JPS5846908B2 - Television imaging method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a television imaging system for improving images and obtaining high-quality images in a color television camera or the like using a solid-state image sensor.

BACKGROUND ART Conventionally, color television cameras using image pickup tubes such as Brambicon and Vidicon have been put into practical use.

In addition, several types of single-tube color television cameras using stripe filters have been developed for the purpose of reducing cost, size, and weight, and some of them are in practical use.

Recently, however, due to rapid advances in semiconductor technology, various solid-state image sensors that have the potential to replace image pickup tubes are being developed.

CCU (Chargc Coupled Device
An example of such an image sensor is an image sensor called an e-=charge-coupled device.

Now, an explanation will be given by taking an example of a CCD, which is a type of solid-state image sensor.

In order to use it as an image sensor in place of an image pickup tube, the semiconductor depletion layers of MOS capacitors, which are the constituent elements of the image sensor, are arranged vertically and horizontally in a regular manner in the required number in one plane, and images are formed on this plane. By optically forming an image, charges corresponding to the intensity of the light are accumulated in each depletion layer.

This charge can be sequentially transferred by externally applied clock pulses or drive pulses, and image information can be extracted as an image signal.

If [tc] is the time during which the accumulated charge is transferred from one depletion layer to the next by a clock pulse, then [tc] is equal to the period of the clock pulse.

Therefore, the transfer time (tc) can be arbitrarily selected by changing the frequency of the clock pulse.

Now, assuming that the number of electrodes in the depletion layer arranged in a horizontal row is (N), the time (TD) for transferring and reading out the charges in one row is expressed by the following equation.

TD=NXtc This (TD) is set to be equal to one period of horizontal scanning in the television standard system, and (N) and Ctc) are determined, and when the transfer of one horizontal line is completed, the transfer is sequentially performed in the vertical direction. Since it is easy to set each constant so that vertical transfer is completed within a time equal to one period of vertical scanning in the John's standard method, it can be used as a substitute for a conventional image pickup tube.

Specifically, considering the horizontal and vertical retrace time, [N],
(tc) and other constants must be determined.

Such a solid-state image sensor has many advantages such as being small and lightweight, consuming little power, and being highly reliable.

Furthermore, a major feature compared to conventional image pickup tubes is that the geometric distortion of the image is extremely small, as is clear from its configuration and operating principle.

This difference in image scanning mechanism allows a new color television imaging system to be realized.

A conventional color television camera uses one to three image pickup tubes, and extracts a video signal by continuously scanning an image formed on a photocathode horizontally and vertically with a single electron beam.

The geometrical distortion of the image at that time is mostly determined by the linearity of scanning of the electron beam.

Furthermore, when a plurality of image pickup tubes are used, each electron beam must be aligned in time and geometrical position, which requires a high degree of skill and skill.

In a color television camera using three image pickup tubes, image information is separated into three colors by a dichroic mirror, and converted into three color signals by the three image pickup tubes.

In this case, the above-mentioned color registration becomes a particular problem.

In addition, in a color television camera that uses two image pickup tubes, one tube is used for the brightness signal and the other one is used for the color signal.For example, a stripe filter is used to extract the red and blue color signals, and the green signal is the brightness signal. The signal is derived from the red and blue signals using a matrix circuit.

Furthermore, in a color television camera using a single image pickup tube, a complicated stripe filter is used to derive a brightness signal and a color signal.

In order to derive color signals using this stripe filter, phase separation using an IH delay circuit is often used.
Alternatively, it is well known that a frequency separation method is used.

Here, H represents the period of the horizontal scanning line frequency.

Single-tube color television cameras using stripe filters have problems with image quality and interference lines and moiré caused by the stripe filter, but there are no registration problems.

On the other hand, stripe filters used in these applications often have a combination of vertical or diagonal stripes.

Additionally, a Yokozuna stripe filter has been proposed, but it has not yet been implemented because it is technically extremely difficult to reliably scan the electron beam over narrow horizontal stripes.

On the other hand, in a solid-state image sensor, color signals can be regularly selected and derived even in vertical and horizontal stripes.

In a solid-state image sensor, as shown in Fig. 1, the electrodes of a part of the sensor are subdivided and regularly arranged vertically and horizontally, and in addition, they are temporally accurate and are regularly transferred by lock pulses, so that the image can be displayed in a straight line. The characteristics and geometric distortion can be made extremely superior to those of an image pickup tube.

Non-linearity and geometric distortion of images in solid-state image sensors are caused by the uneven arrangement of sensor electrodes and by the optical lens system used for imaging, but both are caused by the image pickup tube used. This is negligible compared to those in the case.

In this way, in a solid-state image sensor, the linearity and geometric distortion are uniquely determined by the geometric arrangement of the sensor electrodes and the optical lens system. It is possible to embody a color television camera that is small and lightweight and has a simple configuration and circuit.

In other words, since it is easy to impart color selectivity to each sensor electrode using the method described below, it is possible to arrange the electrodes for selecting the colors required for color television in accordance with predetermined rules, and to accurately or precisely lock them. If transmitted according to rules determined by pulses and, if necessary, further processed by external electrical circuits, the required color television signal can be produced.

There are several ways to provide unique color selectivity to individual sensor electrodes.

Now, we will discuss 1.2 examples of them.

In the first example, for example, the electrode that you want to have a red color is coated with a paint that transmits only red wavelength light, and the electrode that you want to have a blue color is coated with a paint that transmits only blue wavelength light. It is fixed by vapor deposition, baking, etc. on the electrode surface.

As a result, only light of a specific color component in the image reaches the sensor electrode and accumulates charge.

The second example is an image sensor created by assigning pre-designed color selectivity to individual electrodes on one plane of the image sensor.
A filter with the same color as the turn is made and placed in front of the image sensor. First, an image of the subject is formed on this filter 1 using a daking lens, and then the image transmitted through the filter is formed on the image sensor electrode surface using a relay lens. let

By adjusting the image sensor-relay lens and color filter in advance, each electrode has a color-selective function in the designed color arrangement, and the necessary color television signals can be derived as described above.

As described above, when using an image pickup tube, it was technically difficult to create a color television camera using a color filter with horizontal stripes due to non-linearity caused by the deflection of the electron beam, but solid-state image sensors - can be realized very easily.

Color television cameras using solid-state image sensors are similar to those using image pickup tubes, as well as those using one, two, three, and four solid-state image sensors. I can think of things.

In a single color television camera, a wideband three-color color signal is derived by a dot filter or a stripe filter.

In a color television camera that uses two cameras, one camera is used for brightness signals or broadband line signals, and the other camera is used for color signals, and a dot or stripe filter is used to generate two or three color signals. Deduce.

The color signal in this case may be a narrow band.

Furthermore, in a color television camera using 31, there is a method in which an image separated into three colors by a dichroic mirror is derived as a broadband color signal by three image sensors.

Another method is to use one for brightness signals or broadband line signals, and use the remaining two for separate color signals.

In the latter case, the color signal taken out from the image sensor for color signals may have a narrow band.

In a color television camera using four cameras, one camera is used for the brightness signal, and the other three cameras are used for the color signals of the three colors.

In this case as well, the color signal may have a narrow band.

As noted above, there are many cases where a narrow band color signal is sufficient, but in this case, the number of electrodes arranged in the vertical and horizontal directions of the image sensor used should be compared to that of a wide band one. , can be extremely reduced, making manufacturing easier, improving reliability, and lowering costs.

This can be said to be one of the major features when using a solid-state image sensor.

A new color television camera using the above-mentioned solid-state image sensor uses a solid-state image sensor instead of an image pickup tube, transfers charge using clock pulses instead of an electron beam, and derives brightness and color signals. Ta.

In this case, a method is used in which a series of signals are extracted continuously or serially from one image sensor and separated into separate color signals by an external circuit.

However, in a solid-state image sensor, as can be easily determined from the internal structure, electrodes segmented vertically and horizontally are regularly arranged.

Therefore, the internal design and manufacturing can be such that these electrodes are classified into two or three groups according to a certain rule, and each group is transferred with a different series of clock pulses to extract the charge.

Three groups of clock pulses for two groups;
By applying clock pulses for three groups simultaneously to perform charge transfer to the three groups, simultaneous signals for three groups or three groups can be derived.

For example, as described above, a dot or stripe filter may be used to sort the electrodes into two groups, red and blue, so that the red groups can be tied together and transferred in a series of clock pulses, while the blue If you connect them to one blue fumo and transfer them using another series of clock pulses, and transfer red and blue signals at the same time, you can take out the red and blue signals at the same time.

Further, in some cases, it is possible to easily transfer data with a certain time difference between each group by adjusting the clock pulses of each series.

In this way, the ability to divide one image plane into groups of three or more groups and extract signals for each category while maintaining a defined time relationship is a major feature that could not be achieved with conventional imaging methods.

Utilizing this feature, it is possible to propose a simple, inexpensive, and high-performance color television camera imaging system.

For example, the electrodes in each horizontal row of the image sensor are R, G, B, R, G so that they receive red R, green G, and blue B signals one by one in the horizontal direction.

By configuring a group of three systems of R, G, and B with B... and one line, and transferring them simultaneously using three systems of clock pulses, simultaneous signals of R, G, and B can be derived. I can do it.

In the second column, the first horizontal column shares R, the second column shares G, and the third
The rows share B, and the 4th, 5th, 6th, etc. rows are repeated as R, G, B, etc. to form three groups: R group, G group, and B group. group,
Charges can be transferred simultaneously using three series of clock pulses.
, G, and B simultaneously can be derived.

In the third example, as in the second case, each horizontal column is R, G, B,
Arrange them in the order of R, G... and group them in the second order.
Unlike the case of , there are two groups of odd numbers and even numbers,
If charge is transferred using two series of clock pulses, R
, B, G, R.

Line sequential signals in the order of B, G..., and 0. R,B,
G.

Line sequential signals in the order of R, B, . . . can be simultaneously derived.

These two signals include two colors such as (R, G), (B, R), etc. for one horizontal scanning section, but one color is missing.

However, since this missing color is included in the previous scanning section, if this is utilized using a delay circuit of one horizontal period, three signals of R, G, and B can be obtained. .

Although three examples have been shown above, innumerable combinations can be considered when considering the case of two colors, the case of three colors, the case of dots, and the case of lines.

What is common here is that signals are derived for each category while maintaining a time relationship determined by one or more series of clock pulses.

As already mentioned in the previous section, when transferring charges using two or more series of clock pulses, it is possible to control the temporal relationship between the output signals of each series by adjusting the temporal relationship between the mutual clock pulses. This is a major feature when using an image sensor.

If this property is utilized in a color television imaging system, the configuration of the solid-state image sensor can be simplified and an inexpensive color television camera can be manufactured.

For example, N
In the TSC system, a 1:2 interlace system with 525 scanning lines is adopted.

If the vertical blanking period is 21H (H is one horizontal scanning line period), there are 42H in the odd and even fields, so the minimum number of horizontal columns of the image sensor is 525-42=483.

If you leave some margin and have about 500 columns, you can extract a regular interlaced signal.

Furthermore, if the same electrodes are used in the odd and even fields, about 250 columns are sufficient.

Although it is possible to create a regular interlaced raster on a playback monitor or receiver, since the video signal contains information at the same position in both the odd and even fields, this is not preferable because the vertical resolution will deteriorate.

However, depending on the purpose of use, it is possible to make a color television camera that is sufficiently practical.

In order to perform regular interlacing by transferring charges using the above-mentioned two or three series of clock pulses, approximately 1,000 columns of image sensors are required for two series, and approximately 1,500 columns of image sensors are required for three series. Become.

From the viewpoint of production technology, it is better to have fewer rows, and the gable end will also be improved.

Now, taking the case of two lines as an example, we will explain how to reduce the number of columns to 172, or about 500 lines, and still achieve the same effect as regular interlacing.

The first column is R, the second column is G, and the third column is B, and is further arranged in the order of R, G, B, etc. repeatedly, and 500
Suppose it is made up of columns.

The odd numbers are set as one group, and the even numbers are set as another group.

Charges are transferred using two series of clock pulses, and in odd fields, charges are transferred simultaneously.

That is, the first and second columns correspond to the first horizontal scanning line.
, G signals are output, and the third and fourth columns output B and H signals corresponding to the second horizontal scanning line.

Repeating the same transfer, the odd field transfer is 50
Column 0 is sufficient.

Next, the even field starts from the latter half of the 263rd horizontal scanning line, but since this 1/2 section does not have much to do with the problem being dealt with here, the explanation will start from the 264th horizontal scanning line.

For even fields, the odd column groups receive clock pulses such that the first column transfer starts at the beginning of the 263rd horizontal scan line and the second column transfer starts at the beginning of the 264th horizontal scan line. For even column groups, the clock pulse starts 1H period earlier than the clock pulse.

In this way, adjusting the clock pulses can be easily done with external electrical circuitry.

Therefore, in an even field, the second and third columns correspond to the 264th horizontal scanning line and output G and B signals.

Next, the fourth and fifth columns correspond to the 265th horizontal scanning line, and R
, G signals.

Similarly, the process is repeated to complete the transfer of even-numbered fields.

In the next odd field, 1. The first and second columns start transfer at the same time again.

This situation is illustrated in FIG.

In this case, as can be understood from FIG. 14, the combinations of columns that share the horizontal scanning lines of odd and even fields are alternated, and the same effect as regular interlacing can be obtained.

This concept can be similarly applied to the case of three arrays, and is extremely effective because the array of image sensors can be halved.

In a color television camera that uses two or more solid-state image sensors, one of the image sensors
is used for brightness signals or broadband green signals, and the remaining image sensors are used for color signals.

As is well known, all of the color television systems currently in practical use employ a mixed-picture system, so that the brightness signal has a wide band, but the color signal has a narrow band.

For example, in the NTSC system, the brightness signal is approximately 4.2
Although it has a wide bandwidth of MHz, the color signal is approximately 0.5
A narrow bandwidth of M,′ Hz is sufficient.

Therefore, for the brightness signal, there are currently 400 pieces in a row horizontally and 50 pieces vertically.
Assuming that an image sensor with 0 rows is used, 50 color signal sensors in a horizontal row and 500 rows in a vertical row are enough for practical use.

Furthermore, in the vertical direction, the resolution of the color signal may be worse than that of the brightness signal, so it is recommended to use approximately 250 columns or even approximately 125 columns in the vertical direction for the color signal. I can do it.

Therefore, in the case of 150 columns, the same electrode can be used for both odd and even fields.

Further, in the case of 125 columns, it can be implemented by using the same electrode for both odd and even fields and making the color signal of one column correspond to two horizontal scanning lines using an IH delay circuit.

This is effective for producing inexpensive color television cameras.

The present invention will now be described in detail with reference to the accompanying drawings, but prior to this description, in order to better understand the present invention,
An example of a well-known image sensor will be described with reference to FIG.

The image sensor shown in the figure is a model manufactured by Fairchild, Inc. in the United States, which consists of a large number of photosensitive arrays 11, 21, 31, . . . of photo elements arranged in a matrix.・
. . . A photoaccumulation section consisting of 81, etc., a charge transfer section consisting of analog shift registers arranged alternately in pairs with each of the photoelements, and the final stage or last digit of the analog shift register. It is connected to a horizontal transfer section consisting of an output horizontal analog shift register.

All of the photo elements of these optical storage sections are connected in common, and driving clock pulses (φP) are applied at the same time. 1.3.5...
・Rows, etc. are odd numbered rows and 2nd, 4th, 6th, etc.
・Even-numbered rows are commonly connected, and one clock pulse (φVl) for charge transfer is applied to odd-numbered rows.
Similarly, the other clock pulse (φ
■2) is added, and every other output horizontal analog shift register of the horizontal transfer section is commonly connected to each other, and the odd-numbered side has a clock pulse (
While the clock pulse (φH1) is applied to each of the remaining even-numbered output analog shift registers, the clock pulse (
φH2) is added.

The operation based on the above configuration will be described next.

(1) Depending on the intensity of the irradiated light, charge is accumulated in the depletion layer under the photoelement of the light storage section, and a clock pulse (
φP) is added to control the start point, and at the same time this clock pulse (φP) is made zero, a charge transfer clock pulse (φVl) is generated and applied to the above-mentioned odd row side system in the charge transfer section. Thereby, the charge in the photoaccumulator is transferred to the corresponding analog shift register.

However, at this time, the charges on the odd-numbered rows are transferred, but the charges are not transferred on the even-numbered rows because no driving clock pulse is applied to them.

Therefore, after the first transfer, the charge is 1 for each column.
It will always exist.

Therefore, the data can be transferred to the upper row one row at a time.

(2) Next, at the same time as the above clock pulse (φVl) is generated, another clock pulse (φV2) is generated and applied to the even-numbered row side system in the charge transfer section, and the horizontal analog signal in the horizontal transfer section is As a result of applying the clock pulse (φH1) for reading the output signal to the odd-numbered system of the shift register, the charges in each of the first rows are transferred to the horizontal transfer section and transferred to the other rows (3rd, 5th, 7th,・・・・・・
The charges in the rows) are transferred vertically upward one row at a time.

(3) Then, at the same time as setting the clock pulse (φH) to zero, another clock pulse (φH2) is generated and applied to the even numbered side of the horizontal analog shift register in the horizontal transfer section. Each of the charges in the horizontal transfer section is transferred to the left side.

Then, at the same time as making the output of this clock pulse (φH2) zero, the above clock pulse (φH1) is again added to the odd-numbered side system of the horizontal analog shift register, and the clock pulse (φH1) for reading these output signals and When (φH2) repeats positive and zero, each of the charges in the horizontal transfer section is sequentially transferred to the left side, and is sequentially read out as an output signal at the repetition timing of the clock pulse (φH1).

(4) When the charges in each column of the first row are sequentially taken out to the output and the charges in the last column are taken out to the output, the clock pulse (φV2) for charge transfer is made zero, and the clock pulse (φVl ) is added.

At this time, since the charges in the first third row have been transferred to the second row, the clock pulse (φVl) and (φV2
), the charge is transferred to the horizontal transfer section.

After this, horizontal transfer is performed by the same operation as above.

In addition, as other image sensors, R-CA
There is a model developed by the company, but its contents are disclosed in detail in, for example, Japanese Patent Laid-Open Publication No. 49-66082 filed by the company, so a detailed explanation will be omitted here.

Hereinafter, one embodiment of the present invention will be described based on the drawings.

Figure 1 shows the arrangement of sensor electrodes in a wideband solid-state image sensor. N electrodes are arranged horizontally and M vertically, and are used for brightness signals or broadband color signals. .

Further, FIG. 2 shows an arrangement of sensor electrodes of a narrowband solid-state image sensor, in which P electrodes are arranged horizontally in a row and Q electrodes are arranged vertically.

Regarding these solid-state image sensors, it is reasonable to use both types depending on the desired performance.

However, although some of the sensor electrodes in the solid-state image sensors shown in FIGS. 1 and 2 are omitted and simplified in the drawings, they are substantially the same as those in the image sensor shown in FIG. 31 above. - Constructed with almost the same basic arrangement structure as in the case of -, a light storage section consisting of photosensitive rows of a large number of photoelements arranged in a matrix, and an analog shift register arranged alternately in pairs with each photoelement. There is a charge transfer section for temporarily storing charge from the optical storage section and then transferring it by driving with a clock pulse, and a charge transfer section for temporarily storing charge from the charge transfer section by driving with another clock pulse and rearranging the arrangement. It includes a horizontal transfer section for outputting a signal by transfer, and each of these components operates in substantially the same manner as the basic operation of an image sensor using a charge-coupled device.

Therefore, by applying the above-mentioned basic operating principle, the drive method using clock pulses can be used not only in one series, but also in multiple signals by transferring charges using clock pulses such as triple series or triple series. It can be taken out and can be made to correspond to the embodiments shown in and after FIG. 3 according to the present invention.

Figure 3 is an example of a case where signals are extracted using the dot sequential method, where each horizontal electrode is divided into R, G, B by a dot filter.
, R, G, and B in that order, each column in the vertical direction is arranged in the same way, and charges are transferred using one series of clock pulses.

Since B, R, G, and B point-sequential signals can be derived, they can be separated into separate R2O and B color signals using sampling pulses in an external electric circuit.

Combinations of color arrangements only need to follow certain rules, and may be two colors, three colors, or even many colors, and there are countless combinations.

Furthermore, the order of colors may be changed for each column.

Figure 4 is an example of a dot sequential method using two series of clock pulses. In the figure, the electrodes in each column share two colors in the order of R5B, R, and B, and are divided into an R group and a B group. Charge is transferred using two series of clock pulses, and R and B signals are simultaneously extracted.

The combination of colors may be two or more colors, and the order may be changed for each column, but it is sufficient that they are arranged according to a certain rule.

However, it is necessary to separate the necessary color signals using sampling pulses in an external electric circuit.

FIG. 5 is an example of a point-one time order system using three series of clock pulses, but the concept is the same as the two series case of FIG. 4, so a detailed explanation will be omitted.

From FIG. 6 onwards, several examples are mainly shown where the line sequential method is used.

Each column is assigned a specific color, and a horizontal stripe filter is used.

FIG. 6 shows an arrangement that allows regular interlacing, and the NTSO system requires approximately 500 columns of M.

For odd numbers, the 1st, 3rd and 5th are arranged in the order of R, G, B,
For even number rows, arrange the 2nd degree 4.6 in the order of B, R, G,
By sequentially transferring charges of odd numbered columns to odd fields and even numbered columns to even numbered fields by one series of clock pulses, line sequential signals of R, G, and B can be derived.

It is already well known that this R, G, B line sequential signal can be converted into R, G, B simultaneous signals using two IH delay circuits and a switching circuit.

FIG. 7 is a simplified version of FIG. 6, in which the same electrode rows are used in the odd and even fields.

Therefore, in the case of the NTSO system, M is approximately 250 columns.

The output is a line sequential signal of R, G, and B.

FIGS. 8 and 9 correspond to FIGS. 6 and 7, and are a method for extracting two-color line-sequential signals, and are mainly used in color televisions using two solid-state image sensors. Used for camera color signals.

10 to 16 describe an example of a method for simultaneously obtaining two line sequential signals using two series of clock pulses.

In Figure 10, the odd numbered rows are R1, the even numbered rows are B, and the first and second
The column corresponds to the first scanning line, the fifth and sixth columns correspond to the second scanning line,
The 3rd and 4th columns are the 264th scanning line, and the 7th and 8th columns are the 265th scanning line.
If the structure is configured to correspond to the scanning lines, R is set as one group, and B is set as the other group, and charges are transferred using two series of clock pulses, normal interlaced R,
B simultaneous signals are obtained.

In this case, the NTSC system requires about 1000 columns of M, which poses a problem in terms of production technology.

As in the case of FIG. 10, FIG. 11 shows a method in which two color signals are obtained simultaneously by two series of clock pulses, but the same electrode array is used in the odd and even fields.

Therefore, M may be 1/2 of that in the case of FIG.

In Figure 12, each column is R, G, B, R, G, B...
Charges are transferred simultaneously in two rows in each row in this order, with the first and second columns corresponding to the first scanning line, the third and fourth columns corresponding to the second scanning line, and even in even fields. , 2 columns correspond to the 264th scanning line, and the same electrode is used in odd and even fields, and approximately 500 columns of M are required.

In this case, since charges are transferred using two series of clock pulses, the output of the first series is a line sequential signal of R, B, G, R..., and the output of the second series is a line sequential signal of G, R, B, G. The line sequential signal of . . . becomes the output signal.

To obtain a simultaneous three-color signal from these two signals, the 13th
You can do it as shown in the figure.

In FIG. 13, 3 is an IH delay circuit, 4 is a signal switching circuit, and 5 is a switching pulse generation circuit. When the first and second series signals are applied to terminals 1 and 2, the output of delay circuit 3 changes. Since the second series signal comes out with a delay of 1H, the first
, for example, when the signals of the second series are G and B, respectively, the output of the delay circuit 3 becomes R, and the three colors are simultaneously transmitted to the signal switching circuit 4.
Since the color changes in a ring shape for each IH, this switching circuit 4 connects RlG and B to terminals 6 and 7.8, respectively.
continuous simultaneous signals can be derived.

FIG. 14 shows a further improvement to the example of FIG. 12, and the interlacing is improved.

The same solid-state image sensor as shown in FIG. 12 can be used.

That is, each column is arranged in the order of R, G, B, R, G, B, . . . , and charges are transferred two columns at a time.

The difference from the columns in FIG. 12 is that the combination of two columns of even fields is shifted by one column relative to the odd fields.

The 1st and 2nd columns and the 3rd and 4th columns correspond to the 1st and 2nd scanning lines, whereas in the even field, the 2nd and 3rd columns and the 4th and 5th columns correspond to the 264th and 265th scanning lines, respectively. Adjust the start of the clock pulse using an external electrical circuit to correspond to the

In the case of FIG. 14, it is sufficient to start transferring the first stream earlier than the second stream by IH at the start point of the even field.

However, if the two series of signals obtained in this way are added to the circuit of FIG. 13, a problem will occur, so it is necessary to pass them through an additional circuit as shown in FIG. 15.

In FIG. 15, numbers 1 to 8 have the same configuration as in FIG. 13, but odd and even field switching circuits are added in front of terminals 1 and 2.

9 is a signal switching circuit, and 10 is a switching pulse generation circuit.

Then, the signal switching circuit 9 switches the input signals 1L12 to terminals 1 and 2 in the odd field section, and outputs the input signals 11 and 12 to the terminals 2 and 1 in the even field, respectively. .

In this way, the input to the signal switching circuit 4 is always R.
, G, and B are input in a ring shape according to the rules, so that continuous signals of the three colors can be simultaneously output to the terminals 6, 7, and 8.

Note that (even) in the figure means an even field, and an example thereof is shown.

Figure 16 shows a different arrangement of colors in each column of Figure 12.
Even-numbered columns are alternately R and B, and odd-numbered columns are all G.

Therefore, since the signal output of the odd numbered columns is a continuous G signal, and the even numbered columns is a line sequential signal of R and B, it is well known that continuous simultaneous signals of RlB can be derived using an IH delay circuit.

Figure 17 transfers charges using three series of clock pulses and simultaneously extracts three series of signals.
This corresponds to the two series shown in FIG. 11, in which the same electrode is used in both odd and even fields to obtain simultaneous R, G, and B signals.

In order to improve the interlacing, the first
Although an arrangement corresponding to Figure 0 is conceivable, illustration thereof is omitted.

In this case, approximately 1500 columns of M are required, which is technically more difficult.

Figure 18 was devised to obtain a good interlaced image by setting M to about 750 columns, and like the case of Figure 14 in the case of 2 series, each scanning line corresponds to each scanning line in both odd and even fields. This is a different combination of columns.

In the odd field in FIG.
Column 2.3 corresponds to the first scanning line, column 4.5, and 6 correspond to the second scanning line, and the following is repeated sequentially, and in even fields, column 3, 4, and 5 correspond to the 264th scanning line. The 6th, 7th, and 8th columns correspond to the 265th scanning line, and the start of the clock pulse is adjusted by arranging them so that they are repeated sequentially.

In order to do this, in the even field, the first
The clock pulses used to take out the signals of the series R and the second series G may be adjusted by an external electric circuit so that they start earlier than those of the third series B by an interval '1H.

Although not shown, in the even field, only R of the first series starts early in the IH section, and the 2nd and 3rd series start early.
．． The four columns can be combined to correspond to the 264th scanning line.

However, these two methods are not sufficient from the viewpoint of interlacing, and the interlacing can be improved, for example, by the method of FIG. 19.

The image sensor and charge transfer are exactly the same as in the case shown in Fig. 18, and only the even fields are IH controlled by an external electric circuit.
The G signal included in the previous scanning line is used through an IH delay circuit, and four columns of signals correspond to one scanning line.

In the figure, the 2nd 3.4.5th column corresponds to the 264th scanning line, and the 5th, 6th, and 7th 8th columns correspond to the 265th scanning line.

In this way, the scanning lines of the even field are inserted between the scanning lines of the odd field, and good interlacing can be obtained.

Further, to create such a signal, there is a method as shown in FIG. 20, for example.

A second series G signal is applied to the terminal No. 13.

15 is an IH delay circuit, and 14.16 is a buffer amplifier, but these are not essential.

17 is a signal switching circuit, which applies the signal from the buffer amplifier 14 to the signal addition circuit 19 in odd fields, and adds the signal from the buffer amplifier 16 to the signal addition circuit 19 in even fields.
Switch to add to.

Reference numeral 13 denotes a pulse generation circuit for switching between odd and even fields.

In this way, the output signal to the terminal 20 is obtained by adding the input signal to the terminal 13 and the signal from the buffer amplifier 14 in the signal addition circuit 19 in the odd field, but the second and fifth signals, which are originally the same signal, are added. The column G signal comes out as is.

On the other hand, in an even field, the input signal to the terminal 13 and the signal from the buffer amplifier 16 are added in the signal addition circuit 19, so two adjacent G signals are added.

In this way, the method described above can be implemented.

At this time, it is also possible to set the buffer amplifier 16 to have a narrow band (low-pass characteristic) to prevent deterioration in resolution caused by addition of two adjacent G signals.

In FIG. 21, four series of clock pulses are used, and the combinations of odd and even field columns are changed to improve interlacing.

In each column, colors are shared by repeating RlG, B, and G.

In odd-numbered fields, data is transferred four columns at a time, but the G signals of the fourth, 8th, 12th, . . . columns are not used.

1.2.3 rows R, G, B as the first scanning line, 5°6
．． Seven columns of R, G, and B are made to correspond to the second scanning line.

On the other hand, in an even field, the transfer of the first series is started earlier than the others by the LH interval.

The G signal of the second series in Section 2.6.10 is not used in even fields.

Therefore, B, G, and R in columns 3, 4, and 5 correspond to the 264th scanning line, and B, G, and R in columns 7, 8, and 9 correspond to the 265th scanning line.

The G signal of the second series is only for odd fields, the G signal of the fourth series
If the signal is to be used only in the even field, this can be easily implemented using a pulse generator and signal switching circuit for switching between the odd and even fields.

An example thereof is shown in FIG. 22.

In the figure, the second series G signal and the fourth series G signal are applied to terminals 21 and 22, respectively.

23 is a signal gate circuit that allows only odd field signals to pass; 24 is a signal gate circuit that allows only even field signals to pass; 25 is a switching pulse generation circuit; 26 is an addition circuit for the selected two signals; A continuous G signal is output to the terminal.

FIG. 23 shows an example in which two solid-state image sensors are used, one for a brightness signal (hereinafter referred to as a Y signal) and one for a color signal.

As mentioned above, the color signal can be in a narrow band, so
If the same electrodes are used in both the odd and even fields, the number of columns can be approximately 1/2 that for the Y signal, and deterioration in image quality will hardly be a problem.

In the figure, the color signals are shown in which each column is arranged in R, B, R, B, but a dot sequential signal may be used, and many examples are possible.

Figure 24 shows an example in which three solid-state image sensors are used, one for the Y signal and the other two for the color signal, for example, for the R signal and the B signal. Transfers charge.

That is, the first column of color signals is transferred simultaneously with the first column of Y signals.

When the Y signal is being transferred in the third column, the transfer of the chrominance signal is suspended, and in that period, the chrominance signal in the first column is used through the IH delay circuit.

Next, when the Y signal is transferred to the fifth column, the color signal is simultaneously transferred to the third column.

Repeat the same process below. With this method, the number of columns for color signals can be reduced to about 172 for Y signals.

FIG. 25 is a simplified version of the example shown in FIG. 24, in which the same electrode is used for both odd and even fields for color signals.

In this way, approximately 1/4 of the column for the Y signal is used for the color signal.
will be sufficient.

Next, supplementary explanation will be given of a method for improving interlacing in the case where charge is transferred using one series of clock pulses and line sequential signals are extracted and used.

Figure 26 shows R, G, B... according to the method shown in Figure 7.
2 is a block diagram illustrating a method for improving interlacing when a signal is extracted using a line-sequential method.

In the figure, 28 is an input terminal for R, G, and B line sequential signals, 29, 30, and 31 are IH delay circuits, 32 is an odd and even field switching signal generator, 33 is a signal switching circuit, and 34 is an addition terminal. circuit, 35 is a ring counter that generates a pulse every IH in the 3H interval, 36 is a signal switching circuit, 3
7.38° and 39 are output terminals for R, G, and B continuous signals, respectively.

First, the operations of the odd/even field switching signal generator 32, signal switching circuit 33, and addition circuit 34 will be explained.

The signal from the terminal 28 that enters the signal switching circuit 33 and the signal that enters from the delay circuit 31 have the same color, but there is a time difference of 3H.

By adjusting the polarity of the pulse of the odd/even field switching signal generator 32, the signal switching circuit 33 selects the signal input from the terminal 28 and inputs it to the adding circuit 34 when the field is an odd field, and delays it when the field is an even field. The signal input from the circuit 31 is selected and input to the adder circuit 34.

Therefore, the output of the adder circuit 34 is that in odd fields, the signal of each scanning line is output as is, and in even fields, the arithmetic average of two adjacent signals of the same color (with a time difference of 3H) is sequentially output. .

Note that the combination of the delay circuits 29 and 30, the ring counter 35, and the signal switching circuit 36 allows R and G signals to be output.

Since it is well known that the B line-sequential signal is converted into R, G, and B simultaneous signals, a detailed explanation will be omitted.

In this way, performance can be improved by changing the number of electrode columns that combine colors corresponding to each scanning line in both odd and even fields.

FIG. 27 is a block diagram showing a method for improving image quality when R2H line sequential signals are extracted using the methods shown in FIGS. 8 and 9.

In the figure, 40 is R1, Bl, R2, B2...
・Input terminals for line sequential signals, 4L44, 45 are IH delay circuits, 42 is a switching pulse generation circuit, 43 is a signal switching circuit, 46.47 is an addition circuit, 48.49 is an output of R signal and B signal, respectively. It is a terminal.

Among these, delay circuit 41. The switching pulse generation circuit 42 and the signal switching circuit 43 are circuits for converting a well-known line sequential signal into a simultaneous signal, and the outputs of the signal switching circuit 43 are R1 and R1°R.
2, R2... and Bl, Bl, B2. B2...
. . . are two series of simultaneous signals.

These signals are the same signal repeated twice, so when they pass through the delay circuit 44 and addition circuit 46, the arithmetic average of the signals before a certain scan line is output, for example, R1, (R1+
R signals such as R2)/2, R2, and (R2+R3)/2 are obtained.

That is, in the section where B is scanned, conventionally R1 is
Interpolation was performed with a delay of H, but in this method, (R
The image quality is improved by interpolating by 1+R2)/2.

FIG. 28 shows an attempt to improve the interlacing by adding more information to FIG. 27 in the case of FIG. 9.

In the figure, 48 is the same as the R signal output terminal in FIG. 2T.

50 is an IH delay circuit, 51 is an adder circuit, 52 is an odd number,
53 is a signal switching circuit, and 54 is a signal output terminal.

The input signal from terminal 48 is R1, (R1+R2)/2.
Since it is a continuous R signal of R2゜(R2+R3)/2..., the output of the adder circuit 51 that passes through the delay circuit 50 and the adder circuit 51 is delayed by IH and becomes (3R1+R2)/4.
, (R1+3R2)/4 , (3R2+R3”)/4
······become that way.

Therefore, interlacing can be improved by switching these two signals so that in odd fields, the input from terminal 48 is output as is to terminal 54, and in even fields, the output from adder circuit 51 is output to terminal 54.

The method shown in FIGS. 27 and 28 can also be applied to brightness signals, and in the NTSC system, a good image can be obtained by using an image sensor with M of approximately 250 columns.

This method can also be applied to the R and B signals shown in FIGS. 24 and 25.

Furthermore, the concept shown in Fig. 27 can be applied to the case of point sequential signals such as those shown in Figs. 3, 4, and 5, and can be easily implemented by changing the time of the delay circuit. can.

Figures 26, 27, and 28 only show one example of improving image quality, and there are other methods to achieve the same effect.

In the explanations so far, the scanning line numbers, etc., are referred to as NT for convenience.
Although the SG method has been explained as an example, other methods such as FAI and SECAM methods can be handled in exactly the same way.

Furthermore, although the description has been made using R, G, and B for color signals, and R and B for two colors, it is also possible to use other colors without causing any problems.

Furthermore, it is already common knowledge that the G signal is sometimes used as a substitute for the Y signal.

Furthermore, the color of the filter does not necessarily have to match the color of the desired color signal to be extracted, as long as the desired color signal can be obtained by an external electric circuit.

In a television system with 1:2 interlacing, the end of the odd field ends at the 1/2H point, and the beginning of the even field starts from the 1/2H point, so the image sensor has electrodes arranged accordingly. −
Although this point is omitted since it is not directly related to the essence of the present invention, it is possible to leave some margin in the column configuration and process it later using the blanking signal.

In the explanation of the above specific example, the electrodes of the image sensor are arranged vertically and horizontally as shown in Figures 1 and 2 to simplify the explanation, but they can be arranged at any angle as shown in Figure 29. Even if it is arranged diagonally, it can be handled in the same way without detracting from the spirit of the present invention.

Moreover, FIG. 30 shows an example of the relationship between the color assignment and signal extraction of each electrode, and the A and A' series, the B and B' series, and the C and C' series are collectively signaled. If you extract , it will be the same as in Fig. 5, and of course it can be applied to Figs. It may be done in this order.

In any case, since the shared colors can be placed close to each other, it is possible to provide an image with excellent color resolution.

Therefore, a color television camera using the method described above can be extremely small and lightweight, and can operate stably.

In the future, it is expected that a video recording device will be integrated into the device according to the present invention, and it will be realized as a portable imaging and recording device.

[Brief explanation of the drawing]

The drawings are for explaining each embodiment of the present invention,
Figures 1 and 2 are partially omitted arrangement diagrams of the image sensor.
Figures 3 to 5 are partially omitted arrangement diagrams of the image sensor showing the color assignments of each electrode and signal extraction relationships, and Figures 6 to 11 and 19 are arrangement diagrams of the image sensor and scanning lines. Figures 13 and 15 are block diagrams for obtaining simultaneous signals from the image sensor signals, and Figure 20 shows how to extract signals in the case of Figure 19. The block diagrams, FIG. 12, FIG. 14, FIG. 16 to FIG.
The figures are a block diagram showing how to extract signals in the case of Fig. 21, Figs. 23 to 25 are relational diagrams showing the relative positions between multiple image sensors, and Figs. 26 to 28 are diagrams showing how to improve image quality. Fig. 29 is a partially omitted arrangement diagram of another image sensor corresponding to Figs. 1 and 2, and Fig. 30 shows the assignment of each electrode, color and signal extraction relationship FIG. 31 is a partially omitted arrangement diagram of an image sensor showing an example, and FIG. 31 shows the configuration and signal processing relationship of an image sensor according to a conventional method in order to explain the operating principle of the television manipulation method according to the present invention. FIG.

Claims (1)

[Claims] 1. In a television camera using a solid-state image sensor as an imaging element, a 1H delay circuit 41, a switching pulse generation circuit 42, and a signal switching circuit that converts a line sequential signal into a simultaneous signal using the switching pulse. 43, an intermittent signal is derived for every other horizontal scanning line, and in each of the odd and even fields, each intermittent signal derived above is converted into a missing part. The direct signal, which is two adjacent signals, and the signal via the IH delay circuit 4L45 are added to the first adder circuit 46.4.
The first adding circuit output signal 48.49 which was added by 7 to form the missing signal, the first adding circuit output signal 48.49 and the first adding circuit output signal are sent to the delay circuit 50 of the IH. The second addition circuit output signal obtained by adding the delayed signal and the second addition circuit output signal by the second addition circuit 51 is alternately switched for each field by a signal switching circuit 53 that is switched by an odd-even field switching pulse generation circuit 52. A distinctive television imaging method.