GB1601146A - Raster scanner - Google Patents

Description

(54) RASTER SCANNER (71) We, XEROX CORPORATION of Xerox Square, Rochester, New York, United States of America, a corporation organized under the laws of the State of New York, United States of America, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates to raster input scanners and more particularly to, raster input scanners having multiple linear arrays, Scanning technology has progressed rapidly in recent years and today arrays of fairly substantial linear extent are available for use in raster scanners. Indeed, the linear extent of new arrays is in some cases many times the linear extent of earlier array designs. However, the length of even these recent array designs is still not sufficient to enable a single array to span the entire width of the normal-sized line, e.g. 81E inches. Further, it appears improbable that arrays of sufficient length will be developed in the foreseeable future since fabrication of such arrays would appear to require a major breakthrough in semiconductor fabrication technology.

As a result, raster input scanners are forced to rely on shorter arrays and must, therefore, employ a multiplicity of arrays if the entire line is to be scanned in one pass. This raises tile question of how to place the arrays so as to cover the entire line yet provide data representative of the line which is free of aberrations at the array junctures. Recently, interest has been expressed in optically-butted arrays. However, optical and optical/mechanical arrangements often experience difficulty in meeting and maintaining the tight tolerances necessary for aberration-free scanning, particularly in operating machine environments.

Attention is drawn to the disclosure in related application 17247/78 (Serial Number 1 601 145). According to the invention, there is provided an apparatus for scanning an image as claimed in the appended claims.

An embodiment of the invention will now be described by way of example with reference to the accompany drawings in which: Figure 1 is an isometric view showing a raster input scanner incorporating the multiple array arrangement; Figure 2 is a schematic illustrating an exemplary array disposition; Figure 3 is a schematic view of a scanner operating control; Figure 4 is a schematic representation of a memory buffer for temporarily storing image data; Figure 5 is a schematic illustration of a data mapping arrangement to avoid bit shifting on readout from the temporary memory buffer of Figure 4; Figure 6 is a schematic view showing a data readout system; Figure 7 is a schematic illustration of the data readout with crossover and removal of redundant data; Figure 8 is a schematic view illustrating a modified array wherein the center-to-center distances between the photo-sensitive elements ofa portion of one array are changed to provide a vernier useful for aligning arrays; Figure 9 is a schematic view of an alternate array configuration wherein a bridging array is employed to effect continuity between adjoining arrays, and Figure 10 is a schematic view of another alternative array configuration wherein a bridging array is combined with a standard array to form a unitary structure, the photosensitive elements of the bridging array being on different center-to-center distances to provide a vernier.

Referring to Figure 1, an exemplary raster input scanning apparatus 10 is thereshown.

Scanning apparatus 10, as will appear more fully herein, scans an original document 12 line-by-line to produce a video signal representative of the original document 12. The video signal so produced may be thereafter used to reproduce or duplicate the original 12, or stored in memory for later use, or transmitted to a remote source, etc.

Scanning apparatus 10 comprises a box-like frame or housing 14, the upper surface of which includes a transparent platen section 16 on which the original document 12 to be scanned is disposed face down. A displaceable scanning mechanism, designated generally by the numeral 18, is supported on frame 14 below platen 16 for movement back and forth underneath the platen 16 and the original document 12 thereon in the Y direction as shown by the solid line arrow in Figure 1.

Scanning mechanism 18 includes a carriage 20 slidably supported upon parallel rods 21, 22 through journals 23. Rods 21, 22, which parallel the scanning direction along each side of platen 16, are suitably supported upon the frame 14.

Reciprocatory movement is imparted to carriage 20 by means of a screw type drive 24. Drive 24 includes a longitudinally-extending threaded driving rod 25 rotatably journalled on frame 14 below carriage 20. Driving rod 25 is drivingly interconnected with carriage 20 through a cooperating internally-threaded carriage segment 26. Driving rod 25 is driven by means of a reversible motor 28.

A plurality of linear photosensitive arrays 1, 2, 3, 4 are carried on plate-like portion 35 of carriage 20. Arrays 1, 2, 3,4 each comprises a series of individual photosensitive picture elements or pixels 40 arranged in succession along the array longitudinal axis. The arrays scan the original document 12 on platen 14 as scanning mechanism 18 moves therepast, scanning movement being in a direction (Y) substantially perpendicular to the array longitudinal axis (X). As best seen in Figure 2, the arrays 1,2,3,4 may, due to the difficulty in accurately aligning the arrays one with the other, be offset from one another in the direction of scanning movement (the Y direction) so that they scan overlapping parts of the image area. To accommodate the relatively short length of the individual arrays, the arrays overlap. In the exemplary illustration, the end portion of arrays 2, 1, 4 overlap the leading portion of the succeeding arrays 1,4, 3 when looking from left to right in Figure 2 along the X direction.

As will be understood, the length of the individual arrays 1,2, 3, 4 may vary with different types of arrays and from manufacturer to manufacturer. As a result, the number of arrays required to cover the entire width of the original document 12 may vary from that illustrated herein.

Photosensitive elements or pixels 40 of arrays 1, 2,3,4 are normally silicon, with carrier detection by means of photo-transistors, photodiode-MOS amplifiers, or CCD detection circuits. One suitable array is the Fairchild CCD 121-1728 pixel 2-phase linear array manufactured by Fairchild Corporation. As described, arrays 1,2,3,4 are offset from one another in the scanning or sagittal direction (Y direction) but with an end portion of each array overlapping the leading portion of the next succeeding array to form in effect a composite unbroken array.

To focus the image onto the arrays 1,2,3, 4, a lens 43 is provided for each array. Lenses 43 are supported on carriage 20 in operative disposition with the array 1, 2, 3, 4 associated therewith. Mirrors 44, 45 on carriage 20 transmit the light images of the original via lenses 43 to arrays 1, 2,3,4. Lamp 48 is provided for illuminating the original document 12, lamp 48 being suitably supported on carriage 20. Reflector 49 focuses the light emitted by lamp 48 onto the surface of platen 16 and the original document 12 resting thereon.

In operation, an original document 12 to be scanned is disposed on platen 16. When scanning mechanism 18 is actuated motor 28 operates driving mechanism 24 to move carriage 20 back and forth below platen 16. Lamp 48 is energized during the scanning cycle to illuminate the original document 12.

To correlate movement of carriage 20 with operation of arrays 1, 2, 3, 4 an encoder 60 is provided. Encoder 60 generates timing pulses proportional to the velocity of scanning mechanism 18 in the Y direction. Encoder 60 includes a timing bar 61 having a succession of spaced apertures 62 therethrough disposed along one side of the path of movement of carriage 20 in parallel with the direction of movement of carriage 20. A suitable signal generator, in the form of a photocell-lamp combination 64, 65 is provided on carriage 20 of scanning mechanism 18 with timing bar 61 disposed therebetween.

As carriage 20 of scanning mechanism 18 traverses back and forth to scan platen 16 and any document 12 thereon, photo-cell-lamp pair 64, 65 of encoder 60 moves therewith. Movement of the photocell-lamp pair 64, 65 past timing bar 61 generates a pulse-like output signal in output lead 66 of photocell 64 directly proportional to the velocity of scanning mechanism 18.

As can be envisioned by those skilled in the art, supporting arrays 1, 2, 3, 4 in exact linear or tangential alignment (along the X-axis), and maintaining such alignment throughout the operating life of the scanning apparatus, is extremely difficult and somewhat impracticable.

To obviate this difficulty, arrays 1, 2, 3, 4 are initially mounted on carriage 20 in substantial tangential alignment. As can be seen in the exemplary showing of Figure 2, this nevertheless often results in tangential array misalignment along the x-axis. If the disposition of the arrays 1, 2, 3, 4 is compared to a predetermined reference, such as the start-of-scan line 101 in Figure 2, it can be seen that each array 1,2,3, 4 is displaced or offset from line 101 by some offset distance dl, d2, d3, d4, respectively. As will appear more fully herein, the individual offset distance of each array 1, 2, 3, 4 is determined and the result programmed in an offset counter 120 (Figure 3) associated with each array. Offset counters 120 serve, at the start of the scanning cycle, to delay activation of the array associated therewith until the interval dl, d2, d3, d4, therefor is taken up.

Referring to Figure 3, the pulse-like signal output of encoder 60 which is generated in response to movement of carriage 20 in the scanning direction (Y-direction), is inputted to a phase-locked frequency multiplier network 100. Network 100, which is conventional, serves to multiply the relatively low frequency pulse-like signal input of encoder 60 to a high frequency clock signal in output lead 103.

Feedback loop 104 of network 100 serves to phase lock the frequency of the signal in lead 103 with the frequency of the signal input from encoder 60.

As will be understood, changes in the rate of scan of carriage 20 produce a corresponding change in the frequency of the pulse-like signal generated by encoder 60. The frequency of the clock signal produced by network 100 undergoes a corresponding change. This results in a high frequency clock signal in output lead 103 directly related to the scanning velocity of carriage 20, and which accommodates variations in that velocity.

The clock signal in output lead 103 is inputted to programmable multiplexer 106.

The output of a second or alternate clock signal source, such as crystal controlled clock 108, is inputted via lead 109 to multiplexer 106. Multiplexer 106 selects either network 100 or clock 108 as the clock signal source in response to control instructions (CLOCK SELECT) from a suitable programmer (not shown). The selected clock signal appears in output lead 111 of multiplexer 106.

An operating circuit 114 is provided for each array 1, 2, 3, 4. Since the circuits 114 are the same for each array, the circuit 114 for array 1 only is described in detail. It is understood that the number of circuits 114 is equal to the number of arrays used.

Operating circuit 114 includes a line transfer counter 115 for controlling the array imaging line shutter or sample time for each scan.

Counter 115 is driven by the clock signal in output lead 111 of multiplexer 106. It is understood that where the signal input to counter 115 comprises the clock signal produced by network 100, array sample size remains constant irrespective of variations in the velocity of carriage 20. In other words, where carriage 20 slows down, array shutter time becomes longer. If carriage 20 speeds up, array shutter time becomes shorter.

Initial actuation of line transfer counter 115 is controlled by the offset counter 120 associated therewith. Offset counter 120, which is driven by the clock siganl in output lead 111, is preset to toll a count representing the time interval required for array 1 to reach start-ofscan line 101 following start up of carriage 20.

On tolling the preset count, offset counter 120 generates a signal in lead 122 enabling line transfer counter 115.

It will be understood that the offset counters 115 associated with the circuits 114 for the remaining arrays 2, 3, 4 are similarly preset to a count representing the distance d2, d3, d4, respectively by which arrays 2, 3, 4 are offset from start-of-scan line 101.

Referring particularly to Figure 2, each array 1,2,3,4 scans a portion of each line of the original document 12, the sum total of the data (less overlap as will appear more fully herein) produced by arrays 1, 2, 3, 4 representing the entire line. Preferably, arrays 1, 2, 3, 4 are of the same size with the same number of pixels 40. As described, the line transfer counters 115 of circuits 114 control the array imaging line shutter time for each scan, counters 115 being preset to activate the array associated therewith for a preselected period for this purpose. Scanned data from the arrays 1, 2, 3,4 are clocked out by clock signals derived from a suitable pixel clock 118.

Sampled analog video data from the artays 1, 2, 3, 4 are fed to a suitable video processor 148 which converts the video signals to a binary code representative of pixel image intensity.

The binary pixel data from processor 148 are mapped into segments or words by Pixel Data Bit Mapper 149 for storage in offset relation in RAM 175 as will appear. Bit Mapper 149 is driven by clock signals from pixel clock 118.

Data from Bit Mapper 149 are passed via data bus 174 to RAM 175 where the data are temporarily stored pending receipt of data from the array which last views the line. In the exemplary arrangement illustrated, the last array would be array 4.

Multiplexer 150 may be provided in data bus 174 to permit data from other sources (OTHER DATA) to be inputted ot RAM 175.

The binary data are stored in sequential addresses in RAM 175 (see Figure 4, the data being addressed into RAM 175 on a line-by-line basis by the RAM address pointers 165 through Address Bus 180. The clock signal output from pixel clock 118 is used to drive address pointers 165. Line scan counter 170, which is driven by the output from line transfer counter 115, controls the number of full scan lines that will be stored in RAM 175 before recycling. The output of counter 170 is fed to RAM Address pointer 165 via lead 119. It is understood that line scan counters 170 are individually preset to reflect the degree of array offset in the Y-direction.

Ram 175 provides a buffer for scanned data from each array, RAM 175 buffering the data until a full line is completed, following which the data are read out. A suitable priority encoding system (not shown) may be used to multiplex the data input from arrays 1, 2, 3, 4 with the address associated there with. Ram 175 has inlets and outlets for input and output of data thereto.

Since the degree of misalignment of arrays 1, 2, 3,4 in the Y-direction may vary, the storage capacity of RAM 175 must be sufficient to accommodate the maximum misalignment expected. A worst case misalignment is illustrated in Figure 4 wherein it is presumed that arrays 1, 2,3,4 are each misaligned by a full line. In that circumstance, and presuming scanning of line 1 is complete, RAM 175 then stores the line data for lines 1, 1l' 12, 13, 14 from array 1,lines 1,1112,13 from array 2, lines 1,11,12 from array 3, and lines 1, ill from array 4. The blocks of binary data that comprise the completed line 1 are in condition to be read out of RAM 175. In the above example, an extra line of data storage is provided.

Line scan counters 170 are recycling counters which are individually preset for the number of lines of data to be stored for the array associated therewith. As a result, address pointers 165 operate in round robin fashion on a line-by-line basis. On reaching a preset count, the signal from counters 170 recycle the address pointer 165 associated therewith back to the first storage line to repeat the process. It is understood that prior thereto, that portion of RAM 175 has been cleared of data.

As described, data from video processing hardware 148 are stored temporarily in RAM 175 pending completion of the line. In placing the data in RAM 175, the data are preferably mapped in such a way as to avoid the need for subsequent data bit shifting when outputting the data. Referring to Figure 5, wherein mapping of pixel data from arrays 1, 2 is illustrated, data from an earlier array (i.e. array 1) are mapped by Pixel Data Bit Mapper 149 (Figure 3) into segments or words 180 before being stored in RAM 175. The first pixel (P1 - 1) of the array within the array overlap 181 is mapped into a known bit position within the segment or word 180 at the point of overlap.

At the end-of-line transfer, the first pixel (P1 - 2) of the succeeding array (i.e. array 2) is clocked into the bit position (P1 - 1) of the first overlapped pixel of the previous array.

This correlates the first overlapping pixel (P1 - 2) of the succeeding array (i.e. array 2) with the first overlapped pixel (Pi - 1) of the preceding array (i.e. array 1). Crossover from one array to the succeeding array on data readout may then be effected without the need to shift bits.

Referring now to Figures 6 and 7, video data held in RAM 175 are read out to a user (not shown) via RAM output bus 176, in both tangentially and spatially corrected form, lineby-line, through output channel 200. Data readout is controlled by a microprocessor, herein CPU 204 in accordance with address program instructions in memory 206. CPU 204 may comprise any suitable commercially available processor such as a Model 6800 manufactured by Motorola, Inc.

The address program instructions in memory 206 include a descriptor list 207. List 207 contains information identifying the number of bits to be read out (Nn), the address of the first word (A), and other user information (U). The DATA OUT address information is fed to address multiplexer 208 via address bus 209.

As described heretofore, exact tangential alignment and end-to-end abutment of multiple arrays is difficult to achieve. In the arrangement shown, sagittal misalignment (in the Y direction) among the arrays is accommodated by offset counters 120 of the individual array operating circuits 114. The need to abut the arrays accurately end-to-end is obviated by overlapping the areas scanned by the arrays.

As a result of the above, the sequence in which video data are inputted to RAM 175 offsets sagittal misalignments between the several arrays. By outputting the data from RAM 175 on a line-by-line basis, the lines are reconstructed without sagittal misalignment.

Due to the overlapping areas scanned by arrays 1, 2, 3, 4, data within the overlapping areas are redundant. To obviate this and provide a complete line of data without repeated or redundant portions, bit crossover or readout within the overlapping regions is used.

Referring now to the embodiment shown in Figure 7, data bit crossover within the overlapping portions of arrays 1, 2, 3, 4 is effected by an algorithm which picks a last cell in one array to be sampled within the overlapped region, and the first contiguous cell in the succeeding array. In the descriptor list 208 illustrated in Figure 7, the total bit output from the first array is N1 bytes + nl bits with the bit output from the second array N2 bytes - n2 bits. In the example shown in Figure 7, crossover from array 2 to array 1 is effected between bit 4 and bit 5.

In the arrangement described heretofore, the center-to-center distance between successive photosensitive elements or pixels 40 is constant. Referring to Figure 8, wherein like numerals refer to like parts, a pair of arrays 300,301 is there shown with the end portions overlapped. The photosensitive elements or pixels 40 that comprise arrays 300, 301, except for the end 308 of array 300, are on normal center-to-center distances d. The photosensitive elements 40' in the end 308 of array 300 are set on a slightly-reduced center-to-center distance d'. The reduction in center-to-center distances between the photosensitive elements 40' in end 308 of array 300 provides in effect a vernier scale which normally provides at least one point where opposing arrays are in alignment irrespective of the degree of overlap between the arrays. In the exemplary arrangement shown, the end of photosensitive element 40 - 8 of array 301 is in substantial alignment with the start of photosensitive element 40' - 5 of array 300, and crossover would be set at this point.

It will be understood that visual identifica tion of the individual photosensitive elements or pixels 40, 40', to determine the optimum crossover point, may be made through microscopic examination of the arrays. It is further understood that once the optimum crossover point is determined, the descriptor list of memory 206 (Figures 6,7) is programmed to provide crossover from pixels 40 - 8 of array 301 to pixel 40' - 5 of array 300 on readout.

While the vernier scale is illustrated as being at one end 308 of array 300 only, it is understood that vernier scales may be provided at both ends of the array. In that event, in a scanning arrangement employing four arrays such as shown in Figure 2, array 1 may have a vernier scale at one end only, with remaining arrays 2, 4 conventional.

While the vernier scale described is established by reducing center-to-center distances between adjoining pixels, it is understood that a vernier scale may be created by increasing slightly the center-tocenter distances between adjoining array pixels.

Referring to the embodiment shown in Figure 9, there a pair of relatively-long linear arrays 350, 351 are disposed end-to-end. This may be effected optically as by means of lenses 43 in the scanning apparatus 10 of Figure 1 or mechanically through physical contact of the array ends with one another. To accommodate any gaps between the array ends or misalignments along the X axis and to assure continuity of the composite array so formed, a relatively short bridging array 360 is provided to overlap the adjoining ends of each array 350,351.

Bridging array 360 comprises a relativelyshort linear array, preferably with the minimum quantity of pixels 40 needed to provide effective overlap of the adjoining arrays. Typically, bridging array 360 may be comprised of the order of 100 pixels whereas arrays 350, 351 comprise some 1700 pixels.

In use, data from arrays 350,351,360 may be readout as described earlier the data being stored temporarily in RAM 175 (Figure 3) pending completion of the line. By choosing relatively short bridging arrays 360, the amount of data to be stored in RAM 175, and hence the size of RAM 175, may be small. The data held in RAM 175 are, on completion of the line, readout from RAM 175 into bus 176 (Figure 6), with crossover made from array 350 to bridging array 360 to array 351 in the overlapping areas to ensure continuity.

Referring to the embodiment shown in Figure 10, where like numerals refer to like parts, an array structure 400 is thereshown. Array structure 400 includes relatively long and short arrays 402, 404 respectively mounted upon a common substrate or mask 406. Array 404 is disposed in parallel with array 402, with a portion 409 of array 404 overlapping one end 403 of array 402. The remainder of array 404 projects beyond end 403 of array 402 and is adapted to overlap the leading end of the next successive array structure 400' as seen in drawing Figure 10. To accommodate overlapping of successive array structures 400, substrate 406 is inset at 407.

To enhance alignment between the arrays and provide undistorted crossover between arrays, photosensitive elements or pixels 40' of array 404 are disposed on a center-to-center distance d' different from the center-to center distance d of pixels 40 of array 402. This in effect establishes a vernier scale which enables at least one pixel 40' of array 404 to be aligned with a corresponding pixel 40 of array 402. In the examplary arrangement shown, pixel 40 5 of array 402 is in substantial alignment with pixel 40 - 4 of array 404 and crossover would be effected at this point.

Similarly, when associating the array structure 400 with the next succeeding array structure 400', crossover from array 404 to array 402' is selected at the point of closest pixel alignment. In the embodiment shown, crossover would be between pixel 40' - 7 of array 404 and pixel 40 - 3 of array 402.

While the center-to-center distance d' btween pixels 40' of array 404 is illustrated as being less than the center-to-center distance d between the pixels 40 of array 402, it is understood that dimension d' may be greater than dimension d.

WHAT WE CLAIM IS: 1. Apparatus for scanning an image, includ ing: a movable scanning carriage; at least two longitudinally-extending main arrays of photosensitive elements mounted on the carriage transversely to its direction of movement and adapted to scan non-overlapping areas of the image area, and a short linear bridging array of photosensitive elements on the carriage adapted to scan an area which overlaps both areas scanned by the main arrays.

2. The apparatus as claimed in Claim 1, including means for storing data from said arrays pending readout, and means for reading out data from said storage means in sequence, the sequence crossing from the data from a main array to those from the or a bridging array, and from the bridging array to the next main array.

3. The apparatus as claimed in Claim 1 or 2, in which the cenfre-tocentre distances between the photosensitive elements in the end of one array is different from the centre-to-centre distances between the elements in that part of another array which scans the same part of the image area, to provide a vernier-like differential spacing between the elements in the overlapping scan areas of the two arrays to facilitate alignment of a boundary between two elements in one array with a boundary between two elements in the other array.

4. The apparatus as claimed in Claim 3, in which the centre-to-centre distances between

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (7)

**WARNING** start of CLMS field may overlap end of DESC **. tion of the individual photosensitive elements or pixels 40, 40', to determine the optimum crossover point, may be made through microscopic examination of the arrays. It is further understood that once the optimum crossover point is determined, the descriptor list of memory 206 (Figures 6,7) is programmed to provide crossover from pixels 40 - 8 of array 301 to pixel 40' - 5 of array 300 on readout. While the vernier scale is illustrated as being at one end 308 of array 300 only, it is understood that vernier scales may be provided at both ends of the array. In that event, in a scanning arrangement employing four arrays such as shown in Figure 2, array 1 may have a vernier scale at one end only, with remaining arrays 2, 4 conventional. While the vernier scale described is established by reducing center-to-center distances between adjoining pixels, it is understood that a vernier scale may be created by increasing slightly the center-tocenter distances between adjoining array pixels. Referring to the embodiment shown in Figure 9, there a pair of relatively-long linear arrays 350, 351 are disposed end-to-end. This may be effected optically as by means of lenses 43 in the scanning apparatus 10 of Figure 1 or mechanically through physical contact of the array ends with one another. To accommodate any gaps between the array ends or misalignments along the X axis and to assure continuity of the composite array so formed, a relatively short bridging array 360 is provided to overlap the adjoining ends of each array 350,351. Bridging array 360 comprises a relativelyshort linear array, preferably with the minimum quantity of pixels 40 needed to provide effective overlap of the adjoining arrays. Typically, bridging array 360 may be comprised of the order of 100 pixels whereas arrays 350, 351 comprise some 1700 pixels. In use, data from arrays 350,351,360 may be readout as described earlier the data being stored temporarily in RAM 175 (Figure 3) pending completion of the line. By choosing relatively short bridging arrays 360, the amount of data to be stored in RAM 175, and hence the size of RAM 175, may be small. The data held in RAM 175 are, on completion of the line, readout from RAM 175 into bus 176 (Figure 6), with crossover made from array 350 to bridging array 360 to array 351 in the overlapping areas to ensure continuity. Referring to the embodiment shown in Figure 10, where like numerals refer to like parts, an array structure 400 is thereshown. Array structure 400 includes relatively long and short arrays 402, 404 respectively mounted upon a common substrate or mask 406. Array 404 is disposed in parallel with array 402, with a portion 409 of array 404 overlapping one end 403 of array 402. The remainder of array 404 projects beyond end 403 of array 402 and is adapted to overlap the leading end of the next successive array structure 400' as seen in drawing Figure 10. To accommodate overlapping of successive array structures 400, substrate 406 is inset at 407. To enhance alignment between the arrays and provide undistorted crossover between arrays, photosensitive elements or pixels 40' of array 404 are disposed on a center-to-center distance d' different from the center-to center distance d of pixels 40 of array 402. This in effect establishes a vernier scale which enables at least one pixel 40' of array 404 to be aligned with a corresponding pixel 40 of array 402. In the examplary arrangement shown, pixel 40 5 of array 402 is in substantial alignment with pixel 40 - 4 of array 404 and crossover would be effected at this point. Similarly, when associating the array structure 400 with the next succeeding array structure 400', crossover from array 404 to array 402' is selected at the point of closest pixel alignment. In the embodiment shown, crossover would be between pixel 40' - 7 of array 404 and pixel 40 - 3 of array 402. While the center-to-center distance d' btween pixels 40' of array 404 is illustrated as being less than the center-to-center distance d between the pixels 40 of array 402, it is understood that dimension d' may be greater than dimension d. WHAT WE CLAIM IS:

1. Apparatus for scanning an image, includ ing: a movable scanning carriage; at least two longitudinally-extending main arrays of photosensitive elements mounted on the carriage transversely to its direction of movement and adapted to scan non-overlapping areas of the image area, and a short linear bridging array of photosensitive elements on the carriage adapted to scan an area which overlaps both areas scanned by the main arrays.

2. The apparatus as claimed in Claim 1, including means for storing data from said arrays pending readout, and means for reading out data from said storage means in sequence, the sequence crossing from the data from a main array to those from the or a bridging array, and from the bridging array to the next main array.

3. The apparatus as claimed in Claim 1 or 2, in which the cenfre-tocentre distances between the photosensitive elements in the end of one array is different from the centre-to-centre distances between the elements in that part of another array which scans the same part of the image area, to provide a vernier-like differential spacing between the elements in the overlapping scan areas of the two arrays to facilitate alignment of a boundary between two elements in one array with a boundary between two elements in the other array.

4. The apparatus as claimed in Claim 3, in which the centre-to-centre distances between the remaining elements of the same array.

5. The apparatus as claimed in Claim 3, in which the centre-tocentre distance between the photosensitive elements at one end of a main array is greater than the distance between the remaining elements of the same array.

6. The apparatus as claimed in any preceding claim, in which the bridging array is mounted on the same substrate as one of the main arrays so as to be integral therewith.

7. Apparatus as claimed in any preceding claim, substantially as described herein with ref. erence to, and as shown in, the accompanying drawings.