the collimator and detector units. degree of image resolution can be pro- The collimator 76 is supported in the duced. path of the X-ray beam 74 and defines a During each orbit of the source and de- series of narrow slots corresponding in tection assemblies the scanning unit pro- 35 number to the number of detector units 78 duces DATA, READ and POSITION sig- 100 so that a narrow pencil beam of X-rays im- nals which are transmitted to the process- pinges on each detector unit 78. Any rea- ing unit 12 to enable eventual reconstruc- sonable number of detector units can be iton of an image. employed, and twenty such units have The DATA signals are analog electrical 40 been selected for use in one scanning unit. signals which are continuously produced 105 The detector units may be of any suit- by the photomultiplier tube 81 of each X- able or conventional structure and each ray detector unit and have levels which preferably includes an X-ray exictable vary in direct relationship to the intensity scintillation crystal element 80 which is of the X-ray beam impinging on the re- 45 optically coupled to a photomultiplier tube spective detector unit. The DATA signals 110 81 (see FIGURE 5). are individually output from the respective ITie source assembly 46 and the detec- detector units 78a-78n to the processing tion assemblies 48 are fixedwit h respect to unit 12. each other on the framework 44 so that Each scan of the source and detection 50 as the support unit 42 rotates, the source assemblies is broken up into a succession 115 and detection assemblies orbit about the of individual scan segments during which axis 40 while the beam 74 is continuously X-rays impinging on each detector have directed from the source to the detection traversed a narrow path through the sub- assemblies 48. ject. In the preferred embodiment the 55 The housing assembly 24 (see FIGURE READ signal is constituted by a series of 120 1) is detachably connected to the frame pulses, each of which indicates termina- 22 and extends about the scanning assem- tion of one scan segment and the initia- bly 18 to protect and prevent unauthorized tion of the next succeeding scan segment. access to internal components of the scan- The READ signal is produced by a signal " 60 ning system. generator illustrated in FIGURES 2 and 125 The component parts of the scanning 3 as including an annular disc 86 attached unit 10 are formed to define a central to the body flange 54 and bearing a series generally cylindrical opening 82 extending of substantially equally spaced marks 88 through the scanning unit coaxially with about its periphery and a photosensitive 65 the rotation axis 40. The diameter of the signal producing element 90, such as a 130 14 1,555,872 14_

, mounted on the frame 22 ad- the COUNT) corresponding to the inten- jacent the periphery of the disc 86. The sity received by the associated detector markings 88 are accurately located with unit in the scanning system 10 during each respect to the disc and each other and are primary time period. 5 preferably spaced at one-degree intervals The channel processors lOOa-lOOn are 70 around the axis 40. Whenever one of the identically constructed, and are each marks 88 moves to a predetermined loca- coupled to the associated detector unit 78a- tion relative to the element 90, the ele- 78n in the scanning system 10. For ease ment 90 produces a pulse forming part of of description, only the channel proces- 10 the READ signal pulse train. sor 100a is illustrated and described in de- 75 The angular velocity of the support unit tail. 42 is maintained substantially constant The channel processor 100a includes a throughout the period of a complete or- pulse generator 110 which is responsive to bital scan of the object and accordingly the DATA signals and to timing signals 15 the READ signal pulses are of generally from the timing circuit 102 for producing 80 uniform frequency, successive pulses oc- a train of pulses, called PDATA pulses, curring at about 5-millisecond intervals. whose rate of occurrence is indicative of This interval between the READ signal the intensity of the radiation impinging pulses defines the time period during which upon the associated detector unit. A data 20 the radiation paths for each scan segment counter and store 112 is coupled to the 85 are established. The READ signals thus output of the pulse generator 110 and to provide, in effect, timing signals defining the timing circuit 102 for counting PDATA the end and beginning of successive prim- pulses occurring during the primary time ary time periods during which individual period and generating a COUNT signal 25 scan segments occur. indicative of one less than the number of 90 The POSITION signals are produced by pulses received during the primary time a signal generator 92 which is illustrated period. A time counter 114 is coupled to as associated with the transmission 36. The the pulse generator 110 and to the timing signal generator 92 produces signals which circuit 102 for generating a TIME signal 30 are uniquely indicative of the angular dis- whose value is indicative of the time elaps- 95 placement of the unit 42 from a reference ing between the first and last DATA pulses location at any time. The position signals during the primary time period. A time are thus indicative of the location of the store 116 is coupled to the pulse generator X-ray beam paths relative to the subject 110 and to the time counter 114 for stor- 35 at any time. ing the value of the TIME signal upon each 100 It should be noted that, in the illustrated occurrence of a PDATA pulse subsequent embodiment of the invention, the frame- to the first PDATA pulse during the prim- work positioning motor 64 is controlled ary time period. Accordingly, at the ex- from the processing unit 12. As such, the piration of the primary time period, the 40 position of the framework 44 relative to time store 116 contains a value indicative 105 the axis 62 is always known and thus the of the duration of a secondary time period POSITION signals are effective to accu- beginning with the firstPDAT A pulse and rately locate the X-ray beam paths regard- ending with the last PDATA puls.e occur- less of position adjustments of the frame- ring in the primary time period. 45 work 44. The relationship between primary and 110 Referring to FIGURE 1, the processing secondary time periods is exemplified in unit 12 includes a signal processor 94 for FIGURE 10 where a pair of primary time receiving and processing the READ and periods PTP-1, PTP-2 are illustrated and DATA signals from the scanning system nine and three PDATA pulses are illus- 50 10, and a reconstruction processor 96 trated and occur within the respective 115 coupled to outputs of the signal processor periods. The secondary time periods, STP- 94 and to the POSITION signal output 1, STP-2 associated with the primary time from the scanner 10 and effective to pro- periods are shown to encompass nine and duce RECONSTRUCTED IMAGE sig- three PDATA pulses, respectively, but the 55 nals. first pulse in each case is not counted. 120 The signal processor 94 includes a plur- Referring now to FIGURE 5 the pulse ality of channel processors lOOa-lOOn which generator 110 includes an integrator 120 are each responsive to a respective DATA and a preamplifier stage 122 which ampli- signal, and a timing circuit 102 which is fies current levels of the DATA signal for 60 responsive to the READ signals. The tim- input to the integrator 120. The integrator 125 ing circuit 102 generates a plurality of 120 output is coupled to a threshold com- timing signals, including TSET, TRSET, parator 124 which generates an output BRSET, BTSET, BMXFR, and BMRST, whenever the integrator output exceeds the for operating the channel processors 100a- threshold of the comparator 124. An out- 65 lOOn to accumulate a number (hereafter put circuit 126 receives the comparator 130 9 1 564 324 9

output and is connected to a feedback 143. The value of the capacitor 142 is circuit 128 coupled between the output selected according to the rate at which the circuit 126 and the input of the integrator output pulses from the comparator 124 120. The feedback circuit dispenses an will be generated. For a pulse repetition 5 amount of charge to the integrator 120 for rate of the output pulses of one million 70 causing the integrator 120 to reset each PPS, the capacitor 142 is selected to have time its output signal exceeds the threshold a .001 microfarad value. of the threshold comparator 124. Accord- The threshold comparator 124 includes a ingly, the outputs of the integrator 120 gate 144 having an input bias 146 10 and the comparator 124 are pulsating wave- and an output bias resistor 148. The gate 75 forms and the output circuit 126 produces 144 has an approximate 1.5 volts threshold PDATA pulses. and, whenever the integrator output signal The preamplifier stage 122 includes a exceeds that threshold, the gate 144 pro- dark current adjusting circuit 130, a cali- duces an output on a line 149. 15 bration resistor 132, and a current ampli- The output circuit 126 includes an input 80 fier 134. The dark current adjusting cir- NAND gate 150, an RS flip-flop in the cuit 130 adjusts the level of the DATA form of a pair of cross-coupled NAND signal when no light is incident on the gates 152, 154, and an output NAND gate photomultiplier tube of the associated de- 158. The input NAND gate 150 has one 20 tector 94a. This current is known as the input coupled to the line 149 for receiving 85 "dark" current and is preferably main- the output pulses and has another input tained at a predetermined low level so that coupled to a line 155 via an inverter 156 for the ratio of the maximum DATA signal receiving the TSET timing signal. The gate current to the dark current is 1000:1. If 154 has its noncross-coupled input coupled 25 the dark current should fall below the level necessary to maintain the 1000:1 to a line 157 for receiving the TRST tim- 90 ratio, the dark current adjusting circuit ing signal. The output gate 158 has one 130 automatically adds a restoring amount input coupled to the output of the gate 152 of current to the DATA signal. In the and has another input coupled to a line 30 illustrated embodiment, the dark current 159 via an inverter 160 for receiving the compensation circuit maintains 10 nano- TSET timing signal. The output of the 95 amps as the minimum current flowingou t gate 158 is coupled to the feedback cir- of the current 134. cuit 128 and to an inverter 162 for provid- The adjusting circuit 130 comprises a ing the PDATA pulses on a line 164. 35 136 which is connected Referring to FIGURES 9b and 9c, the across a source of reference potential by a trailing edge of the TRST timing signal is 100 fixed resistor 138. The wiper of the poten- seen to be 100 nanoseconds before the ris- tiometer is connected to a junction J1 ing edge of the TSET timing signal. This through a fixed resistor 140. The amount allows the RS flip-flopo f gates 152, 154 to 40 of dark current flowing out of the cur- be reset to a logic one when the TSET rent amplifier 134 into the junction J1 is timing signal goes to a logic zero and an 105 determined by adjustment of the potentio- output pulse is on the line 149. Because the meter wiper position to control the voltage at the junction Jl. TRST timing signal goes to a logic one 45 The current amplifier 134 includes a state before TSET timing signal goes to a conventional operational amplifier 136 hav- logic one state, the logic one is latched ing its inverting input coupled to the junc- into the RS flip-flops o that a logic zero HO tion Jl and its output connected to the is transferred to the feedback circuit 128 integrator 120 via a resistor 138. The oper- when the TSET timing signal goes to a 50 ational amplifier 136 is provided with bias circuitry for adjusting the output to the logic zero state (i.e„ when the TSET tim- resistor 138 to be in the ratio of 1 volt ing signal goes to a logic one state). output for each microamp input to the Conversely the RS flip-flopi s reset to 115 operational amplifier 136. The resistor 138 a logic zero state in the absence of an out- 55 converts the voltage output from the opera- put pulse on the line 149 when the~TRST tional amplifier 136 into a current output. timing signal goes to a logic zero. This The integrator 120 is conventional and produces a logic one state to the feedback comprises an operational amplifier 140 circuit 128. 120 having a feedback capacitor 142 coupling The feedback circuit 128 comprises a 60 its output and its inverting input terminal. noninverting level shifter circuit 170 con- The inverting input terminal of the opera- nected to the output of the output NAND tional amplifier 140 is also connected to gate 158 and a charge supply circuit 172 the resistor 138 at a junction J2 for re- connecting the output of the level shifter 125 ceiving the amplified DATA signal and 170 to the input of the integrator 120 at 65 producing the integrator output on a line the node J2. 10 1 564 324 10

The level shifter circuit 170 includes one to be generated on the line 171, for- three serially connected voltage divider re- ward biasing the 186. The voltage at sistors 174, 176, 178 connected between the junction J3 increases, reverse biasing positive and negative voltage supplies, and the diode 188 and allowing the capacitor 5 a capacitor 180 connected in parallel with 142 to be charged by the current from the 70 the resistor 174. The input to the level current amplifier 134. shifter circuit 170 is the common connec- The illustrated charge pump amplifier tion of the 174, 178, and the capa- is designed to draw approximately 2 milli- citor 180, and is connected to the output amps of current out of the integrator 120. 10 of the gate 158. The output of the level This is sufficient to allow the output pulses 75 shifter circuit 170 is the common connec- to be generated at a one-million-pulse-per- tion of the resistor 174, 176 and the capa- second rate when 10 microamps flow into citor 180. A logic one voltage at the output the anode of the photomultiplier tube 81. of the gate 158 causes the level shifter cir- Referring now to FIGURE 6 the data 15 cuit 170 to generate a logic one output on counter and store 112 includes a plurality 80 a line 171 of approximately four volts. A of binary counter/output latch stages 200a- logic zero produced at the output of the 200d which are connected as a synchron- gate 158 generates a logic zero on the line ous counter. Each of the counter/output 171 of approximately —0.8 volts. latch stages is a model DM 8554 binary 20 The charge supply circuit 172 includes a counter/latch which is commercially avail- 85 potentiometer 182 which is resistively able from the National Semiconductor Cor- coupled to a junction J3 by a resistor 184, poration. Each stage has respective output is capacitively coupled to circuit ground, lines 201a, 201b, 201c, 201d and produce and is coupled to a negative voltage sup- four bits of the COUNT signal. Each stage 25 ply. A diode 186 is connected to the line has a clock pulse terminal CP connected 90 171 and to the junction J3 and is poled to a junction J4 to which the PDATA such that a logic one signal on the line pulses are coupled via the line 164 through 171 forward biases the diode 186. A diode a pair of inverters. Each of the stages 200a- 188 couples the junctions 32, J3 and is 200d also has counter reset terminals CR 30 poled for current flow from the junction coupled to a junction J5 and has a transfer 95 32 into the junction J3. enable terminal TE coupled to a junction Operation of the feedback circuit 128 is 36. Whenever a pulse is applied to the fundamental to operation of the pulse junction J5, the counter is reset, and when- generator 110 as a charge pump integrator. ever a pulse is applied to the junction J6, 35 Assuming the initial conditions that the in- the count of each respective stage is trans- 100 tegrator 120 is initially discharged, then ferred into its output latch. The junction 36 there is no output on the line 164, and the is coupled to a line 214 by an inverter for line 171 is at logic one which forward biases the diode 186 which in turn reverse biases receiving the BMXFR signal for transfer- 40 the diode 188. Upon the detection of radia- ring the contents of the counters of each tion by the associated radiation detection stage to the respective output latch immed- 105 unit current flowsfro m the current gener- iately before each latch is cleared at the ator 134 through the junction 32 and into expiration of a primary time period. the integrator 120, charging the capacitor Each stage also has a pair of count en- 45 142. As the capacitor 142 charges, the in- able terminals CEP, CET, and a terminal tegrator output signal on the line 143 in- count terminal, TC. The TC terminal of 110 creases negatively in value until it falls the stage 200a is coupled in parallel to the below the threshold of the comparator 124. CEP terminals of the stages 200b, 200c, Upon this condition an output is gener- 200d. The TC terminals of the stages 200b, 50 ated on the line 149 which, upon genera- 200c are respectively coupled to the CET tion of the TSET, TRST" and TSET timing terminals of the stages 200c, 200d. When- 115 signals, produces an output on the line 164. ever the count enable terminals CEP, CET This causes a logic zero to be generated are coupled to logic one, occurrences of on the line 171 which reverse biases the the PDATA pulses at the junction J4 in- 55 diode 186 and allows the diode 188 to be- crement the count of each stage. However, come forward biased. Forward biasing of a logic one is generated at the terminal 120 the diode 188 allows the capacitor 142 to count terminal TC only when a logic 1111 rapidly discharge, increasing the value of state is produced in the respective counter the integrator output signal until it is so that synchronous counter operation is 60 above the threshold of the gate 144. This effected. terminates the comparator output and re- Each of the stages 200a-200d also has 125 sults in completion of a PDATA pulse on an output data enable terminal ODI, and the line 164. Upon completion of the the output data terminals of the stages PDATA pulse, the output circuit 126 and 200a, 200b are coupled to a junction J7 65 the level shifter circuit 170 causes a logic and the output data terminals of the stages 200c, 200d are connected to a junction J8. 130 11 1 564 324 11

Whenever FETCH 1, FETCH 2 pulses are the incrementing of each stage. To this received at the junctions J7, J8, from the end, a gate 227 receives on one input the reconstruction processor 20, the latch count in the respective stages is transferred BTRST signal on the line 226 via an inver- 5 onto the lines 201a-201d for transmission ter 228 and receives the PDATA pulses to the reconstruction processor 20. from the junction J4 on its other input. 70 A time flip-flop 202 has one input coupled Upon coincidence of a logic zero state of by a line 204 through an inverter to the the BTRST signal and a PDATA pulse, the junction J4 and has another input coupled gate 227 generates a logic one via an in- 10 by a line 212 to receive the BMRST tim- verter 229 to the junction Jll. This trans- ing signal. It has one output coupled to the fers the contents of the respective coun- 75 junction J5 via a line 206 and has another ters to the respective output latches of the output coupled to a gate 209 by a line 208. stages 220a-220d each time the counter When the BMRST timing signal goes to a 114 is incremented. 15 logic zero state on the line 212, represent- The stages 220a-220d are reset concur- ing the end of a primary time period, the rently with the stages 200a-200d at the end g0 time flip-flop 202 is reset. This produces of each primary time period. Each logic a logic one at the junction J5, which in state of the BMRST signal causes logic turn resets all the stages 200a-200d. Upon ones to be coupled to the junctions J5, J10, 20 the next occurrence of one of the PDATA which resets the stages. pulses at the junction J4 after the begin- The time store 116 is also shown in 85 ning of a subsequent primary time period FIGURE 6 and comprises a plurality of (i.e. the first pulse in the time period), the four bit registers 230a-230d. Each of the time flip-flop 202 is set via the line 204, registers 230a-230d has its input terminals 25 producing a logic zero at the junction J5 coupled to the respective lines 222a-222d so that the stages 200a-200d count the from the associated stages 220a-220d. A go pulses, other than the first pulse, in that clock terminal CK on each of the registers time period. The zero at the junction 5 230a-230d is commonly connected to the also produces a logic one on the line 208 junction J6 to allow the stages 230a-230d 30 and enables the gate 209. The gate 209 has to receive and store the TIME signal on another input coupled to a line 224 for re- the lines 222a-222d upon generation of the ceiving the BTSET timing signal, so that 95 the logic one on the line 208 allows the BMXFR signal at the expiration of each gate 209 to produce a clock signal on a primary time period. The stored TIME 35 line 210 according to occurrences of the signal is output onto respective output lines BTSET timing signals. This is transferred 232a-232d whenever a reductive output by the line 210 to a junction J9 for use by enable OE terminal is pulsed. To this end, jqq the time counter 114. a line 234 is connected to the output en- The time counter 114 includes a plur- able terminals OE of the sections 230a- 40 ality of binary counter/output latch stages 230b and a line 236 is coupled to the out- 220a-220d. Each of these stages 220a-220d put enable OE terminals of the sections has terminals respectively identical to the 230c-230d. The lines 234, 236 are sequen- 105 stages 200a-200d and are interconnected in tially pulsed by FETCH 3, FETCH 4 sig- synchronous counter fashion. The node J9 nals from the reconstruction processor 20 45 corresponds to the node J4, a node J10 for retrieving the stored data TIME _ in corresponds to the node J5 and a node Jll sequential bytes, with each byte consist- corresponds to the node J6 for operating ing of eight bits. 110 the respective stages 220a-220d as a syn- The timing circuit 102 includes a non- chronous counter. However, contrary to buffered signal generator 240 which is 50 the connection of the stages 200a-200d, the shown in FIGURE 7 and a buffered sig- output data terminals ODI of the stages nal generator 242 which is shown in FIG- 220a-220d are conditioned such that the URE 8. The nonbuffered signal genera- contents of the output latches are auto- 115 matically introduced onto the respective tor 240 generates the TSET, TSET, and 55 sets of output lines 222a-222d. TRST timing signals on the lines 155, 159, More specifically, the clock signals from 157 respectively. The buffered signal gener- the gate 209 are coupled to the junction ator 242 is responsive to the TSET timing J9 for incrementing the stages 220a-220d signal on the line 155 and to the READ 120 whenever the time flip-flop20 2 is set by signal from the scanner 16 for producing 60 the occurrence of the first PDATA pulse occurring during a primary time period. the BTRST, BTSET, BMXF'R, and BMRST The contents of the stages 220a-220d is timing signals on the lines 226, 224, 214, transferred to the respective output latches and 212 respectively. as the TIME signal upon the occurrence Referring to FIGURE 7, the nonbuf- 125 65 of each PDATA pulse immediately after fered signal generator 240 includes an os- 12 1 564324 12

cillator 244, a counter 246 coupled to the through fifthpulse s of each ten-pulse cycle oscillator 244 via a line 248, a decoder 250 and is a logic one during the sixth through coupled to the counter 246 via a set of ninth pulse of each ten-pulse cycle. As 65 lines 252, and a pair of latch circuits 254, shown by a comparison of FIGURES 9b, 5 256 connected to outputs of the decoder 9c, there is a 100 nanosecond duration 250. which the TSET and TRST signal are both The oscillator 244 is a 10 megahertz cry- logic zeros. This allows the RS flip-flop stal oscillator which is commercially avail- configuration of the gates 152, 154 to pro- 70 able from Motorola, Inc., under the desig- perly latch upon the occurrence of one of 10 nation K1091A. The oscillator generates a the output pulses on the line 149. ten-megahertz waveform on the line 248 Referring now to FIGURE 8, the buf- for periodically incrementing the counter fered signal generator 242 includes a pair 246. FIGURE 9a illustrates the ten-mega- of one-shot pulse generators 260, 262 re- 75 hertz waveform. sponsive to the TSET signal on the line 155 15 The counter 246 is any suitable decade counter such as a model 74190 from Texas for respectively generating the BTRST Instruments, Inc. The counter 246 gener- timing signal and the BTSET timing signal ates signals on the lines 252 representative on the lines 226, 224 respectively. The of the binary state of the counter 246 as BTRST~and BTSET timing signals are ef- 80 20 it periodically counts between zero and fectively two-phase clock signals operat- nine. ing at a one-megahertz frequency and hav- The decoder 250 is a dual 1:4 decoder ing a 25% duty cycle. FIGURES 9d and which is operatable as a 2: 8 decoder.^ One 9e show the BTRST and the BTSET signals such decoder is commercially available as being relatively phase displaced by 250 85 25 from Texas Instruments, Inc. as Model nanoseconds. No. 74155. The decoder 250 has its input The one-shot pulse generators 260, 262 terminals coupled to the lines 252 and has are commercially available from Texas In- a set of output lines 258 coupled to the struments, Inc. as a model 74221 dual one- latch circuits 254, 256. As the counter 246 shot generator having one generator trig- 90 30 counts between 0 and 9, the decoder 250 gered by positive-going pulses and one sequentially energizes ones of the lines generator triggered by negative-going 258 every 100 nanoseconds, with a period pulses. Each pulse generator is voltage of 300 nanoseconds lapsing between the 111 output state and the subsequent 000 biased to produce the BTRST signal as a 35 output state. 250-nanosecond pulse upon the leading 95 The latch circuit 254 is couped to the de- edge of each TSET pulse, and to produce coder 250 via the lines 258 representing the BTSET signal as a 250-nanosecond the 001 and the 110 decoded states for pulse upon the trailing edge of each TSET pulse. generating the TSET and the TSET output The buffered signal generator 242 also 100 40 signals respectively on the lines 155, 159. includes a master buffer section which is The latch circuit 254 is set via the occur- coupled to the scanner 16 for producing rence of the 001 state to produce the TSET signal and is reset by the occurrence of the the BMXFR and the BMRST signals to be indicative of the end of each primary 110 state to produce the TSET signal. time period during which the average in- 105 45 The latch circuit 256 is coupled to the tensity signal is to be calculated. decoder 250 via the lines 258 to produce The master buffer section includes a gate the TRST signal on the line 157. The latch 264 which is responsive to the BTSET tim- circuit 256 is set by the occurrence of the ing signal and to the READ signal on a 110 state and is reset via the occurrence line 266 from the scanner 10. A latch 268 110 is coupled to the gate 264 by a line 274 50 of the 000 state to produce the TRST and generates an ENABLE signal on a signal. line 276 upon the occurrence of the BTSET Waveforms 9a-9c in FIGURE 9 illus- signal and the READ signal. trate generation of the TSET and TRST A counter 270 and a decoder 272 are 115 signal. FIGURE 9a depicts the output of coupled to the line 276. A line 277 con- 55 the oscillator 244 as a ten megahertz square nects the one-shot pulse generator 262 to wave. The TSET signal is shown in FIG- the counter 270 so that, upon generation URE 9b as a logic one during the first of the ENABLE signal on the line 276, the through fifth pulses of each ten-pulse counter 270 is incremented through four 120 cycle, and is a logic zero during the six states upon four occurrences of the BTSET 60 through tenth pulses of each ten-pulse timing signal. The decoder 272 is coupled cycle. FIGURE 9c shows the TRST sig- to the counter 270 via a pair of lines 280 nal as a logic zero during the tenth for energizing one of its output lines 282a- 13 1 564 324 13

282d for each of the four states of the curring within a secondary period can be counter 270. A gate 284a is coupled to the accurately determined, the number of the line 282a and to the line 226 for produc- intensity representing pulses can be pre- 65 ing the BMXFR signal which is inverted cisely averaged to give an indication of the 5 and suppied on the line 214. average intensity of the beam impinging A gate 284b is coupled to the line 282b upon the detector. The radiation scanning and to the line 226 for producing the system also features a charge pump inte- BMRST signal on the second count of the grator. The charge pump integrator accu- 70 counter 270. The BMRST signal is inverted rately generates the intensity representing 10 and supplied to the line 212. pulses with an extraordinarily high degree A gate 284c is coupled to the line 282c of efficiency and without complicated and and to the line 226 for producing a data expensive electronics. interrupt, DINT, signal which is coupled Although the invention has been de- 75 to the reconstruction processor 20. The scribed in its preferred form with a cer- 15 DENT signal is indicative that the data tain degree of particularity, it is under- accumulated in the signal processor 94 is stood that the present disclosure of the ready to be transferred to the reconstruc- preferred form has been made only by way tion processor 96. The DINT signal is of example. Numerous changes in the de- 80 generated via the third count of the coun- tails of the measuring and processing 20 ter 270. unit, as well as applications other than The latch 268 has an input coupled via with a tomographic scanning system, may the line 282d to the decoder 262 for being be resorted to. reset upon the fourth count of the coun- WHAT WE CLAIM IS: — 85 ter 270. This removes the ENABLE sig- 1. A radiation detection system includ- 25 nal from the line 276 and clears the coun- ing means to produce a series of pulses ter 270, awaiting the expiration of the next having an instantaneous frequency which primary time period. varies in relation to detection radiation intensity; and circuitry for determining the 90 The READ, MBXFR, BMRSfr and rate of occurrence of said pulses during a predetermined period, the circuitry com- DINT signals are shown in FIGURES 9f- prising firstcountin g means to count pulses 30 9i. The READ signal is generated by the which occur during the predetermined scanner 10 as a pulse of approximately period, and having an output representa- 95 two-microseconds duration at the end of tive of one less than the number of pulses each primary time period. The occurrence occurring during the predetermined period, of the first BTSET signal causes the timing means to provide a time signal hav- 35 BMXFR signal to be generated on the line ing a value representative of the time 214 which, in turn, causes transfer of the elapsing between the first and last pulses iqo TIME signals into the time store 116 and occurring during the predetermined time causes the COUNT signals to be latched period, and means responsive to the out- into the output latch of the stages 200a- put of the first counting means and the 40 200d. The ocurrence of the second time signal to produce an output indica- BTSET signal causes the BMRST signal to tive of the average radiation intensity de- 105 be generated on a line 212, causing reset tected during the period. of the time counter 114 and of the stages 2. A system as claimed in claim 1, 200a-200d. The occurrence of the third wherein the timing means includes a timer 45 BTSET signal causes generation of the responsive to the pulses to generate the time signal to characterise the duration 110 DINT signal, and the occurence of the of the elapsed time between the first and fourth BTSET signal resets the latch 268. last pulses which occur during the pre- It is thus apparent that an improved determined period. radiation measuring and processing unit 3. A system as claimed in claim 2, 50 has been provided for a radiation scanning wherein the timing means further includes 115 system. The average intensity of the beam second counting means coupled to the of radiation as it impinges upon the radia- timer for providing the time signal; and tion detector during a primary time period storage means coupled to the second count- is accurately determined by computing the ing means for storing the value of the 55 average rate of a train of intensity repre- time signal whenever the first counting 120 senting pulses occurring during a second- means is incremented. ary time period whose duration is de- 4. A system as claimed in claim 3, fur- termined according to the occurrences of ther including data storage for storing the the first and last pulses within the prim- count accumulated by the first counting 60 ary time period. This allows the duration means during the predetermined period; 123 of the secondary period to be accurately and means for resetting the first counting determined. Since the number of pulses oc- means upon the expiration of the predeter- 14 1564324 14

mined period. to the intensity of the beam as it emerges 5. A system as claimed in any preced- from a subject and impinges upon the re- ing claim, further including a source of ceiving means; and means supporting the penetrative radiation for producing a beam X-ray source and/or the beam receiving 5 of radiation; source supporting means for means for motion along a predetermined 70 moving the source along a predetermined path relative to the subject with the beam path relative to a subject so that the sub- being scanned through the subject during ject is scanned by the beam; beam receiv- the motion. ing means comprising a detector disposed 10. A system as claimed in Claim 9, 10 for receiving the beam of radiation emerg- wherein the counting means produces a 75 ing from the subject during the scanning count signal indicative of the number of motion; and period defining means for pulses counted, the system further includ- establishing the time period during which ing time stores and data stores for storing a predetermined amount of scanning the values of the time signal and the count 15 motion occurs. signal after each of a sequence of said pre- 80 6. A system as claimed in Claim 5, determined periods, the means to produce wherein the source of penetrative radia- an output indicative of the average radia- tion comprises an X-ray source; the detec- tion being responsive to the stored time tor comprises signal producing means for signal and to the stored count signal to 20 producing an analog signal which varies in produce the average intensity signal. 85 accordance with variations in intensity of 11. A method of detecting the intensity radiation impinging on the detector; and of penetrative radiation passing through a the beam receiving means further com- subject during a predetermined period, in- prises a charge-pump integrator circuit for cluding detecting the intensity of radia- 25 producing the pulses in response to the tion emerging from the subject; producing 90 analog signal, the charge-pump integrator a train of signal pulses the rate of occur- comprising an amplifier having an input rence of which varies according to the de- and an output coupled together by a capa- tected radiation intensity; counting pulses citor, the integrator being responsive to which occur during the predetermined 30 the analog signal to produce an integrated period and producing a pulse count repre- 95 data signal at its output, a threshold level sentative of one less than the number of detector connected to the integrator output pulses occurring during said predetermined to produce an output signal in response period; timing an interval defined by the to integrated data signal levels exceeding a time elapsing between the first and last 35 predetermined threshold value, and feed- pulses occurring during the predetermined 100 back means coupling the level detector time period and producing a time signal the output to the integrator input and opera- value of which represents said interval; and tive to discharge the capacitor by a pre- dividing the number of pulses counted by determined amount in response to produc- the time signal value to produce an output 40 tion of a level detector output signal, signal representative of the average radia- 105 whereby the integrator output signal value tion intensity detected during said prede- is reduced below the threshold and an out- termined period. put pulse from the level detector is pro- 12. A method as claimed in claim 11, duced, the output pulse forming one pulse wherein the timing of an interval includes 45 of said series of pulses. storing an indication of time elapsing be- 110 7. A system as claimed in Claim 5 or tween the occurrence of the first and the Claim 6, wherein the radiation source pro- last pulses in the period; and the counting duces a beam of X-radiation, the system of the pulses includes counting the second further including means for producing a and any subsequent pulses occurring with- 50 tomographic image of the subject based in in the period. 115 part on the average radiation intensity out- 13. A method as claimed in claim 12, put. wherein the timing of an interval defined 8. A system as claimed in Claim 5, by the first and last pulses includes initiat- Claim 6 or Claim 7, further including posi- ing timing upon the occurrence of the first 55 tion indicating means for generating index pulse occurring within the predetermined 120 signals indicative of the traversal by said period. source through each of a series of seg- 14. A method as claimed in claim 11, ments constituting said path, the index sig- claim 12, or claim 13, further including nals being effective to determine the ex- producing a beam of penetrative radiation 60 tent of said predetermined period. from a radiation source and directing the 125 9. A system as claimed in Claim 1, fur- beam through the subject; and scanning ther including an X-ray source for pro- the beam through the subject via a pre- ducing a beam of radiation; beam receiv- determined scan path by effecting relative ing means disposed for receiving the beam; motion between the subject and the radia- 65 means to produce said pulses in response tion beam. 130 15 1 564 324 15

15. A method as claimed in any one verting the analog signals to the pulse train of Claims 11-14, further including scan- with the pulse train frequency varying in ning the beam at a predetermined rate; relation to variations of the analog signal 20 producing period defining signals indicat- level. 5 ing the beginning and ending of each scan 17. A system as claimed in Claim 1 and path segment of a succession of scan path substantially as hereinbefore described segments; and performing the counting with reference to the accompanying draw- and timing steps after production of each ings. 25 period defining signal indicating the begin- 18. A method as claimed in Claim 11, 10 ning of a scan path segment and prior to and substantially as hereinbefore described each time period signal indicating the end ofwit h reference to the accompanying draw- the scan path segment. ings. 16. A method as claimed in any one For the Applicants: of Claims 11-15, wherein producing the GILL JENNINGS & EVERY 15 train of signal pulses comprises producing Chartered Patent Agents analog signals having levels proportional to 53-64 Chancery Lane the intensity of detected radiation and con- London WC2A 1HN

Printed for Her Majesty's Stationery Office by The Tweeddale Press Ltd., Berwick-upon-Tweed, 1980. Published at the Patent Office, 25 Southampton Buildings, London, WC2A 1AY, from which copies may be obtained. 1564324 COMPLETE SPECIFICATION - -ypr-rr This drawing is a reproduction of 6 2>Mttlb the Original on a reduced scale Sheet 1

Lu 1564324 COMPLETE SPECIFICATION

6 SHEETS drawing Is a reproduction of the Original on a reduced scale Sheet 2 1564324 COMPLETE SPECIFICATION 6 SHEETS drawing is a reproduction of the Original on a reduced scale Sheet 3 1564324 COMPLETE SPECIFICATION - This drawing is a reproduction of

D anttli the Original on a reduced scale Sheet 4

DATA C0lMTtS_/)M0 STORSr

TO eecousTeucriotJ -18? mcessozza 1564324 COMPLETE SPECIFICATION

6 SHEETS 77*f$ drawing is a reproduction of 0 the Original on a reduced scale Sheet 5

??6 07K57T BTRST

BTSBT.

,270

6i 5 - z,coimec. lit 149b D 0MX$B Ai ft 214

BMBST ZiZ

ZcW^li" wrr kK> TO 26 5 BBKONSrmaM

Fig. 8 1564324 COMPLETE SPECIFICATION

I clicctc This drawing is a reproduction of the Original on a. reduced scale Sheet 6

o -Q O "O