JPH11196429A - Character picture restoration device, electronic still camera system and electronic still camera - Google Patents

Abstract

PROBLEM TO BE SOLVED: To restore a picture that is highly detailed and excellent in readability from a character picture consisting of plural kinds of colors. SOLUTION: A processing change-over part 40 transfers a video signal to a character picture restoration part 42 when the video signal which has a missing color signal restored from the video signal from a single plate CCD and is retained in an output part 38 is a character picture and a background color estimation part 52 estimates the background color by making the video signal within a specified luminance range from the video signal as a whole average. A local area extraction part 56 divides the video signal into local areas, converts it into a single plate state and an R signal maximum and minimum calculation part 64, a G signal maximum and minimum calculation part 66 and a B signal maximum and minimum calculation part 68 calculate the maximum and minimum range for each local area. A character color decision part A100 estimates a character color from the maximum and minimum ranges and the background color on a decision table 70 and a background restoration part 74 and a missing color restoration part A104 are made to restore the character picture.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic still camera using a single-chip or two-chip type solid-state image pickup device and an electronic still camera system using the same as a component. The present invention relates to a character image restoring device for restoring an image with excellent readability. The present invention also relates to an electronic still camera and an electronic still camera system to which such a character image restoring device is applied.

[0002]

2. Description of the Related Art In recent years, electronic still cameras using inexpensive and lightweight single-chip solid-state imaging devices (CCD) have become widespread. In such a single-chip solid-state imaging device, color filters are arranged in a mosaic pattern on a light receiving surface in order to obtain color information of a subject from one imaging device.

[0003] For example, FIG. 10 (B) shows commonly used cyan (Cy), magenta (Mg), and yellow (Y).
5E is a diagram showing the arrangement of green (G) complementary color mosaic filters. In the figure, n lines and n
The luminance signal corresponding to the +1 line is represented by Y _{e, n} , Y _{e, n + 1} , and the color difference signal is represented by Ce _{, n} , Ce _{, n + 1.} Assuming that the luminance signals to be _processed are Y _{o, n} and Y _{o, n + 1} and the color difference signals are C _{o, n} and C _{o, n + 1} , these signals are expressed by the following equations.

_{Yo, n} = _{Yo, n + 1} = _{Ye, n} = _{Ye, n + 1} = 2R + 3G + 2B (1) _{Co, n} = Ce _{, n} = 2R-G (2) C _{o, n + 1} = Ce _{, n + 1} = 2B-G (3) where Cy, Mg, and Ye are represented by G and red (R) and blue (B) by the following equations.

Cy = G + B (4) Mg = R + B (5) Ye = R + G (6)

[0006]

As shown by the above equation (1), a luminance signal is generated for all lines of even and odd fields. In contrast, the above equations (2) and (3)
As shown by the equation, two color difference signals are generated only for each line. Therefore, conventionally, the missing line is supplemented by linear interpolation. And after this,
By performing a matrix operation, three primary colors of R, G, and B can be obtained.

However, in such a method, the color difference signal has only one half the information amount of the luminance signal, and a false signal called color moire occurs at an edge portion. Therefore, in the case of capturing an image having a steep edge portion such as a character image, the readability is significantly reduced.

[0008] Such a problem also occurs in an electronic still camera using a two-chip solid-state imaging device.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and has been made in consideration of an electronic still camera using a single-chip or two-chip solid-state imaging device and an electronic still camera system including the same as a constituent element. The present invention provides a character image restoring apparatus capable of restoring a character image composed of a high-definition image with excellent readability, and by providing such a character image restoring apparatus, a character image composed of a plurality of colors can be enhanced. It is an object of the present invention to provide an electronic still camera and an electronic still camera system capable of restoring an image with excellent readability in detail.

[0010]

According to a first aspect of the present invention, there is provided a character image restoring apparatus according to the first aspect of the present invention, comprising a character image composed of a plurality of colors from a video signal of a single-plate or two-plate solid-state imaging device. A character image restoration device for restoring, a background color estimating means for estimating a background color from the video signal of the solid-state imaging device, extraction means for extracting a local area of a predetermined size from the video signal of the solid-state imaging device, A character color estimating means for estimating the character color in the local area extracted from the extracting means; and the local area based on the background color estimated by the background color estimating means and the character color estimated by the character color estimating means. Separating means for separating the video signal in the image into a background and characters,
A missing color restoring means for restoring a color signal missing by the background color when the background is set by the separating means, and restoring a color signal missing by the character color when the character is set by the separating means. , Is provided.

That is, according to the character image restoring device of the present invention, the background color is estimated by the background color estimating means from the video signal of the solid-state image sensor, and the extracting means is used to estimate the background color from the video signal of the solid-state image sensor. Is extracted, and the character color in this local region is estimated by the character color estimating means.
Then, based on the background color estimated by the background color estimating unit and the character color estimated by the character color estimating unit, the separating unit that separates the video signal in the local region into the background and the character is set as the background. In this case, the missing color restoration means
When the color signal missing due to the background color is restored, and when the color signal is regarded as a character, the missing color restoring means restores the color signal missing due to the character color.

According to a second aspect of the present invention, there is provided a character image restoring apparatus for restoring a character image composed of a plurality of colors from a video signal of a single-chip or two-chip solid-state imaging device. Background color estimating means for estimating a background color from a video signal of the image sensor, extracting means for extracting a local region of a predetermined size from the video signal of the solid-state image sensor, and a background color estimated from the background color estimating means. Separating means for separating the video signal in the local region into a background and others,
If the background is set by the separating means, the color signal missing due to the background color is restored, and if otherwise, the color signal is lost due to the averaged video signal of the other area in the local area. A missing color restoring means for restoring a color signal.

That is, according to the character image restoring device of the present invention, the background color estimating means estimates the background color from the video signal of the solid-state imaging device, and the predetermined size from the video signal of the solid-state imaging device. The video signal in the local area extracted by the extracting means for extracting the local area is separated into a background and others by the separating means based on the background color estimated by the background color estimating means. When the background is set by the separating unit, the missing color restoring unit restores the missing color signal due to the background color. The missing color signal is restored by the averaged video signal of the other areas in the area.

An electronic still camera system according to a third aspect of the present invention is an electronic still camera system including an electronic still camera using a single-chip or two-chip solid-state imaging device and a post-processing device. A general-purpose image restoration unit that restores a color signal missing from a video signal by interpolation, and a character image composed of a plurality of colors from the video signal restored by the general-purpose image restoration unit is suitable for a character image different from the interpolation. Character image restoring means for restoring by the restoring method, and transmission means for transmitting the video signal restored by the general-purpose image restoring means to the character image restoring means. It is characterized by being comprised in.

That is, according to the electronic still camera system of the third aspect, the color signal missing from the video signal of the solid-state image sensor is restored by interpolation by the general-purpose image restoration means, and the general-purpose image is restored by the transmission means. The video signal restored by the restoration unit is transmitted to the character image restoration unit configured in the post-processing device. Then, the character image restoring means restores a character image composed of a plurality of colors from the video signal restored by the general-purpose image restoring means by a restoration method suitable for a character image different from the interpolation.

An electronic still camera according to a fourth aspect of the present invention is an electronic still camera using a single-chip or two-chip solid-state imaging device, wherein a color signal missing from a video signal of the solid-state imaging device is interpolated. General-purpose image restoration means for restoring,
A character image restoring means for restoring a character image composed of a plurality of types of colors from the video signal of the solid-state imaging device by a restoring method suitable for a character image different from the interpolation, the general-purpose image restoring means, and the character image restoring means And processing switching means for switching between the two.

That is, according to the electronic still camera of the fourth aspect, the processing switching means includes a general-purpose image restoration means for restoring, by interpolation, a color signal missing from the video signal of the solid-state imaging device, and a processing device for the solid-state imaging device. A character image restoring means for restoring a character image composed of a plurality of colors from a video signal by a restoration method suitable for a character image different from the interpolation is switched.

[0018]

Embodiments of the present invention will be described below with reference to the drawings.

[First Embodiment] FIG. 1 is a diagram showing a configuration of an electronic still camera system according to a first embodiment of the present invention. The electronic still camera system according to the first embodiment is a general-purpose camera. An electronic still camera 10 having an image restoring function, and an adapter 1 having a character image restoring function and a printer output function, which is appropriately connected to the electronic still camera 10
And 2.

Here, in the electronic still camera 10, the subject image transmitted through the lens system 14 and the low-pass filter 16 is guided to a single-plate CCD 18. The CCD1
The output image signal 8 is connected to process circuits 24, 26, and 28 via an A / D conversion circuit 20 and a color separation circuit 22. Output signals of the process circuits 24, 26, and 28 are transferred to an R signal buffer 32, a G signal buffer 34, and a B signal buffer 36 via a matrix circuit 30. Each output signal of the R signal buffer 32, the G signal buffer 34, and the B signal buffer 36 is transferred to an output unit 38 such as a memory card.

The signal from the output section 38 is transferred to the connected adapter 12. In this case, the adapter 12
May be attached to, for example, the back cover of the electronic still camera 10 to transmit and receive signals by abutting each other's connection terminals, or provided to a personal computer in the form of application software. Signals may be transmitted and received by connecting a predetermined output terminal of the electronic still camera 10 to an interface unit of a personal computer via a serial cable or a parallel cable. Further, in the latter case, if the output unit 38 of the electronic still camera 10 is a removable storage medium such as a memory card, signals are transmitted and received via a drive device for the storage medium provided in a personal computer. Such a configuration is also possible.

In the adapter 12, a signal from the output section 38 of the electronic still camera 10 is selectively transferred to a printer output section 44 by a processing switching section 40 via a character image restoring section 42 or directly.

The CCD 18 of the electronic still camera 10
Is supplied with a drive signal generated by a CCD drive circuit 48 based on a clock signal from a clock generator 46. The control unit 50 such as a microcomputer includes the output unit 38, the clock generator 46, and the adapter 12 in the electronic still camera 10.
It is connected to a processing switching unit 40, a character image restoring unit 42, and a printer output unit 44, and controls these units.

Next, the flow of signals in the electronic still camera system having such a configuration will be described. The arrows in FIG. 1 indicate the signal flow direction, the bold lines indicate video signals, and the thin lines indicate control signals.

The electronic still camera 10 includes a lens system 14
And a single-plate CCD 18 via a low-pass filter 16. Here, the CCD 18 is, for example, as shown in FIG.
(A) is a CCD having an RGB primary color type filter arrangement. The CCD 18 has a CCD driving circuit 48 based on a clock signal from a clock generator 46.
Is driven by a drive signal from the camera to capture a subject image. The video signal from the CCD 18 is digitized by an A / D conversion circuit 20 and then converted to a known color separation circuit 2.
2. R in a three-plate state in which missing pixels are interpolated via the process circuits 24, 26, 28 and the matrix circuit 30.
It is converted into a GB signal, and stored in the R signal buffer 32, the G signal buffer 34, and the B signal buffer 36 for each color signal. These color signal buffers 32, 34,
The color signal in 36 is transferred to the output unit 38 under the control of the control unit 50.

The RGB signals in the three-plate state from the output unit 38 of the electronic still camera 10 are transferred to the character image restoration unit 42 or the printer output unit 44 via the processing switching unit 40 in the adapter 12. This switching is performed according to an instruction transmitted from the outside to the control unit 50, for example, by a manual mode switching switch (not shown) such as a character mode / general image mode.

The character image restoring unit 42 restores a predetermined small number of colors, for example, seven color characters of red, blue, green, cyan, magenta, yellow, and black, to a more readable image. The details will be described later. The RGB signals restored to a more readable image by the character image restoration unit 42 are output to the printer output unit 44.

FIG. 2A shows an example of a specific configuration of a filter arrangement in the CCD 18 of FIG. here,
There is a basic arrangement having a size of 2 × 2 pixels, and this basic arrangement is repeatedly repeated to fill all pixels on the CCD. In the present embodiment, the appearance frequencies of the three RGB signals are set to be 1: 2: 1.

FIG. 3 is a diagram showing a basic configuration of the character image restoring unit 42 in the adapter 12. The symbol a in FIG.
a is a three-plate video signal in which missing color signals are interpolated,
bb is a background color estimation signal, cc is a single-plate state video signal (for each color) in the local area, dd is the maximum value and the minimum value of each color signal in the local area, ee is a signal indicating the contents of the determination table, ff denotes a signal indicating a method of calculating a character color and a missing color in the local region, and gg denotes a video signal of a character image of a three-plate state in which the missing color signal is restored.

Thus, the control unit 50 of the electronic still camera
When the processing image is switched to the character image restoration unit 42 by the processing switching unit 40 based on the above control, all the signals from the output unit 38 are transferred to the background color estimation unit 52. In the background color estimation unit 52 calculates R regions within a predetermined luminance range, G, B averages of the respective color signals _{(R B, G B, B} B) , and character color determination unit as a background color Transfer to A100.

After the calculation of the background color, the processing switching section 40 again controls the output section 38 under the control of the control section 50.
Are transferred to the local area extraction unit 56. The local area extracting unit 56 sequentially extracts a rectangular area of a predetermined size, for example, an 8 × 8 size, without duplication based on the control of the control unit 50, and blocks the R signal block buffer 58, the G signal block buffer 60, Transfer to the B signal block buffer 62. At this time, the transferred signal is converted into a single-plate state and transferred based on the CCD filter arrangement shown in FIG. The color signals in each of the signal block buffers 58, 60, and 62 are transferred to an R signal maximum / minimum calculation unit 64, a G signal maximum / minimum calculation unit 66, and a B signal maximum / minimum calculation unit 68. Maximum value (R
_max , G _max , B _max ) and the minimum value (R _min , G _min , B
_min ) is calculated, and the results are transferred to the character color determination unit A100.

The character color determination unit A100 determines the distribution ranges ΔR, ΔG, and ΔB of the R, G, and B color signals in the local area.

[0033] _{ΔR = R max -R min ΔG =} G max -G min ... (7) ΔB = B max -B min thus determined distribution range [Delta] R, .DELTA.G, .DELTA.B is, the background color R _B, G _B, B _In the case of 1/2 or more of _B, the flags R_flag, G_flag, B_flag relating to the color signal are turned on, and in other cases, they are turned off.

[0034] _{if (ΔR> R B / 2} ) then R_flag = ON else R_flag = OFF if (ΔG> G B / 2) then G_flag = ON else G_flag = OFF if (ΔB> B B / 2) then B_flag = ON else B_flag = OFF (8) Next, the character color determination unit A100 obtains the character color in the local area from the flag for each color signal based on the determination table 70 as shown in FIG. When all the flags are turned off, it is determined that the local region is filled only with the background color without the character.

When the character color is determined in this way, a method for calculating a missing color signal is determined. Thus, the character color determination unit A1
00 indicates the switching result of the determination result of the character color of the local area.
Transfer to The switching unit A102 determines the single-plate color signals in the R signal block buffer 58, the G signal block buffer 60, and the B signal block buffer 62 based on the control of the control unit 50, based on the character color determination. Part A10
When the determination result from 0 is the background color, the processing proceeds to the background restoring unit 74.
In other cases, it is transferred to the missing color restoring unit A104.

In the background restoration section 74, the background color estimation section 5
2 from the background color _{(R B, G B, B} B) receive, to fill the missing color signals based on this. On the other hand, the missing color restoring unit A104 receives the method of calculating the missing color signal from the character color determining unit A100 and the background color from the background color estimating unit 52, and calculates the missing color signal based on these. .

The background restoration section 74 or the missing color restoration section A
The R, G, and B color signals in which the missing color signals have been restored by the 104 and are in a three-plate state are output to the printer output unit 44.

Here, the background color estimating operation in the background color estimating section 52 will be described with reference to the flowchart shown in FIG.
Further explanation will be given. Note that in this figure, variables R _B , G _B , B
_B represents the background color, and Count represents the total number of valid rectangular areas (local areas).

That is, first, all signals from the output unit 38 are read (step S10), and each of the variables R _B , G _B , B _B ,
Count is initialized to “0” (Step S1)
2). Then, the image is scanned (step S14), and rectangular areas of a predetermined size, for example, 8 × 8 size, are sequentially extracted without overlap (step S16).

Next, R, G, B in the extracted rectangular area
Mean value R _av of each color signal, G _av, the B _av is calculated (step S18), and also, the average luminance Y of the rectangular region is calculated by _{Y = 0.3R av + 0.6G av +} 0.1B av (Step S20).

Then, whether the calculated average luminance Y is within a predetermined range is determined by determining whether the lower limit threshold value Th1 or the upper limit threshold value T
The judgment is made at h2 (step S22). These thresholds are set, for example, when the signal to be handled is an 8-bit level (0 to 255).
Then, Th1 = 10, Th2 = 230, etc. are used.

If the calculated average luminance Y is not within the threshold range, it is determined whether scanning of the entire screen has been completed (step S24). If not, the flow returns to step S14 to return to the next rectangular area. The above processing is repeated.

On the other hand, when the calculated average luminance Y is within the threshold range, the variables R _B , G _B , B
Mean value R _av of each color signal, G _av, the B _av integrated to _B (step S26), also the variable Count is incremented to "1" (step S28). Thereafter, the process proceeds to step S24, where it is determined whether the scanning of the entire screen is completed.

After the scanning of the entire screen is completed, the number of valid rectangular areas is then set to a predetermined number Th3, for example, Th3 =, by a variable Count indicating the number of valid rectangular areas.
10 is determined (step S3).
0). As a result, if exceeds the predetermined number Th3 is variable R _B, when G _B, to calculate the background color by dividing B _B variable Count (step S32), also not surpassed, If the reference background color, for example, the signal to be handled is at an 8-bit level, “255” corresponding to white is assigned to the variables R _B , G _B , and B _B (step S34). Then, the step S32 or S34 set background color R _B in, G _B, a B _B, and outputs as the background color estimated (step S36).

As described above, in order to recover a missing color signal from a color signal existing in a character image composed of a plurality of colors, blurring of an edge portion and generation of a false color caused by the conventional linear interpolation processing are avoided. It is possible to generate a character image with good readability. In addition, since this processing can be executed by a simple operation by referring to the table, low-cost and high-speed processing can be performed. Further, the present embodiment
Since there is no need to improve the electronic still camera itself, it is highly versatile.

In this embodiment, the arrangement shown in FIG. 2A is used for the color filters of the CCD. However, the present invention is not limited to this, and processing is performed for an arbitrary arrangement. It is also possible. Also, the character colors to be used need not be limited to the seven colors shown in FIG. 4, and any color can be registered. Furthermore, although processing is performed so that local regions do not overlap, if processing time does not matter, processing may be performed so as to overlap at one pixel intervals, and the results may be averaged to further improve performance. It is possible.

[Second Embodiment] Next, a second embodiment of the present invention will be described.
An embodiment will be described. The configuration of the electronic still camera system according to the second embodiment is the same as that of the first embodiment except for the filter arrangement used for the CCD 18 and the configuration of the character image restoration unit 42.

That is, in the second embodiment, the character image restoring section 42 calculates a background color for each color signal input via the processing switching section 40, and then calculates the background color and the color signal. By calculating the difference, the color signal is separated into the background and the others, and a process for restoring the image is performed based on the color signal, and is output to the printer output unit 44.

FIG. 6 is a diagram showing the configuration of the character image restoring section 42 in the present embodiment, and the same reference numerals are given to the same components as those in the first embodiment. In addition,
In the drawing, the symbol hh indicates a signal (for each color) indicating a difference from the background color in the local area, and mm indicates an estimated signal of a non-background color (non-background color) in the local area. I have.

That is, the processing switching section 40 outputs all signals from the output section 38 to the background color estimation section 52 under the control of the control section 50.
Transfer to In the background color estimating unit 52, the background color (R _B , G
_B , B _B ). The method of estimating the background color uses, for example, the first embodiment (FIG. 5). Then, the estimated background color is transferred to the R signal background color component difference calculator 78, the G signal background color component difference calculator 80, and the B signal background color component difference calculator 82.

After the background color is calculated in this way, the control unit 5
Based on the control of 0, the process switching unit 40 transfers the signal to the local area extracting unit 56 again. The local area extracting unit 56 sequentially extracts rectangular areas of a predetermined size, for example, an 8 × 8 size, without overlap based on the control of the control unit 50, and blocks the R signal block buffer 58, the G signal block buffer 60, and the B signal Transfer to the block buffer 62. At this time, the signal to be transferred is converted into a single-plate state and transferred based on the filter arrangement of the CCD shown in FIG. The color signals in the signal block buffers 58, 60, 62 are
The signals are transferred to an R signal background color component difference calculation unit 78, a G signal background color component difference calculation unit 80, and a B signal background color component difference calculation unit 82.

The background color component difference calculation sections 78, 80,
In 82, the background color _{(R B, G B, B} B) and the difference between the respective color signals in the local region is calculated, and the results are switching unit B1
06.

The switching unit B106 determines whether or not a portion other than the background exists in the local area based on the difference between the background color and the color signal transferred from the background color component difference calculation units 78, 80, and 82 for each signal. Find out. If the difference between all color signals in the local region and the background color is smaller than a predetermined value, for example, “16”,
The background color is assumed to be nonexistent, and otherwise the non-background color (R
_N , G _N , B _N ). Further, the switching unit B106
Under the control of the control unit 50, the single-plate color signals in the color signal block buffers 58, 60, and 62 are sent to the background restoration unit 74 when the color determined as described above is the background color. In this case, it is transferred to the missing color restoring unit B108.

[0054] Background restorer 74, receiving the background color from the background color estimation section _{52 (R B, G B,} B B) and to fill the missing color signals based on this.

On the other hand, in the missing color restoring unit B108, the non-background colors (R _N , G _N ,
B _N ) and the background color (R _B , G
_B , B _B ), and the missing color signal is calculated based on the received color signals. The calculation of the missing color signal includes, for example, judging from the color signal in a single-plate state whether or not it is the background, and restoring the missing portion to a color signal corresponding to the judgment result.

Specifically, when recovering the missing G signal G ′ and B signal B ′ from the single-plate R signal R ′,
'As a criterion to determine whether the background, the R signal R' R signal R and the absolute difference absolute difference R _Bs respective following formula R _Ns and R signal R 'as the background color R _B of the non-background colors R _N Is calculated as follows.

R _Ns = | R′−R _N | R _Bs = | R′−R _B | (9) Then, by comparing the magnitudes of the absolute difference R _Ns and the absolute difference R _Bs , the following equation is obtained. Thus, the missing G signal G 'and B signal B' are recovered.

[0058] if this way _{(R Ns ≦ R Bs) then} G '= G N B' = B N else G '= G B B' = B B ... (10), missing color signal is restored becomes three-plate state The respective color signals are output to the printer output unit 44.

FIG. 2B is a diagram showing C in the present embodiment.
An example of a specific configuration of the filter arrangement of the CD 18 is shown. Here, there is a basic arrangement of a 4 × 4 pixel size, and this basic arrangement is repeatedly repeated to fill all the pixels on the CCD. In the present embodiment, the appearance frequencies of the three RGB signals are set to be 3: 10: 3.

FIG. 7 shows a background color component difference calculator 7 for each color signal.
8, 80, 82, the non-background color (R _N ,
G _N , B _N ). This flowchart is started after an image of a rectangular area is read in a single-plate state into the block buffers 58, 60, and 62 of each color signal. In FIG. 7, for the sake of simplicity, processing corresponding to the portion surrounded by the dotted line (step S4)
4 to S54), only the case of the R signal is described. However, when the read color signal is the G signal or the B signal,
A similar process is performed by changing the portion described by the R signal to a G signal or a B signal. Therefore, in the following description, in steps S44 to S54, each color signal is handled collectively.

That is, first, each variable is initialized to "0" (step S40). Here, what is initialized, variable representing the difference between the background color and the color signals _{(R S, G S, B} S) , the variable _{(R S, G S, B} S) variable representing the maximum value of ( R
_{Smax, G Smax, B Smax)} , the non-background color (R _N, G _N,
B _N ) and a variable (R_R_) representing a counter that counts the number of portions where the above variables (R _S , G _S , B _S ) are equal to or greater than a certain value.
C, G_C, B_C).

Thereafter, the image in the rectangular area (local area) is scanned (step S42), and the target color signals (R ', G', B ') in the rectangular area are read (step S4).
4). Then, the read color signals (R ′, G ′,
Background color corresponding to the _{B ') (R B, G} B, calculates a difference between B _B), variable (R _S, G _S, and stores the B _S) (step S46). Here, the calculated variables (R _S ,
It is determined whether or not the absolute value of G _S , B _S ) is greater than the absolute value of the maximum difference (R _Smax , G _Smax , B _Smax ) calculated from the difference up to the previous time (step S48). Variables (R _S ,
G _S , B _S ) are the maximum values (R _Smax , G _Smax , B
If it is smaller than the absolute value of ( _Smax ), a step S52 to be described later is performed.
To the maximum difference (R _Smax ,
G _Smax, B _Smax) in the variable (R _S, G _S, after the maximum value updating by substituting the B _S) (step S50), the process proceeds to the next step S52.

In step S52, it is determined whether or not the absolute value of the calculated variable (R _S , G _S , B _S ) is a constant value Th31, for example, Th31 = 16 or more. Then, the absolute value of the variable (R _S , G _S , B _S ) is equal to this constant value T.
proceeds to step S56 to be described later is smaller than h31, if a constant value Th31 above, variable (R _S, G _S,
B _S ) is determined to be a non-background color, and the read color signals (R ′, G ′, B ′) are converted to non-background colors (R _N , G _N , B _N ).
, And at the same time, the counter variables (R_C, G
_C, B_C) is incremented by “1” (step S5).
4), proceed to the next step S56.

In step S56, it is determined whether the scanning of the rectangular area has been completed. If the scanning has not been completed, the process returns to step S42.

When the scanning of the rectangular area is completed, the counter variables (R_C, G_C, B_
C) to determine whether all are “0” (Step S)
58). These variables (R_C, G_C, B_C) if it is all "0", the rectangular region is determined that there is no portion other than the background, the background color _{(R B, G B, B} B) a non-background color _{(R N, G N, B} N) is substituted for (step S6
0). Then, the non-background color _{(R N, G N, B} N) for outputting a (step S62).

On the other hand, if it is determined in step S58 that the variables (R_C, G_C, B_C) are not all "0", then the counter variables (R_C, G_C, B_C) for each color signal are determined. Is determined to be "0" (step S64). When the variables (R_C, G_C, B_C) are “0” for each color signal, the maximum value of the difference from the background color (R _Smax , G _Smax , B _Smax )
Is determined to represent a non-background color, and the maximum value (R
_Smax, G _Smax, B _Smax) to the background color _{(R B, G B, B} B)
Are added to calculate the non-background color (R _N , G _N , B _N ) (step S66). If the variables (R_C, G_C, B_C) are not “0” for each color signal,
Variables (R _N ,
G _N , B _N ) are converted into counter variables (R_C, G_C, B).
_C), the non-background colors (R _N , G _N ,
B _N ) is calculated (step S68). Then, the process proceeds to step S62, the non-background color calculated in step S66 or _{S68 (R N, G N,} B N) to output a.

As described above, with respect to a character image composed of a plurality of colors, a difference between the estimated background color and the video signal is obtained to recover a color signal that is missing from a portion other than the background (including characters) at once. Therefore, a character image composed of an arbitrary type of color can be restored with high accuracy. In addition, since this embodiment can be realized with a simple configuration, high speed and low cost can be realized. Further, in the present embodiment, since it is not necessary to improve the electronic still camera itself, it becomes highly versatile.

In the present embodiment, the arrangement shown in FIG. 2B is used for the color filters of the CCD. However, the arrangement is not limited to this. For example, the arrangement shown in FIG. It is also possible to perform the processing for the arrangement of. Also,
By using the difference between the background color and the color signal in the local area and the background color to recover the missing color, it is possible to perform processing using a linear relationship between the color signals as in the determination table shown in FIG. is there. Furthermore, although processing is performed so that local regions do not overlap, if processing time does not matter, processing is performed so that they overlap at one pixel intervals, and the results are added and averaged to further improve performance. Is also possible.

[Third Embodiment] Next, a third embodiment of the present invention will be described.
An embodiment will be described. The electronic still camera according to the present embodiment is an example in which the functions configured in the adapter 12 in the first and second embodiments are configured in the electronic still camera.

FIG. 8 is a diagram showing the configuration of an electronic still camera according to the third embodiment. Components similar to those of the first embodiment are denoted by the same reference numerals.

That is, the input image transmitted through the lens system 14 and the low-pass filter 16 is guided to the CCD 18 and
It is supplied to the D conversion circuit 20. A / D conversion circuit 20
Is transferred to the character image restoring unit 42 or the general-purpose image restoring unit 84 via the process switching unit 40. Output signals of the character image restoration unit 42 and the general-purpose image restoration unit 84 are transferred to the output unit 38. Further, a signal of the clock generator 46 is supplied to the CCD 18 as a drive signal via a CCD drive circuit 48. The control unit 50 such as a microcomputer is connected to the processing switching unit 40, the character image restoration unit 42, the output unit 38, the clock generator 46, and the general-purpose image restoration unit 84. Here, a signal flow in such a configuration will be described. That is, a single-plate CCD through the lens system 14 and the low-pass filter 16
18 are arranged. The CCD 18 is, for example, a CCD having an RGB primary color type filter arrangement shown in FIGS. 2A and 2B. The video signal from the CCD 18 is digitized by the A / D conversion circuit 20. Further, a CCD drive circuit 48 that generates a drive signal based on a clock generator 46 is connected to the CCD 18. The single-plate video signal from the A / D conversion circuit 20 is transferred to the character image restoration unit 42 or the general-purpose image restoration unit 84 via the process switching unit 40 under the control of the control unit 50. This change is transmitted from the outside to the control unit 50 by a manual change switch (not shown) such as a character mode / general image mode. Character image restoration unit 42
In, a process of restoring a color character composed of a plurality of colors into a more readable image is performed and output to the output unit. On the other hand, in the general-purpose image restoration unit 84, after the missing color signal is interpolated by a known interpolation process (for example, linear interpolation),
Output to the output unit 38.

FIG. 9 is a block diagram of the character image restoring unit 42 according to the present embodiment. The same reference numerals as in the first embodiment denote the same parts. Note that, in the figure, a symbol ii represents a video signal in a single-plate state,
jj is a signal indicating the distribution of the histogram (for each color), k
k indicates a signal (for each color) indicating a color signal to be the peak of the histogram, and 11 indicates a signal indicating a character color in the local area.

The processing switching section 40 transfers the single-plate video signal from the A / D conversion circuit 20 to the background color estimation section C110 under the control of the control section 50. Background color estimation unit C110
Then, for example, the filter arrangement shown in FIG.
Based on the fact that it is established from repetition of × 2 pixel size,
By calculating RGB data for each 2 × 2 size, a three-plate image signal is generated. The video signal in the three-plate state has a size of 1/2 × 1/2 of the video signal in the single-plate state. Thereafter, similarly to the above-described first embodiment, the average value of each of the R, G, and B color signals in an area within a predetermined luminance range is calculated to obtain the background color (R _B , G _B , B). _B ) is estimated, and the estimated background color is transferred to the character color determination unit C112.

After the background color is transferred, the process switching unit 40 again controls the local region extraction unit C114 under the control of the control unit 50.
To transfer the signal. The local area extraction unit C114 separates the RGB signals into RGB signals based on the filter arrangement shown in FIG. 2A for a rectangular area of a predetermined size, for example, 8 × 8 size, based on the control of the control unit 50. Block buffer 58, G signal block buffer 60, B
The signal is transferred to the signal block buffer 62. The color signals in the signal block buffers 58, 60, and 62 are transferred to the R signal histogram generation unit 86, the G signal histogram generation unit 88, and the B signal histogram generation unit 90, and the distribution for each color signal is calculated. Is done. The distribution of each signal is represented by an R signal character color component detector 92 and a G signal character color component detector 9.
The color signal which is transferred to the 4, B signal character color component detection unit 96 and which is the peak of the distribution is detected, and is detected by the character color determination unit C.
Transferred to 112.

[0075] In the character color determining unit C112, the color signal and the background color as the peak of the distribution of each color signal _{(R B, G B, B} B)
And the character color (R _T ,
G _T , B _T ). And when the text color is determined,
The character color determination unit C112 determines the character color (R _T , G
_T, B _T) of the switching unit C116 transfers.

The switching section C116 outputs the single-plate color signals in the signal block buffers 58, 60 and 62 based on the control of the control section 50, and outputs the background signal when the result from the character color determination section C112 is the background color. The data is transferred to the restoration unit 74, and in other cases, to the missing color restoration unit C118. Background restorer 74, the background color from the background color estimation unit _{C110 (R B, G B,} B B) receive, to fill the missing color signals based on this. On the other hand, in the missing color restoring unit C118, the character color determining unit C112
Character color in the local area from the _{(R T, G T, B} T) the background color from the background color estimation unit _{C110 (R B, G B,} B B) receive respectively, the color signals missing on the basis of these calculate.

The method of calculating missing color signals utilizes the relationship between the color signals. When a character and a background are present in a local area, distributions are obtained for two color signals (S ₁ , S ₂ ), and the relationship is, as shown in FIG. Converge. From this state, four combinations of the character color and the background color are conceivable, but since the background color is calculated in advance, the combination of the background colors can be uniquely determined (in FIG. 10A,
(S _1B , S _2B ) is used as the background color.) Therefore, the combination of the character colors is also determined (in FIG. 10A, (S _1T , S
_2T ) is the text color). From this relationship, if it is determined whether the color signal in the single-plate state belongs to the background or the character, the missing color signal can be recovered as a color signal corresponding to the result.

The R, G, and B color signals in which the missing color signals have been restored and are in the three-plate state are output to the output unit 38.

[0079] Figure 11 shows a flowchart of processing from the histogram generation unit 86, 88 and 90 of the respective color signals to be output character color is the character color determination unit _{C112 (R T, G T,} B T) a. The operation of this flowchart is started after an image of a rectangular area (local area) is read in a single-plate state into the block buffers 58, 60, and 62 of each color signal.

That is, first, each variable is initialized to "0" (step S70). Here, sequences representing the frequency of the histogram _{(R H [], G H} [], B H []), sequence representing pixel values a peak value _{(R L [], G L} [], B L [] ) And variables (R _{L —} C, G _{L —}
C, B _L _C) is used. Array (R _H [], G _H [], B
The size of _H []) depends on the setting of the interval according to the signal level. For example, when the signal is at the 8-bit level (0 to 255), if the interval is set to “16”, the size of each array is “256”.
/ 16 = 16 ".

Next, histograms (R _H [], G _H [], B _H []) corresponding to the respective color signals in the rectangular area are created (step S72), and the frequency is determined from the histogram of the respective color signals. In the case of a rectangular area having a size equal to or more than the value, for example, in the case of a rectangular area of 8 × 8 size, a portion where the R signal and the B signal are “5” or more and the G signal is “10” or more are detected as peak values. Then, the pixel value corresponding to the peak value (for example, the median of the range) variables _{(R L [], G L} [], B L []) , a variable number of peak values (R _L _C, G _L _C, B _L _C) to store respectively (step S74). As a specific example, FIG. 12 shows the state of the pixel value corresponding to the peak value of the histogram of the R signal.

[0082] Next, where the number of peak values of each color signal (= variable _{(R L _C, G L _C} , B L _C)) is "2"
It is checked whether it is below (step S76). If this number is equal to or less than "2", color background is estimated from the background color estimation unit _{C110 (R B, G B,} B B) and a variable indicating the peak value _{(R L [], G L} [], B _L []) by referring to the text color in the rectangular area _{(R T, G T, B} T) to determine (step S78). Here, a variable having a large difference from the background color is defined as a character color.

For example, when the number of peak values is “2”, the luminance signal of the background color and the variables (R _L [], G _L [],
The brightness signal of B _L []) is calculated and compared, and a variable having a large absolute difference is defined as a character color. If the number of peak values is "1", the background color is set to the text color if the absolute difference between the variable and the luminance signal of the background color is within a predetermined value, for example, "15". And When the number of peak values is “0”, the luminance signal is calculated with respect to the average value or the median of the histograms (R _H [], G _H [], B _H []), and the number of peak values is “1”. The same processing as in the case of is performed. Such a method to determine the character color in the rectangular area _{(R T, G T, B} T) a.

Then, the character colors (R _T ,
G _T , B _T ) is output (step S80), and this operation ends.

On the other hand, if it is determined in step S76 that the number of peak values of each color signal is not equal to or less than "2", the recovery processing in the target rectangular area is not performed, and the target is set to the next rectangular area. Control is performed to move to the area (step S82), and this operation ends.

FIG. 13 is a flowchart of the processing operation of the background color estimating unit C110. Where the variables R _B , G _B ,
B _B is the background color, Count is the total number of effective three-chip pixel, C_
bw indicates the total number of three white pixels.

That is, first, a single-plate video signal transferred from the A / D conversion circuit 20 is read (step S9).
0), the variable _{R B, G B, B B} , Count, and is initialized to "0" C_bw (step S92).

Thereafter, the image is scanned (step S9).
4), the array size of the filters of the single CCD, for example, FIG.
The 2 × 2 rectangular areas shown in FIG. 7A are sequentially extracted without overlap (step S96). The color signals read at this time are R, G1, G2, and B, respectively, and are hereinafter collectively referred to as veneer data.

Next, the color components R ', G', B 'of one pixel in the three-plate state are calculated from the information of the single-plate data (step S98). This is hereinafter collectively referred to as three-plate data.
The filter array size of the single CCD is, for example, as shown in FIG.
In the case of the 2 × 2 size shown in the above, three-plate data is calculated from the following equation.

R ′ = RG ′ = (G1 + G2) / 2 (11) B ′ = B Next, using the three-plate data, a luminance signal Y is calculated as follows: Y = 0.3R ′ + 0.6G ′ + 0. 1B '(step S100).

Here, it is determined whether or not the luminance signal Y is within a predetermined range based on the lower threshold value Th1 and the upper threshold value Th2 (step S102). As this threshold,
For example, if the signal to be handled is an 8-bit level (0 to 255), Th1 = 10, Th2 = 230, etc. are used.

[0092] If the luminance signal Y is within this range, variable R _B, G _B, color signals R three-chip data B _B ', G',
B ′ is integrated (step S104), and after incrementing the variable Count by “1” (step S106), it is determined whether scanning of the entire screen has been completed (step S108). yet,
If the scanning of the entire screen has not been completed, the process returns to step S94.

On the other hand, when it is determined in step S102 that the luminance signal Y is not within the predetermined range, it is further determined whether the luminance signal Y is equal to or more than the upper limit threshold Th2 (step S110). If the luminance signal Y is not equal to or greater than the upper threshold value Th2, the process proceeds to step S10.
8, if the value is equal to or more than the upper threshold value Th2, the variable C_bw
Is incremented by "1" (step S112), and the process proceeds to step S108.

If it is determined in step S108 that the scanning of the entire screen has been completed, the number of valid three-plate data Count is set to a predetermined number Th3, for example, Th3 = 10.
Is determined (step S11).
4). If exceeds, the variable R _B, by dividing G _B, a B _B a variable Count, calculates the background color (step S116), the calculated background color R _B, G _B, outputs B _B (Step S118).

On the other hand, if it is determined in step S114 that the number C_bw does not exceed the predetermined number Th3, it is further determined whether the number C_bw of white three-plate data exceeds the predetermined number Th3 ( Step S1
20) If it exceeds, for example, if the signal to be handled is an 8-bit level, the value “255” corresponding to white is set to the variable R _B ,
G _B, by substituting the B _B (step S122), or if not surpassed, the variable R _B value "0" corresponding to black,
G _B, by substituting the B _B (step S124), the background color R _B above step S118, the outputs of G _{B, B B.}

As described above, with respect to a character image composed of a plurality of colors, a histogram of each color component is created from a single-plate image signal for each local region, and the character color is determined using the image signal having the peak value of this distribution. Is estimated, a character image composed of any kind of color can be restored. In addition, since the missing color signal of the portion corresponding to each of the character and the background is recovered based on the video signal that is the peak value of the histogram, the character image can be restored with high accuracy while minimizing the influence of noise.

In the present embodiment, the arrangement shown in FIG. 2A is used for the color filters of the CCD. However, the arrangement is not limited to this. For example, the arrangement shown in FIG. It is also possible to perform the processing for the arrangement of. Also,
Although processing is performed so that local regions do not overlap, if processing time does not matter, processing can be performed so as to overlap at one pixel intervals, and the results can be averaged to further improve performance. is there. Further, in the present embodiment,
As the restoration of the character image, a process of estimating the background color and the character color from the video signal in the single-plate state and restoring the missing color is performed. However, as in the second embodiment, after estimating the background color, A process of restoring missing colors by referring to a single-plate video signal is also possible.

Although the present invention has been described based on the above embodiments, the present invention is not limited to the above embodiments, and various modifications and applications are possible within the scope of the present invention. Of course. For example, in the above-described embodiment, a single-chip CCD has been described as an example of the solid-state imaging device.
The present invention is similarly applicable to a two-chip CCD.

The summary of the present invention is as follows.

(1) A character image restoring apparatus for restoring a character image composed of a plurality of colors from a video signal of a single-panel or two-panel solid-state imaging device, wherein a background color is estimated from the video signal of the solid-state imaging device. Background color estimating means, extracting means for extracting a local area of a predetermined size from the video signal of the solid-state imaging device, character color estimating means for estimating the character color in the local area extracted from the extracting means, Separating means for separating the video signal in the local area into a background and a character based on the background color estimated by the color estimating means and the character color estimated by the character color estimating means; And a missing color restoring means for restoring a color signal missing due to the background color, and restoring a color signal missing due to the character color when converted to a character by the separating means. Character image restoration device.

An embodiment relating to this configuration corresponds to the first embodiment. As a method for restoring a character image, a method of estimating a background color from a given video signal, and The signal is divided into local regions, the character color is estimated for each local region, the background and characters are separated for each local region, and the background color is estimated from the estimated background color, and the character is Includes a method for restoring missing color signals from the estimated character color. The method of estimating the background color includes a method of generating a luminance signal based on a given video signal and estimating the background color from a video signal in an area where the luminance signal is within a predetermined range. The method of estimating the character color includes a method of estimating the character color from a character color table in which the distribution range of the maximum and minimum values of the local region and the background color and the character color are registered from the given video signal.

For example, in the character image restoring section 42 shown in FIG. 3, when the given video signal is a character image, the background color estimating section 52 averages the video signal within a predetermined luminance range from the entire video signal. By estimating the background color, the local area extracting unit 56 divides the video signal into local areas and converts it into a single-plate state, and outputs a maximum / minimum calculating unit 64 for an R signal and a maximum / minimum calculating unit 66 for a G signal. , B signal maximum / minimum calculation unit 68 calculates the maximum / minimum range for each local area, and character color determination unit A100 and determination table 70 estimate the character color from the maximum / minimum range and the background color. The character image is restored by the background restoration unit 74 and the missing color restoration unit A104.

That is, the given video signal is restored again only when it is a character image. In the restoration processing, first, a background color is obtained from a video signal within a predetermined luminance range. Next, the video signal is returned to the single-plate state and divided into local regions of a predetermined size, and a character color is obtained for each local region. The text color is
This is determined by comparing the maximum and minimum signal distribution forms, the background color, and the character color table registered in advance. When the background color and the character color are determined, the local portion is separated into the background portion and the character portion, and the missing color signal is restored.

As described above, the character color is estimated for each local area based on the character color registered in advance, and thereafter, the missing color signal is restored because there are only two types of colors, the background color and the character color. In addition, the false color of the edge portion generated when the image is restored by the conventional linear interpolation can be reduced, and a character image with good readability can be generated. In addition, since the estimation of the character color is performed by referring to a table registered in advance, the process is performed at high speed, and there is no need to improve the electronic still camera itself, so that the versatility is high.

The embodiment relating to the configuration (1) also corresponds to the third embodiment. As a method of restoring a character image, a background color is estimated from a video signal in a single-plate state, a video signal in a single-plate state is divided into local regions, and a character color is estimated for each local region. The method includes a method of separating the background and the character from each other, and restoring a color signal missing from the estimated background color in the case of the background color and the estimated character color in the case of the character. As a method of estimating the background color, a video signal having no reduced-size missing color signal is generated from a video signal in a single-plate state, a luminance signal is generated based on the reduced-size video signal, and the luminance signal is within a predetermined range. The method includes a method of estimating a background color from a video signal in an area within the image. In addition, the method of estimating the character color includes a method of generating a histogram of a local region from a video signal in a single-plate state and estimating a character color from a video signal having a peak value of the histogram and a background color.

For example, in the character image restoring section 42 shown in FIG. 9, when the given video signal is a character image, the background color estimating section C110 generates a reduced-size video signal and then outputs a predetermined signal from the entire video signal. The background color is estimated by averaging the video signals within the luminance range, and the local region extraction unit C114
Divides the video signal into local regions, and generates a histogram for each local region in an R signal histogram generation unit 86, a G signal histogram generation unit 88, and a B signal histogram generation unit 90. Color component detector 92, G
Signal character color component detector 94, B signal character color component detector 9
6, a video signal having a histogram peak value is detected, a character color determination unit C112 estimates a character color from the peak value video signal and a background color, and a background restoration unit 74 and a missing color restoration unit C118 detect a character color. Restore the image.

That is, a single-plate video signal (corresponding to a character image) is converted into a reduced-size video signal with no missing signals for background color estimation, and then a reduced-size video signal within a predetermined luminance range. Find the background color from the signal. Next, the video signal in the single-plate state is divided into local regions of a predetermined size, and a character color is obtained for each local region. The character color is determined by generating a histogram for each color signal and comparing the video signal having its peak value with the background color. When the background color and the character color are determined, the color signal which is separated into the background portion and the character portion for each local region and is missing is restored.

As described above, since the histogram of the video signal is generated for each local area and the character color is estimated using the video signal having the peak value, a character image composed of any kind of color can be restored. In addition, since the missing color signal in the character color and the background color estimated from the histogram of the video signal is restored, the character image can be restored with high accuracy while minimizing the influence of noise.

(2) A character image restoring device for restoring a character image composed of a plurality of colors from a video signal of a single-chip or two-chip solid-state imaging device, wherein a background color is estimated from the video signal of the solid-state imaging device. Background color estimating means, extracting means for extracting a local area of a predetermined size from the video signal of the solid-state imaging device, and extracting the video signal in the local area as a background or the like based on the background color estimated from the background color estimating means. Separation means for separating the color signal which is missing due to the background color when the background is set by the separation means, and averaging the other areas in the local area when the others are selected A character image restoring device, comprising: a missing color restoring means for restoring a missing color signal by the extracted video signal.

An embodiment relating to this configuration corresponds to the second embodiment. As a method for restoring a character image, a background color is estimated from a given video signal, and a given video signal is restored. The signal is divided into local regions, and the background and others are separated for each local region, and the background color is estimated from the estimated background color, otherwise the average of the other regions in the local region And a method for restoring a missing color signal from the converted video signal.

For example, in the character image restoring section 42 shown in FIG. 6, when a video signal corresponding to a character image is given, the background color estimating section 52 converts a video signal within a predetermined luminance range from the entire video signal. By averaging, the background color is estimated, and the local area extracting unit 56 divides the video signal into local areas and converts it into a single-plate state.
The G signal background color component difference calculation unit 80 and the B signal background color component difference calculation unit 82 calculate the difference between the video signal and the background color for each local area, and the switching unit B106 sets the local area to the background color. And an averaged video signal of the other area is calculated, and the background restoration unit 74 and the missing color restoration unit B1 are calculated.
At 08, the character image is restored.

That is, in the restoration process of a character image, a background color is first obtained from a video signal within a predetermined luminance range. next,
The video signal is returned to the single-plate state and divided into local regions of a predetermined size, the difference between the video signal and the background color is determined for each local region, and the local and other regions are separated for each local region. The missing color signal is restored by the background color of the background, and by the video signal averaged for the other area in the local area.

As described above, by taking the difference between the estimated background color and the video signal, a color signal in which portions other than the background (including characters) are collectively lost is restored, so that the image signal is composed of an arbitrary type of color. Character images can be restored with high accuracy. Further, the character image restoring device having such a configuration has a simple configuration, and can achieve high speed and low cost. Further, since the character image restoring device can be configured independently of the electronic still camera main body, there is no need to improve the electronic still camera itself, so that the versatility is high.

(3) In an electronic still camera system including an electronic still camera and a post-processing device using a single-chip or two-chip solid-state imaging device, a color signal missing from a video signal of the solid-state imaging device is restored by interpolation. A general-purpose image restoration unit, a character image restoration unit that restores a character image composed of a plurality of colors from a video signal restored by the general-purpose image restoration unit by a restoration method suitable for a character image different from the interpolation, A transmission means for transmitting a video signal restored by the general-purpose image restoration means to the character image restoration means, wherein at least the character image restoration means is configured in the post-processing device. system.

An embodiment relating to this configuration corresponds to the first embodiment shown in FIGS. here,
The general-purpose image restoration means includes a color separation circuit 22 shown in FIG.
The process circuits 24, 26, and 28 and the matrix circuit 30 correspond to this. The character image restoring unit corresponds to the character image restoring unit 42 in FIG. As a method of restoring a character image, a background color is estimated from a video signal from a general-purpose image restoration unit, a video signal from the general-purpose image restoration unit is divided into local regions, and a character color is estimated for each local region. And a method of separating a background and a character for each local region, and restoring a missing color signal from the estimated character color from the estimated background color in the case of the background color. The method of estimating a background color includes a method of generating a luminance signal based on a video signal from a general-purpose image restoration unit and estimating the background color from a video signal in an area where the luminance signal is within a predetermined range.
As a method of estimating the character color, a method of estimating the character color from a character color table in which the distribution range of the maximum and minimum values of the local area and the background color and the character color are registered from the video signal from the general-purpose image restoration unit is used. including. The configuration switching unit corresponds to the process switching unit 40 shown in FIG.

A preferred application example of the electronic still camera system having this configuration is that an image signal from the single CCD 18 shown in FIG. 1 is converted into an A / D conversion circuit 20, a color separation circuit 22, process circuits 24, 26, 28, and a matrix circuit. At 30, the missing color signal is restored and stored, and only when the video signal is a text image is transferred to the text image restoration unit 42 shown in FIG. Estimating the background color by averaging the video signal within a predetermined luminance range from the entire signal, dividing the video signal into local regions by the local region extracting unit 56 and converting it into a single-plate state, The maximum / minimum calculation unit 64, the maximum / minimum calculation unit 66 for G signal, and the maximum / minimum calculation unit 68 for B signal calculate the maximum / minimum range for each local area. The largest It estimates the character color from a small range and the background color, an electronic still camera system for restoring a character image by missing color restoration unit A104 and the background restoration unit 74.

That is, the image signal from the general-purpose image restoration means is restored again only in the case of a character image. In the restoration processing, first, a background color is obtained from a video signal within a predetermined luminance range. Next, the video signal is returned to the single-plate state and divided into local regions of a predetermined size, and a character color is obtained for each local region. The character color is determined by comparing the maximum and minimum distribution forms of the signal, the background color, and a character color table registered in advance. When the background color and the character color are determined, the color signal which is separated into the background portion and the character portion for each local region and is missing is restored.

The character image restoring means estimates the character color for each local area based on the character color registered in advance. afterwards,
In order to restore a missing color signal because only two types of colors, the background color and the character color, exist, false colors at edges are reduced when a conventional linear interpolation is used to generate a readable character image. be able to. In addition, since the estimation of the character color is performed by referring to a table registered in advance, the process is performed at high speed, and there is no need to improve the electronic still camera itself, so the versatility is high.

The embodiment relating to the above configuration corresponds to the second embodiment shown in FIG. 1 and FIGS. 2, 6 and 7. The configuration general-purpose image restoration unit, the configuration character image restoration unit, the configuration transmission unit, and the method of estimating the background color are as described above. As a method of restoring a character image, a background color is estimated from a video signal from a general-purpose image restoration unit, and a video signal from the general-purpose image restoration unit is divided into local regions. Perform the separation,
In the case of the background color, a method of restoring a missing color signal from the averaged video signal of the other region in the local region from the estimated background color otherwise.

A preferred application example of the electronic still camera system having this configuration is to convert an image signal from the single-chip CCD 18 shown in FIG. 1 into an A / D conversion circuit 20, a color separation circuit 22, process circuits 24, 26, 28, and a matrix circuit. At 30, the missing color signal is restored and stored, and only when the video signal is a character image is transferred to the character image restoring unit 42 shown in FIG. Estimating the background color by averaging the video signal within a predetermined luminance range from the entire signal, dividing the video signal into local regions by the local region extracting unit 56 and converting it into a single-plate state, Background color component difference calculator 78, G signal background color component difference calculator 8
The 0, B signal background color component difference calculation unit 82 calculates the difference between the video signal and the background color for each local area, and the switching unit B106
An electronic still camera that separates a local area into a background color and others, calculates an averaged video signal of the other areas, and restores a character image by the background restoration unit 74 and the missing color restoration unit B108 System.

That is, the video signal from the general-purpose image restoration means is restored again only in the case of a character image. In the restoration processing, first, a background color is obtained from a video signal within a predetermined luminance range. Next, the video signal is returned to the single-plate state, and divided into local regions of a predetermined size, a difference between the video signal and the background color is obtained for each local region, and the background and others are separated for each local region. . The missing color signal is restored by the background color of the background, and by the video signal averaged for the other area in the local area.

The character image restoring means restores a color signal in which portions other than the background (including characters) are collectively lost by taking the difference between the estimated background color and the video signal. Can be restored with high precision. Further, the character image restoring means has a simple configuration, and can achieve high speed and low cost. Furthermore, the character image restoring means can be configured independently of the electronic still camera body, and has high versatility since there is no need to improve the electronic still camera itself.

(4) In an electronic still camera using a single-chip or two-chip solid-state imaging device, general-purpose image restoration means for restoring, by interpolation, a color signal missing from a video signal of the solid-state imaging device; Character image restoring means for restoring a character image composed of a plurality of colors from the video signal by a restoration method suitable for a character image different from the interpolation, and switching between the general-purpose image restoring means and the character image restoring means An electronic still camera, comprising: switching means.

An embodiment relating to this configuration corresponds to the third embodiment shown in FIGS. The general-purpose image restoration unit corresponds to the general-purpose image restoration unit 84 shown in FIG. The character image restoring unit is a character image restoring unit 4 shown in FIG.
2 corresponds. As a method of restoring a character image, a background color is estimated from a video signal in a single-plate state, a video signal in a single-plate state is divided into local regions, and a character color is estimated for each local region. The method includes a method of separating the background and the character from each other, and restoring a color signal missing from the estimated background color in the case of the background color and the estimated character color in the case of the character. As a method of estimating the background color, a video signal having no reduced-size missing color signal is generated from a video signal in a single-plate state, a luminance signal is generated based on the reduced-size video signal, and the luminance signal is within a predetermined range. The method includes a method of estimating a background color from a video signal in an area within the image. In addition, the method of estimating the character color includes a method of generating a histogram of a local region from a video signal in a single-plate state and estimating the character color from a video signal having a peak value of the histogram and a background color. The processing switching means corresponds to the processing switching unit 40 shown in FIG.

In a preferred application example of the electronic still camera having this configuration, a video signal from the single CCD 18 shown in FIG.
Via the / D conversion circuit 20 to the process switching unit 40,
Only when the image signal is a character image in the processing switching unit 40, the image signal is transferred to the character image restoring unit 42 shown in FIG. The background color is estimated by averaging the video signals in the luminance range, the video signal is divided into local regions by the local region extraction unit C114, and the R signal histogram generation unit 86 and the G signal histogram generation unit 88 , B signal histogram generation unit 90 generates a histogram for each local area, and an R signal character color component detection unit 92, a G signal character color component detection unit 94, and a B signal character color component detection unit 96 generate a histogram. Is detected, and the character color determination unit C112 estimates the character color from the peak value video signal and the background color, and the background restoration unit 74 and the missing color restoration unit C1
An electronic still camera for restoring a character image at 18.

That is, the video signal in a single-plate state is restored into a general image and a character image separately. In the character image restoration processing, first, the video signal in a single-plate state is converted into a reduced-size video signal without a missing signal for background color estimation, and then the background color is converted from the reduced-size video signal within a predetermined luminance range. Ask. Next, the video signal in the single-plate state is divided into local regions of a predetermined size, and a character color is obtained for each local region. The character color is determined by generating a histogram for each color signal and comparing the video signal having its peak value with the background color. When the background color and the character color are determined, the color signal which is separated into the background portion and the character portion for each local region and is missing is restored.

The character image restoring means generates a histogram of the video signal for each local region and estimates the character color using the video signal having the peak value, so that a character image composed of any kind of color can be restored. In addition, since the missing color signal in the character color and the background color estimated from the histogram of the video signal is restored, the character image can be restored with high accuracy while minimizing the influence of noise.

(5) The character image restoring means includes: a background color estimating means for estimating a background color from a video signal of the solid-state imaging device or a video signal restored by the general-purpose image restoring device; Extracting means for extracting a local area of a predetermined size from a signal or a video signal restored by the general-purpose image restoring means; and extracting a video signal in the local area as a background or the like based on the background color estimated by the background color estimating means. Separation means for separating the color signal which is missing due to the background color when the background is set by the separation means, and averaging the other areas in the local area when the others are selected An electronic still camera system according to (3) or an electronic still camera according to (4), comprising: a missing color restoring means for restoring a missing color signal by the extracted video signal. Child still camera.

According to such a configuration, as described in the above (1), an inexpensive electronic still camera system and an inexpensive electronic still camera system that can restore a character image of an arbitrary type of color with high accuracy. can do.

(6) The character image restoring means comprises: a background color estimating means for estimating a background color from a video signal of the solid-state imaging device or a video signal restored by the general-purpose image restoring device; Extracting means for extracting a local area of a predetermined size from a signal or a video signal restored by the general-purpose image restoring means; character color estimating means for estimating a character color in the local area extracted from the extracting means; Separating means for separating the video signal in the local area into a background and a character based on the background color estimated by the color estimating means and the character color estimated by the character color estimating means; In the case, a color signal missing by the background color is restored, and a missing color restoring means for restoring a color signal missing by the character color when the character is formed by the separating means, The electronic still camera system according to (3) or the electronic still camera according to (4), comprising:

According to such a configuration, as described in (2) above, an electronic still camera system and an electronic still camera for restoring a character image composed of an arbitrary type of color with high accuracy while eliminating the influence of noise. Can be provided. Further, it is possible to provide an electronic still camera system and an electronic still camera that can quickly restore a character image composed of limited types of colors.

(7) The background color estimating means generates, from the video signal of the solid-state imaging device, a video signal without a color signal having a reduced size and no luminance signal, and generates a luminance signal from the video signal without a color signal having a reduced size. (1) generating a signal and estimating a background color from a video signal having no color signal having the reduced size and missing according to the range of the luminance signal (1).
Or the character image restoration device according to (2), or the electronic still camera system and the electronic still camera according to (5) or (6).

That is, according to such a configuration, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera which are capable of restoring a character image of any kind of color with high accuracy and being inexpensive. it can. In addition, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera that restore a character image composed of an arbitrary type of color with high accuracy while eliminating the influence of noise.
Further, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera that restore a character image composed of limited types of colors at high speed.

(8) The background color estimating means generates a luminance signal from the video signal in which the missing color signal of the general-purpose image restoring means has been restored, and the missing color signal is generated in accordance with the range of the luminance signal. The character image restoration device according to (1) or (2), or the electronic still camera system and the electronic still camera according to (5) or (6), wherein the background color is estimated from the restored video signal. .

That is, according to such a configuration, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera which can restore a character image of an arbitrary type of color with high accuracy and are inexpensive. it can. In addition, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera that restore a character image composed of an arbitrary type of color with high accuracy while eliminating the influence of noise.
Further, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera that restore a character image composed of limited types of colors at high speed.

(9) The character color estimating means generates a histogram from the video signal of the solid-state image pickup device or the video signal restored by the general-purpose image restoring means, and obtains a video signal having a peak value of the histogram and the background color. The character image restoration device according to (1) or the electronic still camera system and the electronic still camera according to (6), wherein the character color is estimated from the background color estimated by the estimation unit.

That is, according to such a configuration, there is provided a character image restoring apparatus, an electronic still camera system, and an electronic still camera that restore a character image of an arbitrary type of color with high accuracy while eliminating the influence of noise. be able to.

(10) The character color estimating means calculates the maximum / minimum value from the video signal of the solid-state imaging device or the video signal restored by the general-purpose image restoring means, and calculates the distribution range of the maximum / minimum value. The character image restoration device according to (1) or the electronic still image according to (6), wherein the character color is estimated based on the background color estimated by the background color estimation unit and a character color table registered in advance. Camera system and electronic still camera.

In other words, according to such a configuration, it is possible to provide a character image restoring device, an electronic still camera system, and an electronic still camera that restore a character image composed of limited types of colors at high speed.

[0140]

As described in detail above, according to the present invention,
In an electronic still camera using a single-plate or two-plate solid-state imaging device and an electronic still camera system including the same as a component, a character capable of restoring a character image composed of a plurality of colors to a high-definition image with excellent readability. An electronic still camera and an electronic still camera capable of restoring a character image composed of a plurality of types of colors into a highly readable image by providing an image restoring device and thus including such a character image restoring device A system can be provided.

[Brief description of the drawings]

FIG. 1 is a block configuration diagram of an electronic still camera system according to a first embodiment of the present invention.

FIG. 2A shows a single-chip CCD according to the first embodiment;
FIG. 7B is a filter arrangement diagram of a single-chip CCD according to the second embodiment.

FIG. 3 is a block diagram of a character image restoring unit according to the first embodiment.

FIG. 4 is a diagram showing the contents of a determination table.

FIG. 5 is a flowchart of a background color estimation operation according to the first embodiment.

FIG. 6 is a block diagram of a character image restoring unit according to a second embodiment of the present invention.

FIG. 7 is a flowchart of a non-background color determination operation according to the second embodiment.

FIG. 8 is a block diagram of an electronic still camera according to a third embodiment of the present invention.

FIG. 9 is a block diagram of a character image restoring unit according to a third embodiment.

FIG. 10A is a diagram for explaining a correlation between color signals in a local region, and FIG. 10B is an explanatory diagram of a filter arrangement of a conventional single-chip solid-state imaging device.

FIG. 11 is a flowchart of a character color estimation operation according to the third embodiment.

FIG. 12 is a diagram illustrating a state of a pixel value corresponding to a peak value of a histogram.

FIG. 13 is a flowchart of a background color estimation operation according to the third embodiment.

[Explanation of symbols]

Reference Signs List 10 electronic still camera 12 adapter 18 CCD 20 A / D conversion circuit 22 color separation circuit 24, 26, 28 process circuit 30 matrix circuit 32, 34, 36 buffer for R signal, G signal, B signal 38 output section 40 processing Switching unit 42 Character image restoration unit 44 Printer output unit 46 Clock generator 48 CCD drive circuit 50 Control unit 52 Background color estimation unit 56 Local area extraction unit 58, 60, 62 Block buffer for R signal, G signal, B signal 64 , 66, 68 R signal, G signal, B signal maximum / minimum calculation section 70 Judgment table 74 Background restoration section 78, 80, 82 R signal, G signal, B signal background color component difference calculation section 84 General-purpose image restoration unit 86,88,90 R signal, G signal, B signal histogram generation unit 92,94,96 R signal , G signal, B signal character color component detecting unit 100 character color determining unit A 102 switching unit A 104 missing color restoring unit A 106 switching unit B 108 missing color restoring unit B 110 background color estimating unit C 112 character color determining unit C 114 Local area extraction unit C 116 Switching unit C 118 Missing color restoration unit C

Claims (4)

[Claims] 1. A character image restoration apparatus for restoring a character image composed of a plurality of colors from a video signal of a single-chip or two-chip solid-state imaging device, wherein a background color is estimated from a video signal of the solid-state imaging device. Color estimating means, extracting means for extracting a local area of a predetermined size from the video signal of the solid-state imaging device, character color estimating means for estimating the character color in the local area extracted from the extracting means, and the background color Separating means for separating the video signal in the local area into background and characters based on the background color estimated by the estimating means and the character color estimated by the character color estimating means; And a missing color restoring means for restoring a color signal missing due to the background color, and restoring a color signal missing due to the character color when the character is formed by the separating means. Character image restoration device. 2. A character image restoration device for restoring a character image composed of a plurality of colors from a video signal of a single-chip or two-chip solid-state imaging device, wherein a background color is estimated from a video signal of the solid-state imaging device. Color estimating means, extracting means for extracting a local area of a predetermined size from the video signal of the solid-state imaging device, video signal in the local area based on the background color estimated from the background color estimating means, Separating means for separating, when the background is set by the separating means, a color signal which is missing due to the background color is restored, and when it is set as other, the other area within the local area is averaged. And a missing color restoring means for restoring a missing color signal by the video signal. 3. An electronic still camera system comprising an electronic still camera and a post-processing device using a single-chip or two-chip solid-state imaging device, wherein a color signal missing from a video signal of the solid-state imaging device is restored by interpolation. Image restoration means; and a character image restoration means for restoring a character image composed of a plurality of colors from a video signal restored by the general-purpose image restoration means by a restoration method suitable for a character image different from the interpolation; Transmission means for transmitting a video signal restored by the image restoration means to the character image restoration means, wherein at least the character image restoration means is configured in the post-processing device. . 4. An electronic still camera using a single-chip or two-chip solid-state imaging device, wherein: a general-purpose image restoration means for restoring color signals missing from a video signal of the solid-state imaging device by interpolation; A character image restoring means for restoring a character image composed of a plurality of colors from a video signal by a restoration method suitable for a character image different from the interpolation, and a process switching for switching between the general-purpose image restoring means and the character image restoring means Means, and an electronic still camera.