Solid State Imaging Device

1. Field of the Invention

The present invention relates to a solid state imaging device, and more particularly to a solid state imaging device suitable to produce a color image signal.

2. Description of the Prior Art

In the prior art, when a color image signal with three or more color signals is to be produced by one or two solid state imaging devices, image sensing parts of the imaging devices receive light through color separating optical members such as color filters arranged in stripe or mosaic to form electrical information representative of the respective colors in picture cells and the electrical information of the picture cells are time-serially read out through a common transfer path.

FIG. 1 shows an example of a well-known prior art solid state imaging device which utilizes a frame transfer (FT) type charge coupled device (CCD).

In FIG. 1, numeral 1 denotes an image sensing part having a plurality of photo-electric converting picture cells arranged in rows and columns. Numeral 2 denotes a memory part for storing charge information of the picture cells of thee image sensing part 1. Numeral 3 denotes a horizontal register which functions as a read transfer path. It reads out the information of the memory part 2 one horizontal line at a time and transfers the line information horizontally to time-serially produce point-sequential signals.

By arranging color separating stripe color filters as shown in FIG. 2 in front of the imaging part 1 with the pitch of the color filters R (red), G (green) and B (blue) being coincident with the pitch of the picture cells of the image sensing part 1, the picture cells in the respective columns produce signals representative of the respective colors and the point-sequential color signals are time-serially produced from the horizontal shift register 3.

The color signals thus produced are converted to, for example, an NTSC signal by a signal processing circuit as shown in FIG. 3.

The point-sequential image output signals from a CCD amplifier 4 are sampled and held by a signal separation circuit 8 comprising three sample-and-hold circuits 5, 6 and 7 so that a red signal E.sub.R, a green signal E.sub.G and a blue signal E.sub.B are separated. The color signals E.sub.R, E.sub.G and E.sub.B are level-adjusted by variable gain amplifiers 9, 10 and 11, respectively, so that a white balance is controlled. The level-adjusted color signals are than processed by processing circuits 12, 13 and 14 each including a clamp circuit, a .gamma. correction circuit and an aperture correction circuit, and the signals are converted to a luminance signal and two color difference signals by a matrix circuit 15, and they are converted to the NTSC signal by an encoder 16.

With such an arrangement, the horizontal register 3 sequentially reads out three primary colors. In order to read them with a carrier of 3.58 MHz, a clock of 3.58 MHz.times.3=10.74 MHz is required. However, as the clock frequency is high, the transfer efficiency is reduced and the power consumption increases. As a result, a problem is encountered when the number of picture cells of the horizontal shift register or the number of horizontal picture cells of the image sensing part 1 is large.

It is an object of the present invention to provide a solid state imaging device suitable for color imaging, which overcomes the difficulties encountered in the prior art solid state imaging device.

It is another object of the present invention to provide a color imaging device which separates color signals to produce the color signals directly.

It is an other object of the present invention to provide a solid state imaging device which has reduced noise, is capable of driving a horizontal register at a low rate and has a high transfer efficiency.

It is a further object of the present invention to provide a solid state imaging device which allows a significant increase of the number of horizontal picture cells and significantly improves horizontal resolution.

In accordance with a feature of the present invention, there is provided a solid state imaging device comprising imaging means having a plurality of picture cells arranged in rows and columns for producing electrical information in response to incident radiation, readout means having m (m.gtoreq.3) read sections for reading the electrical information produced by the imaging means and input means for dividing one row of information in the imaging means into m groups and reading the information by parallel-to-serial conversion for each group.

In accordance with another feature of the present invention, controllable isolate means for isolating the read sections from each other in shifting the information in the read sections is additionally provided. In this manner, the information can be synchronously read from the m read sections.

The solid state imaging device of the present invention is suitable for color imaging. In accordance with still another feature of the present invention which takes a combination of color separating optical members and the solid state imaging device into consideration, there is provided a solid state imaging device comprising imaging means for receiving incident light through a color separating optical means which separates the incident light into a plurality of color light components to produce electrical information representative of the respective color light components, readout means having a plurality of read sections for reading the electrical information produced by the imaging means and input means for classifying one row of electrical information in the imaging means to the respective color information and supplying it to the readout means while converting it by parallel-to-serial conversion.

Again, isolate means for isolating the read sections from each other in shifting the information in the read sections may be additionally provided so that the color signals can be synchronously read out.

In accordance with the present invention, the solid-state imaging device is characterized by the input means. Considering particularly the solid-state imaging device comprising the imaging means having the plurality of picture cells arranged in rows and columns to produce the electrical information representative of the incident light, the readout means having the m read sections for reading the electrical information produced by the imaging means and input means for dividing the arrangement in the imaging means into m groups with respect to the columns and supplying the electrical information to the readout means while converting it by parallel-to-serial conversion for each group, the input means is arranged between the imaging means and the readout means and transfers the information of the respective column groups while imparting different delays to the m picture cell column groups of the imaging means so that the information of the respective column groups are supplied to the corresponding read sections.

In a specific embodiment, when m is three, the input means transfers the information while it imparts zero picture cell or zero bit of delay to the information of the first column group, one picture cell or one bit of delay to the information of the second column group and two picture cells or two bits of delay to the information of the third column group so that the information of the respective column groups are supplied to the corresponding read sections.

In other words, the input means is constructed to have a function of parallel-to-serial conversion.

The read sections, when imparted with the different delays, read the information of the corresponding column groups sent at the different timings.

The other objects and features of the present invention will be apparent from the description of the preferred embodiments of the invention taken in conjunction with the accompanying drawings.

FIG. 1 shows a configuration of a prior art frame transfer type CCD,

FIG. 2 shows an example of a stripe color filter,

FIG. 3 shows a configuration of a prior art color image signal processing circuit,

FIG. 4 shows one embodiment of a solid state imaging device of the present invention,

FIG. 5 shows detail of the device of FIG. 4,

FIGS. 6(a), 6(b) and 6(c) show a vertical transfer timing and a horizontal transfer timing of the device,

FIG. 7 shows a configuration of a color imaging system which uses the above device,

FIG. 8 shows another embodiment of the solid state imaging device of the present invention,

FIG. 9 shows detail of the device of FIG. 8,

FIGS. 10(a) and 10(b) show a vertical transfer timing and a horizontal transfer timing of the device of FIG. 8, and

FIG. 11 shows a further embodiment of the present invention, shown in a similar manner to FIGS. 5 and 9.

One embodiment of the present invention is shown in FIG. 4 in which the like numerals to those shown in FIGS. 1 to 3 denote the like elements or like functional means.

In FIG. 4, a color stripe filter for color separation as shown in FIG. 2 is arranged in front of the image sensing part 1. Numerals 31, 32 and 33 denote horizontal shift registers which function as horizontal read sections, numerals 41, 42 and 43 denote charge voltage conversion amplifiers, and numeral 17 denotes separating input section arranged between the memory part 2 and the three horizontal shift registers 31-33 for distributing three color information contained in the last horizontal line of the memory part 2 to the corresponding one of the three horizontal shift registers 31-33 to parallel-to-serial convert the charges.

In the present embodiment, three horizontal shift registers 31-33 are provided, one for each of the color signals to be produced, as the read transfer path and the charges representative of the respective colors are distributed to the respective horizontal shift registers 31, 32 and 33 to read out the information.

Accordingly, the respective color signals are essentially sampled by the horizontal registers 31, 32 and 33 and the amplifiers 41, 42 and 43 produce the separated color signals.

A charge clear drain CD is arranged at the lowermost end of the imaging device, that is, under a charge clear gate CL adjacent to the horizontal shift register 31. The drain CD is connected to a power supply level.

FIG. 5 shows an electrode structure of the imaging device of FIG. 4. It shows a lower end of the memory part 2 to the three horizontal shift registers 31-33.

Hatched areas show channel stops, numerals 31E-33E denote transfer electrodes of the horizontal shift registers 31-33, numeral 17E denotes a transfer electrode of the separating input section 17 and numeral 2E denotes a transfer electrode of the memory parts 2.

While the present embodiment shows a single-phase driven transfer, two-phase, three-phase or even four-phase driven transfer may be used.

A set of portions A, B, C and D form a unit cell and potentials of the portions A-D are represented by P(A)-P(D). Virtual electrodes are formed by ion injection such that P(A)>P(B) and potential levels are fixed. Potentials of the portions C and D under the transfer electrodes are set such that P(C)>P(D). When low level potentials are applied to the respective electrodes, a relation of P(A)>P(B)>P(C)>P(D) is met, and when high level potentials are applied, a relation of P(C)>P(D)>P(A)>P(B) is met. It is of opposite relation to that of the potentials P(A)-P(D).

.phi.1-.phi.3 denote clock pulses applied to the electrodes 31E-33E, .phi.T denotes a clock pulse applied to the electrode 17E, .phi.S denotes a clock pulse applied to the electrode 2E and .phi.CL denotes a clock pulse applied to the electrode CE of the clear gate CL.

FIGS. 6(a) and 6(b) show a vertical transfer clock timing and a horizontal transfer timing, respectively.

FIG. 7 shows a block diagram of a color imaging system which uses the imaging device of the present invention. Numeral 18 denotes a driver which functions as control means and supplies the clock pulses .phi.I, .phi.S, .phi.T, .phi.1, .phi.2, .phi.3 and .phi.CL as shown in FIGS. 6(a) and 6(b). .phi.I denotes the clock pulse applied to the electrode of the image sensing part 1. Numeral 19 denotes a reference signal generator. The like numerals to those shown in FIG. 3 denote the like elements. As seen from FIG. 7, the present embodiment omits the sample and hold circuit for separating the color information so that the circuit configuration is simplified.

The operation of the configuration shown in FIG. 5 is explained. In the vertical transfer of the charges from the image sensing part 1 to the memory part 2 as shown in FIG. 6(a), synchronized clock pulses of substantially the same phase (except the clock pulse .phi.S which slightly leads to the other clock pulses as shown), which are equal in number to at least the number of vertical picture cells of the image sensing area 1 are supplied in synchronism with a vertical synchronizing signal V.SYNC during a time period t.sub.1 -t.sub.2 as the clock pulses .phi.I, .phi.S, .phi.T, .phi.3, .phi.2, .phi.1 and .phi.CL so that the charges remaining in the memory part 2 are drained to the clear drain CD through the horizontal registers 33, 32 and 31 and the clear gate CL and the charges in the image sensing part 1 are transferred to the memory part 2 and stored therein. At and after time t.sub.3, the clock pulses .phi.T, .phi.3, .phi.2 and .phi.1 are supplied as shown so that the horizontal information in the last line of the memory part 2 are distributed to the three horizontal shift registers 31-33 through the separating input section 17, and at and after time t.sub.4, the information in the horizontal registers 31, 32 and 33 are sequentially read out by supplying the clock pulses .phi.1, .phi.2 and .phi.3 to the horizontal registers 31-33 as shown.

Regarding the information in the next and following lines of the memory part 2, the clock pulse .phi.S is applied immediately before the time t.sub.3 to shift one line of information to be read to the last line of the memory part 2 and then the above operation is repeated (FIG. 6(b)).

The operation during the time period t.sub.3 -t.sub.4, that is, the operation to distribute the information in the last line of the memory part 2 to the three horizontal shift registers 31-33 through the separating input section 17 is now explained in detail with reference to FIGS. 5, 6(a) and 6(b). For the sake of simplification, the shift of the charge information in only three columns I, II and III of the memory part 2 shown in FIG. 5 is explained although the same operations are carried out for the columns of other three-column sets.

At time t.sub.3, the clock pulse .phi.T assumes a high level in substantial synchronism with the horizontal synchronizing signal H.SYNC and the charges stored in the portions 116, 117 and 118 in the last line of the memory part 2 are shifted to portions 111, 114 and 115 in the separating input section 17. When the clock pulse .phi.T thereafter assumes a low level, the charges in the portions 111, 114 and 115 are shifted to portions 110, 113 and 106, respectively. When the clock pulses .phi..sub.3, .phi..sub.2 and .phi..sub.1 are sequentially applied slightly later than the clock pulse .phi.T, the charge in the portion 106 of the separating input section 17, that is, the charge initially stored in the portion 118 in the column I of the memory part 2 is shifted to a portion 100 of the horizontal register 31 through portions 105 and 104 of the horizontal register 33, portions 103 and 102 of the horizontal register 32 and a portion 101 of the horizontal register 31 and stored in the portion 100 of the horizontal register 31.

When the clock pulse .phi.T is again applied, the charges in the portions 110 and 113 of the separating input section 17 are shifted to the portions 108 and 106 through the portions 109 and 112. When the clock pulses .phi..sub.3 and .phi..sub.2 are sequentially applied slightly later than the clock pulse .phi.T, the charge in the portion 106 of the separating input section 17, that is, the charge initially stored in the portion 117 in the column II of the memory part 2 is shifted to the portion 102 of the horizontal register 32 through the portions 105, 104 and 103 and stored in the portion 102.

When the clock pulse .phi.T is again applied, the charge in the portion 108 of the separating input section 17 is shifted to the portion 106 through the portion 107. When the clock pulse .phi..sub.3 is applied slightly later than the clock pulse .phi.T, the charge in the portion 106 of the separating input section 107, that is, the charge initially stored in the portion 116 in the column III of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 and stored therein.

In this manner, the charges stored in the last line of the memory part 2 are distributed to the groups of the columns I, II and III through the separating input part 17. Thus, if the R, G and B stripe filters are arranged such that the R stripe filter corresponds to the column I group, the G stripe filter corresponds to the column II group and the B stripe filter corresponds to the column III group, the charges corresponding to R, G and B are stored in the horizontal registers 31, 32 and 33, respectively.

At and after the time t.sub.4, the charges stored in the horizontal registers 31, 32 and 33 are read out (OUT1-OUT3 in FIG. 6(b)).

When the charges in one horizontal line of the memory part 2 have been read out through the horizontal registers 21-33, the clock pulse .phi.S is applied to the memory part 2 as shown in FIG. 6(b) and the charges stored in the respective horizontal lines are vertically shifted by one horizontal line so that new charges are stored in the last line, and then the operation of the time period t.sub.3 -t.sub.4 is carried out to distribute the new last line of charges to the horizontal registers 31-33.

By repeating the above operations, the charges stored in the respective lines of the memory part 2 are separated to the respective colors and then read out.

In order to prevent the charges of the registers 31, 32 and 33 from being mixed when the charges of the horizontal registers 31-33 are horizontally shifted, control gates for isolating the registers 31, 32 and 33 from each other in a horizontal charge transfer mode may be additionally provided as shown in another embodiment to be described later, or the waveforms of the clock pulses .phi..sub.1, .phi..sub.2 and .phi..sub.3 in the horizontal transfer mode may be slightly modified so that the charges are horizontally transferred without being mixed among the horizontal registers 31-33 without adding the isolating means. For example, at the time t.sub.4, the clock pulse .phi..sub.3 is set to the high level in advance to the clock pulses .phi..sub.2 and .phi..sub.1 to shift the charge stored in the portion 104 of the horizontal register 33 to the portion corresponding to C under the left adjacent electrode 33E, and the clock pulse .phi..sub.2 is then set to the high level while keeping the clock pulse .phi..sub.3 at the high level to shift the charge stored in the portion 102 of the horizontal register 32 to the portion corresponding to C under the left adjacent electrode 32E, and the clock pulse .phi..sub.1 is then set to the high level while keeping the clock pulses .phi..sub.3 and .phi..sub.2 at the high level to shift the charge stored in the portion 100 of the horizontal register 31 to the portion corresponding to C under the left adjacent electrode 31E, and the clock pulse .phi..sub.1 is then set to the low level in advance to the clock pulses .phi..sub.2 and .phi..sub.3 to shift the charge stored in the portion corresponding to C under the electrode 31E of the horizontal register 31 to the portion corresponding to A on the left side thereof, and the clock pulse .phi..sub.2 is then set to the low level to shift the charge stored in the portion corresponding to C under the electrode 32E of the horizontal register 32 to the left adjacent portion, and the clock pulse .phi..sub.3 is then set to the low level to shift the charge stored in the portion corresponding to C under the electrode 33E of the horizontal register 33 to the portion corresponding to A on the left side thereof. In this manner, the transfer of the charges from the portions A to the portions C is carried out sequentially from the horizontal register 33 through the horizontal register 32 to the horizontal register 31, and the transfer of the charges from the portions C to the portions A is carried out in the opposite sequence from the horizontal register 31 through the horizontal register 32 to the horizontal register 33. Thus, the charges are horizontally transferred without being mixed among the horizontal registers 31-33 without the additional isolating means. The clock pulses .phi..sub.1, .phi..sub.2 and .phi..sub.3 and the three outputs derived are shown in FIG. 6(c).

Another embodiment of the present invention is now explained. As described above, in the present embodiment, controllable isolate sections for isolating the horizontal registers 31, 32 and 33 from each other in the horizontal charge transfer mode are additionally provided.

Referring to FIG. 8, the like numerals to those shown in the previous drawings designate the like elements. Numerals 51, 52 and 53 denote controllable isolate sections associated with the horizontal registers 31, 32 and 33 to isolate the horizontal registers 31, 32 and 33 from each other in the horizontal charge transfer mode of the horizontal shift registers 31, 32 and 33. More particularly, they are constructed as shown in FIG. 9, in which mumerals 51E, 52E and 53E denote control electrodes of the isolate sections 51, 52 and 53, respectively, to which the clock pulse .phi.T is applied. In the present embodiment, transfer electrodes 31E', 32E' and 33E' of the horizontal shift registers 31, 32 and 33 are separated in the respective horizontal registers 31, 32 and 33 as shown although they are connected in common for each of the horizontal registers 31, 32 and 33 by well-known means such as Al substrate.

While the present embodiment shows a single-phase drive, two-phase, three-phase or even four-phase drive may be used.

The other portions are identical to the previous embodiment. A color imaging system which utilizes the imaging device of the present invention may be constructed in the same manner as shown in FIG. 7.

The operation of the present embodiment is now explained. As shown in FIG. 10(a), in the vertical transfer mode of the charges from the image sensing part 1 to the memory part 2, synchronous clock pulses of essentially the same phase (except the clock pulse .phi.S which slightly leads the other clock pulses) which are equal in number to at least the number of vertical picture cells in the image sensing part 1 are supplied as the clock pulses .phi.I, .phi.S and .phi.T during the time period t.sub.1 -t.sub.2 in substantial synchronism with the vertical synchronizing signal V.SYNC so that the charges remaining in the memory part 2 are drained to the clear drain CD through the isolate sections 51-53, the horizontal registers 31-33 and the clear gate CL and the charges in the image sensing part 1 are transferred to the memory part 2 and stored therein. Then, as shown in FIG. 10(b), at and after the time t.sub.3, the clock pulse .phi.T is supplied as shown to distribute the horizontal information in the last line of the memory part 2 to the separating input section 17, the isolate sections 51-53 and the horizontal shift registers 31-33, and at and after the time t.sub.4, the clock pulses .phi..sub.1, 100.sub.2 and .phi..sub.3 are applied to the horizontal registers 31, 32 and 33 as shown to sequentially read out the information.

Regarding the information in the next and following lines of the memory part 2, the clock pulse .phi.S is applied immediately before the time t.sub.3 to shift one line of information to be read to the last line of the memory part 2 and then the above operation is repeated (FIG. 10(b)).

The operation during the time period t.sub.3 -t.sub.4, that is, the operation to distribute the information in the last line of the memory part 2 to the three horizontal shift registers 31-33 through the separating input section 17 and the isolate sections 51-53 is now explained in detail with reference to FIGS. 9 and 10(b). For the sake of simplification, the shift of the charge information in only three columns I, II and III of the memory part 2 shown in FIG. 9 is explained although the same operations are carried out for the columns of other three-column sets.

At time t.sub.3, the clock pulse .phi.T assumes a high level in substantial synchronism with the horizontal synchronizing signal H.SYNC and the charges stored in the portions 116, 117 and 118 in the last line of the memory part 2 are shifted to portions 111, 114 and 115 in the separating input section 17. When the clock pulse .phi.T thereafter assumes a low level, the charges in the portions 111, 114 and 115 are shifted to portions 110, 113 and 106, respectively. When the second clock pulse .phi.T is applied, the charge in the portion 106 of the separating input section 17, that is, the charge initially stored in the portion 118 in the column I of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 of the isolate section 53, and the charges stored in the portions 110 and 113 of the separating input section 17 are shifted to the portions 108 and 106 through the portions 109 and 112, respectively. When the third clock pulse .phi.T is applied, the charge in the portion 104 of the horizontal register 33 is shifted to the portion 102 of the horizontal register 32 through the portion 103 of the isolate section 52, and the charge in the portion 106 of the separating input section 17, that is, the charge initially stored in the portion 117 in the column II of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 of the isolate section 53. The charge in the portion 108 of the separating input section 17 is shifted to the portion 106 through the portion 107. When the fourth clock pulse .phi.T is applied, the charge in the portion 102 of the horizontal register 32 is shifted to the portion 100 of the horizontal register 31 through the portion 101 of the isolate section 51 and stored therein, and the charge in the portion 104 of the horizontal register 33 is shifted to the portion 102 of the horizontal register 32 through the portion 103 of the isolate section 52 and stored therein. The charge in the portion 106 of the separating input section 17, that is, the charge initially stored in the portion 116 in the column III of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 of the isolate section 53 and stored therein.

In this manner, the charges stored in the last line of the memory part 2 are distributed to the shift registers 31-33 in the groups of the columns I, II and III through the separating input part 17. Thus, if the R, G and B stripe filters are arranged such that the R stripe filter corresponds to the column I group, the G stripe filter corresponds to the column II group and the B stripe filter corresponds to the column III group, the charges corresponding to R, G and B are stored in the horizontal registers 31, 32 and 33, respectively.

At and after the time t.sub.4, the charges in the horizontal registers 31, 32 and 33 are read out by applying the clock pulses .phi..sub.1 -.phi..sub.2. Since the clock pulse .phi.T is kept at the low level at this time, the isolate sections 51-53 function as barriers so that the mixing of the charges among the horizontal registers 31, 32 and 33 in the horizontal transfer of the charges in the horizontal registers 31, 32 and 33 is prevented.

When the charges in one horizontal line of the memory part 2 have been read out through the horizontal registers 31-33, the clock pulse .phi.S is applied to the memory part 2 as shown in FIG. 10(b) and the charges stored in the respective horizontal lines are vertically shifted by one horizontal line so that new charges are stored in the last line, and then the operation of the time period t.sub.3 -t.sub.4 is carried out to distribute the new last line of charges to the horizontal registers 31-33.

By repeating the above operations, the charges stored in the respective lines of the memory part 2 are separated to the respective colors and read out (OUT1-OUT3 in FIG. 10(b)).

In the present embodiment, in the horizontal transfer of the input charges in the horizontal registers 31-33, the phases of the clock pulses .phi..sub.1, .phi..sub.2 and .phi..sub.3 to the registers 31, 32 and 33, respectively, are varied as shown in FIG. 10(b) so that the three color signals are produced at different phases, although the clock pulses .phi..sub.1 -.phi..sub.3 may be in phase as shown by chain lines in FIG. 10(b) so that the three color signals are simultaneously produced. The manner of reading out the three color signals may be selected depending on the subsequent signal processing.

A further embodiment of the present invention is explained with reference to FIG. 11, which shows an improvement over the embodiment shown in FIGS. 8-10. In the present embodiment, the separating input section is formed by the last line of the memory part 2 and a portion of the separating input section is provided with the function of the isolate section 53 shown in FIGS. 8 and 9 so that the configuration is simplified.

Referring to FIG. 11, the like numerals to those shown in FIG. 9 designate the like elements and the like numerals with primes designate the like functional means.

Numeral 17' denotes a separating input section, and numeral 17E' denotes an electrode of the separating input section 17'. The clock pulse .phi.T is applied to the electrode 17E'. As shown, the separating input section 17' is formed by a region B of the last line of the memory part 2 and a portion of the separating input section 17' has a function of the isolate section to the memory part 2, for the horizontal register 33.

The other portions are identical to FIG. 9. In the arrangement described above, when the charges have been vertically transferred from the image sensing part 1 to the memory part 2 (corresponding to the time t.sub.2 in FIG. 10(a)), the charges in the last line of the memory part 2 are distributed in the following manner. The charge in the column I is stored in a portion 118', the charge in the column II is stored in a portion 117' and the charge in the column III is stored in a portion 116'.

A read mode starts from the above status at the time t.sub.3 of FIG. 10(b). When the clock pulse .phi.T is applied, the charge in the portion 118' of the column I of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 of the separating input section 17', the charge in the portion 117' in the column II of the memory part 2 is shifted to the portion 118' of the memory part 2 through the portion 112' of the separating input section 17' and the charge in the portion 116' in the column III of the memory part 2 is shifted to the portion 108' of the separating input section 17' through the portion 109' of the separating input section 17'. When the second clock pulse .phi.T is applied, the charge in the portion 104 in the column I of the horizontal register 33 is shifted to the portion 102 of the horizontal register 32 through the portion 103 of the isolate section 52, the charge in the portion 118' in the column II of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 of the separating input section 17', and the charge in the portion 108' in the column III of the separating input section 17' is shifted to the portion 118' of the memory part 2 through the portion 107' of the separating input section 17'. When the third clock pulse .phi.T is applied, the charge in the portion 102 of the horizontal register 32 in the column I is shifted to the portion 100 of the horizontal register 31 through the portion 101 of the isolate section 51, the charge in the column II in the portion 104 of the horizontal register 33 is shifted to the portion 102 of the horizontal register 32 through the portion 103 of the isolate section 52, and the charge in the column III in the portion 118' of the memory part 2 is shifted to the portion 104 of the horizontal register 33 through the portion 105 of the separating input section 17'.

In this manner, the information in the last line of the memory 2 is distributed to the horizontal registers 31-33. In the present embodiment, one line of information is distributed to the horizontal registers 31-33 with one fewer clock pulses than the embodiment of FIG. 9, that is, with three clock pulses .phi.T. Accordingly, in the present embodiment, the number of clock pulses .phi.T applied is one fewer than that shown in FIG. 10(b). The waveforms of the clock pulses shown in FIG. 10(a) are also applicable to the present embodiment. The operation of the horizontal charge transfer after the distribution of the charges is the same as that in FIG. 9.

In the above three embodiments, the separating input sections 17 and 17' impart different amounts of delay to the respective groups of the memory part 2 such that they impart an effective delay of zero picture cells (corresponding to zero bits) to the information in the column I group of the memory part 2, an effective delay of one picture cell (corresponding to one bit) to the information in the column II group and an effective delay of two picture cells (corresponding to two bits) to the information in the column III group, in order to parallel-to-serial convert one line of information. The read-in of the horizontal registers 31-33 is controlled in synchronism with the serial outputs from the separating input sections 17 and 17' so that one line of information in the memory part 2 is read into the horizontal registers 31-33 by column group.

In the above embodiment, one horizontal line of information in the image sensing part 1 is divided into three parts, which are read out by the three horizontal registers 31-33. Thus, the number of bits of each of the horizontal registers 31-33 may be approximately one third of the number of horizontal picture cells of the image sensing part 1 and hence the frequency of the clock pulses .phi..sub.1 -.phi..sub.3 applied to the horizontal registers 31-33 may be reduced by a factor of approximately three. As a result, the power consumption is saved, the noise is reduced and the transfer efficiency is improved. In the embodiments of FIGS. 8-11, the mixing of charges among the horizontal registers 31, 32 and 33 and the memory part 2 in the horizontal charge transfer mode is prevented by the isolate sections 51-53 or the isolate sections 51 and 52 and the separating input section 17' so that the read operation is well carried out.

While the frame transfer type two-dimensional CCD array has been specifically described, it should be understood that the present invention is equally applicable to an interline type two-dimensional CCD or CPD (charge priming device) array.

The horizontal registers serve to separately read in and read out the charges of the respective colors separated by the color separating optical members such as color separation filters. The color separation filter may be a combination of complementary color filters. Instead of the stripe color filter, a mosaic color filter may be used.

If the number of colors separated by the color separating optical member is larger than three, the number of horizontal registers may be more than three.

While the clear drain CD and the clear gate CL for clearing the unnecessary charges are provided in the present embodiment, the color information may be separated without them. The separating input section 17 for the horizontal registers 31-33 may be constructed by the gate electrode.

In accordance with the solid state imaging device of the present invention, the sample and hold circuit for separating the color signals is not necessary or can be extremely simplified and the mixing of the color information in reading out the color information is prevented and high quality of color information is provided. Since the read clock frequency in the horizontal read circuit is significantly lowered, a high transfer efficiency can be maintained even if the number of horizontal picture cells of the image sensing part is increased, that is, when horizontal resolution is increased. In addition, the noise is reduced and the power consumption is saved.