The invention relates to a solid-state imaging device comprising a matrix, referred to as imaging matrix, of charge-storage sites arranged in rows and columns in which, during an integration period, a field of charge packets generated in dependence upon the local intensity of light captured, via a color filter, at a photosensitive surface can be stored, and read-out means by means of which a field of charge packets can be read row by row, the color filter used being such that, within a common column, charge packets in two successive rows correspond to two different colors.

The imaging device may be of the type commonly referred to in the literature as "CMOS Imager". The imaging device of this type includes a system of intersections in which the cells or pixels are formed by a photodiode, said cells or pixels being coupled via a MOS switch to bit lines, the word lines used to select the rows being connected to the gates of the MOS switches. Most of the imaging devices, however, are of the charge-coupled-device type.

Such a charge-coupled imaging device as well as a method of operating this device are described in patent application WO 98/17051, in the name of the current applicant, which application is laid open to public inspection on 23 April 1998, bearing publication number WO 98/17051. In said document, a description is given of a charge-coupled imaging device comprising a first matrix, referred to as imaging matrix, of charge-storage sites arranged in rows and columns, and, coupled thereto, a second matrix of charge-storage sites arranged in rows and columns, which is a memory matrix in which charge packets of a first field are temporarily stored, while a second field of charge packets is generated in the imaging matrix, the number of columns in the memory matrix corresponding to the number of columns in the imaging matrix, and the number of rows in the memory matrix being smaller than the number of rows in the imaging matrix, so that upon transport of a field of charge packets from the imaging matrix to the memory matrix, only a limited number of rows of the entire field is stored in the memory matrix, and the remaining lines are dumped via a drain region. Although the following description will specifically relate to a charge-coupled imaging device, it should be kept in mind that the invention is not limited thereto and can also suitably be used in the above-described CMOS imager.

The invention is of importance, in particular, to imaging devices having a large number, for example several millions, of picture elements (hereinafter referred to as pixels), which are used, for example, for digital photography. For direct picture display, for example on a LCD screen in the viewfinder of the camera, it is often desirable to use a memory matrix in addition to the imaging matrix. On account of the large number of pixels, it is not very well possible to display all pixels in the viewfinder during operation. For this reason, it is generally not necessary to provide the device with a memory matrix having an equally large number of storage sites as the imaging matrix. In addition, it is desirable to employ a smaller memory matrix in order to preclude the dimensions of the device from becoming too large. As described in the above-mentioned application, a much smaller memory is sufficient, so that the overall dimensions of the imaging device can be kept within reasonable limits. Since the memory can hold only a part of the pattern of charge packets generated in the imaging matrix, only a part of the rows is stored in the memory matrix while the rest of the rows are not read out (for example in the case of the CMOS imager or CCD imager of the interline type) or are dumped. In the literature, this method is referred to as"sub-sampling". As described hereinabove, it is possible, for example, to drain off the redundant charge packets via the substrate. The manner in which the unused lines are drained is not relevant to the invention in question; other ways of draining can also be used in combination with the invention.

The above-mentioned application also describes the possibility of sub-sampling if the surface of the imaging device is provided with a color filter, whereby two successive rows in the imaging section contain different color information. By way of example, a description is given of an imaging device including a filter of the following composition: row 1: RGRGRG row2: GBGBGB row3: RGRGRG row4: GBGBGB etc. In this case, the pattern of the color filter recurs every other row. In order to preserve the color information during sub-sampling (and hence be able to display the imaged picture in color in the viewfinder of the camera) it has already been proposed in the above-cited application to store each time two successive rows of the imaged picture in the memory matrix, then dump a number of lines and subsequently store two successive lines again. In practice, due to picture distortion, the reproduction quality is often found to be indifferent or at least not as good as one would wish it to be. It is an object of the invention to provide, inter alia, a charge-coupled imaging device of the type described in the opening paragraph, in which

the above-mentioned distortion upon displaying a sub-sampled picture is at least substantially overcome.

To achieve this, a solid-state imaging device of the type described in the opening paragraph is characterized in accordance with the invention in that means are available by means of which, for the purpose of reading out, a part of the rows are selected, while the other rows remain unread, and, if a first row is read, at least two succeeding rows are not read and the next row to be read contains color information which differs from that of the above-mentioned first row that has been read, and, in two successive fields, rows of the second field which are to be read, correspond to positions in the imaging matrix which are situated between positions of successive rows of the first field, which have been read.

The invention is, inter alia, based on the recognition that the above-described distortion upon displaying the picture is connected with the interlacing mode for scanning the display screen, in which in a first field the odd lines are written, and in a second field, the even lines are written, or conversely. Due to interlacing, two successive rows of the imaging matrix are not displayed as two successive rows or lines but as two lines which are separated by at least a third line which corresponds to a row in the imaging matrix which is not situated between said two rows and hence is not displayed at the proper location on the display screen.

By, in accordance with the invention, not reading out successive rows of the imaging matrix (or not storing successive rows in the memory matrix in the case of a charge-coupled imaging device), but instead reading out (or storing in the case of a charge-coupled imaging device) rows which are farther apart, it is possible, as will be explained hereinafter by means of an example, to display a sub-sampled picture practically without distortion, while retaining the color information. An important embodiment of a device in accordance with the invention is characterized in that the device constitutes a charge-coupled device comprising a first matrix, referred to as imaging matrix, of charge-storage sites arranged in rows and columns, and, coupled thereto, a second matrix of charge-storage sites arranged in rows and columns, which second matrix constitutes a memory matrix in which charge packets of a first field are temporarily stored, while a second field of charge packets is generated in the imaging matrix, the number of columns in the memory matrix corresponding to the number of columns in the imaging matrix, and the number of rows in the memory matrix being smaller than the number of rows in the imaging matrix, so that upon transport of a field of charge packets from the imaging matrix to the memory matrix only a limited number of rows of the entire field is stored in the memory matrix, and the remaining lines are dumped via a drain region, means being present by means of which charge packets are transported from the first field and the

second field of the imaging matrix to the memory matrix, in such a manner that when a first row is stored in the memory matrix at least two succeeding rows are dumped and the next row that is stored includes color information which differs from that of the above-mentioned first row stored in the memory matrix, rows of the second field which are stored in the memory matrix corresponding to positions in the imaging matrix which are situated between positions of successive rows of the first field which are stored in the memory matrix.

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

In the drawings: Fig. 1 is a diagrammatic plan view of a charge-coupled device in accordance with the invention; Fig. 2 shows a part of the color filter used in the device shown in Fig. 1.

Fig. 1 diagrammatically shows a charge-coupled imaging device as described in the above-mentioned patent application PCT/IB 97/01201, the contents of which is considered to be incorporated in the current application by way of reference. The device comprises an imaging matrix 1 of charge-storage sites arranged in rows 2 and columns 3. The drawing shows a few rows (in continuous lines) and a few columns (in broken lines). The device may be of the field-transfer type in which the imaging elements are formed by charge-storage sites of the charge-transport ducts themselves, as described in the above-mentioned application. The device may alternatively be of the interline type, in which the matrix 1 includes, as is known, columns of separate photosensitive diodes alternated with vertical charge-transport ducts screened against radiation. The surface of the imaging matrix is provided with a color filter 4, not shown in Fig. 1, a part of which is shown in Fig. 2. The filter is embodied so as to be a matrix which coincides with the imaging matrix 1. The letters R, G and B represent the colors red, green and blue, respectively. As shown in Fig. 2, the color pattern is repeated every alternat line, so that, in the imaging matrix 1, different color information is formed during an integration period in successive lines or rows.

The device includes a second matrix 5, referred to as memory matrix, which is adjacent to the imaging matrix 1 and in which charge packets of a first field, generated in the imaging matrix, are stored for further processing, such as display on a display screen in the viewfinder of a camera while a second field is being formed in the imaging matrix. The matrix 5 ends in a horizontal read-out register 9. The block 10 represents voltage means by means of

which the proper voltages are applied to the gates and, if necessary, regions in the semiconductor body. As indicated in the drawing, the memory is smaller than the imaging matrix, which can be attributed to the fact that the number of rows in the memory matrix is smaller than in the imaging matrix. The number of columns in both matrices is the same. Since it is not possible to store all lines of a field formed in the imaging matrix in the memory matrix, during the charge transport only a part of the information stored in the imaging matrix is stored in the memory, while the rest is dumped via a drain. The way in which redundant lines are dumped is described in greater detail in the above-mentioned patent application PCT/IB 97/01201, in the name of the current applicant. In the method described therein, the drain is formed by the substrate situated below the charge-transport ducts, to which substrate a sufficiently high voltage is applied. Excess charge is dumped at the transition between the imaging matrix and the memory matrix, for example by activating the gates of the imaging matrix and, simultaneously, not activating in the gates of the memory matrix. It is noted that the invention is not limited to this way of draining charge. Other methods, such as the one described in European patent application EP-A 0 720 388, can also be applied within the scope of the invention.

By means of Table I, a description is given of the operation of the device.

colour<BR> sensor fine<BR> sensor line<BR> display line<BR> sensor line<BR> display line<BR> sensor line<BR> display line<BR> sensor line<BR> display line<BR> sensor line<BR> display line<BR> sensor line<BR> display line 1R1 1111 B 22#2#2 R 3 # 1 3 (i) 1 3 # 1 3 #344444B4 555#25#2R5 66#36#36B6 7#27777R7 8#48#48#4B8 R 9 9 9 9 9 9 9 B 101010101010 R 11 11#511#5115 B 12 (g) 7 12 12 12 12 12 R 1313#613#613 B 14 14 14 14 14 14 14 R 15 15 # 6 15 15 15 15 B 16 16 # 8 # 7 16 16 7 16 17R17 17171717 B 1818#818#818 R 191919#91919 B 202020202020 R 21212121#1021 B 22 22 22 22 22 (g) 11 22 R 23 23 23 23 23 23 B 24242424#1224 laid 1 field 2 field 1 field 2 field 1 field 2 4/16x9606/22x9609604/16x960 = 240 = 240 = 262 (a) (b) (c) (d) table 1

In the following description, it has been assumed that the imaging matrix includes 960 lines or rows. In column (a) only the first 24 lines of said 960 lines are indicated. The symbols R and G in the first column (a) indicate whether in the relevant row the color red or the color green occurs in the filter. The next column in (a) gives the number of the relevant row or line in the imaging matrix. In part (b) of Table I, the most obvious way of sub-sampling is indicated for two successive fields. Here, it is assumed that the capacity of the memory matrix is large enough to store 240 lines, which means one in four lines of the imaging matrix. In the Table, the encircled (underlined) digits indicate the number of the lines that will be stored. In this method, to preserve the color information, each time two succeeding lines are stored in the memory matrix, whereafter a number of subsequent lines are removed. For example, in the example shown in (b) in Table I, the lines 3 and 4 of field 1 are stored and subsequently the lines 11 and 12. The intermediate lines 5-10 are drained. The lines of field 2 that are stored and read out are situated between the selected lines of field 1; in the example, the lines 7 and 8 and the pair 15,16 are stored. It can be readily ascertained that, in the current system, out of every 16 lines, 4 lines are stored in the memory. Consequently, if an imaging matrix has 960 lines, then a memory of 240 lines is sufficient. If the picture is displayed by means of a display system in which the display screen is written in accordance with an interlaced mode, then, as indicated in the Table, line 7 of field 2 is written as line 2 on the display screen, while, for example, line 4 in field 1 appears as line 3 on the display screen, i. e. after line 7 of field 2.

Table I (c) indicates the way in which in a device in accordance with the invention this distortion can be, at least substantially, precluded. Instead of storing two succeeding lines in the memory each time, now lines with different color information are successively stored in the memory, which lines are farther apart. In the example shown in Table I (c), for example the lines 3 and 6 are stored, in field 1, while the intermediate lines 4 and 5 are dumped. The following lines to be stored of field 1 are the lines 11 (R) and 16 (B); the intermediate lines are all dumped. The lines 3,6,11 and 16 of the first field, which are to be stored, correspond to the lines 1,3,5,7 etc., of the display screen. Of the second field, the lines 5,8,13,18 etc., are stored in the memory while the intermediate lines are dumped. The lines to be stored of the second field correspond to the lines, respectively, 2,4,6, etc., of the display screen. As shown in Table I, these lines are situated, in the correct order in conformity with the lines on the display screen, between the lines of the first field, so that the above- described distortion is precluded.

If the way of sub-sampling indicated in Table I (c) is employed, then, in the case of an imaging matrix of 960 lines, a memory of 240 lines, in conformity with the NTSC

standard, is sufficient. Table I (d) gives a way of sub-sampling 288 lines, in conformity with the PAL standard, of the 960 lines in the imaging matrix. In field 1, alternately the R and B lines etc., are stored, which correspond respectively to the lines 1,3,5,7 etc., of the display screen. In the second field, alternately R and B lines are selected, which are situated between the lines of the first field, for example the lines etc., which correspond to the even lines 2,4,6,8, etc., on the display screen. As shown in the Table, the selected lines of the second field are indicated on the display screen in the proper sequence relative to the indicated lines of the first field, while retaining the color information.

As is further shown in the Table, the number of lines stored is 6/20*960 = 288, in conformity with the PAL standard.

Fig. 3 schematically shows a solid-state imaging device of the CMOS type without a memory matrix, which imaging device is known per se. Said imaging device includes rows indicated by Ml, M2, M3 etc., and columns indicated by C1, C2, C3, etc. The imaging device further includes a mosaic color filter of, for example, the same composition as in the preceding example, the lines M1, M3, M5, etc., corresponding, for example, to the lines "R", that is the lines 1,3,5 in Table I, and the lines M2, M4, M6 etc., corresponding to the lines"B"in Table I. The elements in an imaging element include a diode D which is coupled to one of the main electrode areas of a MOS switch T. The anode of the diode D is connected to a reference voltage. The cathode of the diode D can be connected, via the switch T, to one of the (vertical) bit lines 6 upon reading and resetting the imaging element. The rows are addressed via word lines 7 which are connected to the gates of the transistors T of a row. The word lines 7 are connected to addressing means 8 by means of which a voltage is applied to a selected line 7, thereby causing the switches or transistors T of the relevant line 7 to be closed, and enabling the charge condition of the diodes D to be determined via the bit lines 6. The manner in which, in general, an imaging device of this type is operated is assumed to be known. In this imaging device, the above-described principle of sub-sampling can be employed, in the manner indicated in Table I, with this difference that the encircled numbers are not stored in a memory, but instead read out, and that the intermediate lines, instead of being dumped, are not selected during reading.

It will be obvious that the invention is not limited to the examples given herein, and that within the scope of the invention many variations are possible to those skilled in the art. For example, a color filter other than the filter described can be employed, for example a filter of the composition:

CyYeCyYe Mg Gr Mg Gr Cy Ye Cy Ye Mg Gr Mg Gr The invention can also be used in combination with a color filter in which repetition occurs every three, or more, lines. It is also possible to use a charge-coupled imaging device of the"interline"type.