JP2532975B2 - Color video camera - Google Patents

Description

The present invention relates to a white balance adjusting device for a color video camera that controls white balance based on an imaged video signal obtained from an image sensor.

(B) Conventional Technology In a color video camera, it is necessary to control the white balance in order to correct the difference in the wavelength distribution of light depending on the light source.

This control is performed by adjusting the gain of each color signal so that the ratio of the three primary color signals of red (hereinafter R), blue (hereinafter B), and green (hereinafter G) is 1: 1: 1. Generally, for example, as shown in JP-A-62-35792 (H04N9 / 73), a method of adjusting the gain so that the integrated value of the color difference signals RY and BY of the screen becomes zero is used. ing.

FIG. 2 is a block diagram of a white balance adjusting circuit using this method.

The light that has passed through the lens (1) is the image sensor (CCD).
After photoelectric conversion in (2), R,
The signals are extracted as three primary color signals of G and B, the G signal is directly input, and the R and B signals are input to the color process and matrix circuit (6) through the R amplification circuit (4) and the B amplification circuit (5). , Luminance signal Y, and color difference signals RY and BY of red and blue are generated and sent to the video circuit. At the same time, the two color difference signals are respectively fed to the integrating circuit (1
7) In (18), the gain control circuits (13) and (14) are integrated for a sufficiently long time, and the gains of the amplifier circuits (4) and (5) of R and B are adjusted so that the result becomes zero. Adjust.

In this method, various color distributions of the screen shot by the video camera are devised such that the time constants of the integrating circuits (17) and (18) are lengthened and the color distributions are averaged. For example, it is premised that the color components that make up the color distribution cancel each other out and can be approximated to a substantially white screen state. However, when the color of the subject itself is biased, for example,
When a green grass or blue sky occupies a large area on the screen, the gain changes in the direction of canceling the biased color, and the white balance shifts to the complementary color side.

As a measure against this, by not inputting the color information signal of the part that is not included in the predetermined range on the color plane on the screen or the part where the brightness level exceeds a certain value to the integrating circuit (17) (18). , A method of eliminating the influence of a special part on the screen is often used.

(C) Problems to be Solved by the Invention In the above-described method, in order to properly adjust the white balance for various subjects, it is necessary to determine whether or not a predetermined condition is satisfied based on the color information signal including the brightness level. It is determined whether or not, and white balance adjustment is performed using the color information signal selected here.

However, in the case of judging whether or not the condition is satisfied, in the vicinity of the boundary of the condition, a small change in the subject has a great influence on the evaluation of the entire screen, and it is difficult to obtain a stable image. On the other hand, in order to prevent this instability, subdividing the conditions and setting intermediate processing is severely restricted in terms of system capacity and processing capacity.

(D) Means for Solving the Problems The present invention is based on a color information signal in an imaged video signal, a representative value determination means for determining a representative value of the color information signal of the entire imaging screen, and based on this representative value In a white balance adjusting device having a gain control means for controlling the gain of each color signal, fuzzy inference is used to determine the representative value in the representative value determining means. Color evaluation value detection means for obtaining the color information signal level for each of a plurality of areas set by dividing the screen as a color evaluation value of each color, and the color evaluation value of each area is weighted by a weighting amount for each area. A screen color evaluation value calculation means for calculating a screen color evaluation value representative of the entire screen from the weighted color evaluation value, and a gain control means for controlling the amplification gain of each color signal based on the screen color evaluation value. , Screen color evaluation value calculation means And an input variable values and brightness level of each region based on at least the color evaluation values of the respective regions, and performs fuzzy inference the weighting amount as a conclusion portion.

(E) Operation The present invention is configured as described above, and performs inference based on some preset rules using fuzzy inference without setting detailed conditions corresponding to various conditions. The representative value of the screen color information can be determined, and stable white balance adjustment can be easily performed.

(F) Embodiment An embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a circuit block diagram of an automatic white balance adjusting circuit according to this embodiment.

The light passing through the lens (1) is imaged on the CCD (2) and photoelectrically converted, and then R,
It is taken out as three primary color signals of G and B. The R and B signals of these three primary color signals are input to the camera process and matrix circuit (6) together with the G signal through the R and B amplifier circuits (4) and (5), respectively, and the luminance signal ( Y) and the color difference signals (RY) of red and blue (B-
Y) is created and supplied to the video circuit (7) and subjected to well-known processing. Further, the respective signals (Y) (RY) (BY) are simultaneously supplied to the selection circuit (21).

The selection circuit (21) receives the luminance signal (Y) and the color difference signal (R−) according to the selection signal (S1) from the timing circuit (25).
One of the three signals Y) and (BY) is sequentially selected for each field. (Y) → (RY) →
(B−Y) → (Y) → (R−Y) → ... Is output to the A / D converter (22) in the subsequent stage for each field. The selection signal (S1) is created based on the vertical sync signal obtained from the sync separation circuit (24) as described later. A / D converter (22)
Is one of the signals (Y) (R-Y) (B-Y) selected by the selection circuit (21) at a predetermined sampling period, converts the sampled signal into a digital value, and this value is integrated ( 2
Output to 3). By the way, the timing circuit (25) shows an image pickup screen of 8 × 8 shown in FIG. It is divided into 64 rectangular areas (A11), (A12), (A13) ... (A88), that is, (Aij) (i, j = 1 to 8), and these areas are divided for each area. A switching signal (S2) for time-divisionally extracting the output of the selection circuit (21) therein is output to the integrator (23).

The integrator (23) receives the switching signal (S2), adds the A / D converted value of the output of the selection circuit (21) over one field period for each region, that is, digitally integrates for every 64 regions, When the integration for one field is completed, the integrated value is held in the memory (26) as a brightness or color evaluation value. As a result, the digital integrated value of the luminance signal (Y) corresponding to 64 regions in a given field can be obtained as 64 luminance evaluation values (yij) (i, j: 1 to 8). Become. In the next field, since the color difference signal (RY) is selected by the selection circuit (21), the result of integration in each area of the integrator (23) results in each area of the color difference signal (RY). Digital integration values are obtained as 64 color evaluation values (rij). Further, in the next field, since the color difference signal (BY) is selected by the selection circuit (21), as a result of integration in each area of the integrator (23), the color difference signal (BY) is obtained.
Y) the digital integrated value for each area is 64 color evaluation values (bi
obtained as j). Thus, when the integration of the three fields of the luminance signal (Y) and the color difference signals (RY) (BY) is completed, the luminance evaluation value (yij) and the color evaluation value (rij)
The 64 × 3 value of (bij) is held in the memory (26). After that, the same operation as described above is repeated, and the luminance evaluation value (yij) in the next field and the color evaluation value (rij) in the next field are sequentially updated.

FIG. 11 shows the internal structure of the integrator (23) in more detail. Each A / D conversion data is supplied to the switching circuit (61). This switching circuit (61) receives the switching signal (S2) and
Adder (F11) (F12) with A / D conversion value prepared for each area
It has a role of supplying the adder for the area where the sampling point of the relevant data exists in (F88). That is, if the sampling point of certain arbitrary data is included in the area (A11), this data is supplied to the adder (F11) for the area (A11). In the same way, adder (Fij)
(I, j = 1 to 8) is set for the area (Aij), and a total of 64
Individual adders are provided. A holding circuit (Qij) is arranged after each adder, and each added value is temporarily held in each holding circuit. The data held in each holding circuit is input to the adder again and added to the data input next. Further, each holding circuit is reset for each field based on the vertical synchronizing signal, and only the held data immediately before this reset is supplied to the memory (26). Therefore, one digital integrator circuit is configured by one set of adder and holding circuit, and a total of 64 integrator circuits constitute the integrator (23), which means that 64 holding circuits are used for each field. The digital integrated value is input to the memory (26) for each of the regions.

When the integration for this one field is completed, this integrated value is held in the memory (26) as a brightness evaluation value or a color evaluation value.

The reference level of the color difference signals input to the A / D converter (22), that is, the zero level is preset to the level obtained when a perfect achromatic surface is photographed.
It goes without saying that the A / D conversion value can be a negative value as well as a positive value.

Latest evaluation value (yij) (rij) obtained from the above
(Bij) is input to the weighting determination circuit (27). The weighting determination circuit (27) determines the weighting amount of each area based on each evaluation value.

In the weighting decision circuit (27), so-called fuzzy inference that treats ambiguous boundary information while being ambiguous is used, and the following rules are used for each region.

[Rule 1] "if rij + bij is close to zero and rij-bij is close to zero then the weighting amount is large" [Rule 2] "if rij + bij is a little negative and rij-bij is close to zero "Slightly small" [Rule 3] "If yij is fairly large then the weighting amount is fairly small." These rules are as shown in Figs.
The expressions "close" and "big" are "rij + bij",
It is defined by membership functions for input variables such as “rij-bij” and “yij” and output variables such as “weighting amount”.

[Rule 1] is defined by a membership function as shown in FIGS. 4 (a), (b) and (c). FIG. 4 (a) is for determining the membership value of the condition (1) of the rule (1) that "rij + bij is very close to zero". ) (Bij) is a membership function for the sum of the two color evaluation values, and is a mountain-shaped function that takes a maximum value of "1" when the sum of the evaluation values of both colors is zero. Is set to be relatively steep. From this function, the membership value (u ₁₁ ) corresponding to the sum of the two color evaluation values can be obtained.

FIG. 4 (b) is for obtaining the membership value of the condition (2) of the rule (1) that "rij-bij is close to zero", and the color evaluation value (rij ) Is a membership function for the subtraction value obtained by subtracting the color evaluation value (bij) from the), and is a mountain-shaped function that takes the maximum value “1” when the subtraction value is zero. From this function, the membership value according to the subtraction value (U ₁₂ ) is obtained.

FIG. 4 (c) is a membership function for the weighting amount of each region, which is an output variable of the rule (2) that "the weighting amount is large", which is a function including a simple increasing straight line.
Membership value (u ₁₁ ) of both conditions of rule (1)
When the smaller of (u ₁₂ ) is determined, this value is specified on the vertical axis indicating the membership value, and the shaded area obtained by cutting off the head of the membership function is rule (1). It is an inference result obtained by illuminating the facts. For example, if the sum of the evaluation values of both colors and the subtraction value of a certain area are the values shown by (Δ) in FIGS. 4A and 4B, the membership values are u ₁₁ <u ₁₂ , U ₁₁ is specified on the vertical axis of FIG. 4 (c), the head of the membership function is cut off, and the center of gravity of the shaded portion corresponds to the weighting amount according to the degree of establishment of rule (1). It will be.

[Rule 2] is defined by a membership function as shown in FIGS. 5 (a), (b) and (c). FIG. 5 (a) is for obtaining the membership value of the condition (1) of the rule (2) that "rij + bij is a little negative". ) (Bij) is a membership function for the sum of both color evaluation values, and is a mountain-shaped function that takes a maximum value of "1" when the sum of the two color evaluation values is slightly smaller than zero. The membership value (u ₂₁ ) is calculated according to.

FIG. 5 (b) is for determining the membership value of the condition (2) of the rule (2) that "rij-bij is close to zero", and the condition of the rule (1) of FIG. 4 (b). It is the same as the function of (2), and the membership value (u ₂₂ ) according to the subtraction value is obtained from this function.

FIG. 5 (c) is a membership function for the weighting amount of each area, which is an output variable of the rule (2) that "the weighting amount is slightly small", and the maximum value is "1" when the weighting amount is around 0.25. If the smaller of the membership values (u ₂₁ ) (u ₂₂ ) of both conditions of rule (2) is determined by the type function, specify this value on the vertical axis indicating the membership value, The shaded area obtained by cutting off the head of the membership function is the inference result obtained by illuminating the fact in Rule (2). For example, if the sum and subtraction values of both color evaluation values of a certain area are values indicated by (Δ) in FIGS. 5 (a) and 5 (b), each membership value is u ₂₁ > u ₂₂ and U ₂₂ is specified on the vertical axis in Fig. 5 (c), the head of the membership function is scraped off, and the center of gravity of the shaded area corresponds to the weighting amount according to the degree of establishment of rule (2). Become.

[Rule 3] is defined by the membership function as shown in FIGS. 6 (a) and 6 (b). FIG. 6 (a) is for obtaining the membership value of the condition (1) of the rule (3) that "yij is considerably large", and it is a member for the brightness evaluation value (yij) in each region which is an input variable. The ship function is a function including a monotonically increasing straight line, and the membership value (u ₃₁ ) according to the brightness evaluation value is obtained from this function.

FIG. 6 (b) is a membership function for the weighting amount of each region, which is an output variable of the rule (3) that "the weighting amount is quite small". Specifying the membership value (u ₃₁ ) of (1) on the vertical axis, the shaded part obtained by cutting off the head of the membership function is the inference result obtained by illuminating the fact in rule (3). Become. For example, if the brightness evaluation value of a certain area is a value indicated by (Δ) in FIG. 6A, the head of the membership function is scraped off as shown in FIG. The center of gravity corresponds to the weighting amount according to the degree of establishment of the rule (3).

Next, a method of determining the weighting amount of the corresponding area will be described after considering all three rules. Fig. 4 (c), 5
Each of the shaded rectangles in FIGS. 6 (c) and 6 (b) is superimposed on the same coordinates as those in these diagrams, that is, the coordinates in which the horizontal axis represents the weighting amount and the vertical axis represents the membership value. The function in Fig. 7 obtained by this superposition becomes the membership function that shows the final inference result, and the position of the center of gravity of this function is the weighting amount (wij) of the relevant area, considering all the conditions of the three rules. Becomes Further, instead of the method of determining the weighting amount based on the position of the center of gravity described above, the position on the horizontal axis that halves the area surrounded by the membership function to the left and right can be used as the weighting amount. Needless to say.

Next, the influence of each rule on the white balance adjustment operation will be described, but before that, consider the relationship between the color temperature change of the light source and each color difference signal. Generally, the color difference signal when the color temperature of the light source illuminating a white subject changes
It changes like the color temperature axis of the light source in FIG. Therefore, the color difference signals appearing in the first and third quadrants that do not include the light source color temperature axis do not reflect the color temperature of the light source and are preferably not considered as information when performing white balance adjustment.

By the way, the color evaluation value (bi) of (BY) is plotted on the horizontal axis (X axis).
j) is the color evaluation value (rij) of (RY) on the vertical axis (Y axis), and when the light source color temperature axis is approximated, as shown in (L1) of FIG. 9 and FIG. Is a straight line with -1 (rij =
-Bij).

Therefore, in [Rule 1], regarding the screen evaluation at the time of white balance adjustment, when the color evaluation value of the corresponding region deviates from the origin, the component of the deviation in the direction orthogonal to the light source color temperature axis (L1) , Even if the deviation amount is small,
For the components in the direction of the light source color temperature axis (L1) while reducing the weighting amount of the corresponding area, at the same time, the degree of reduction of the weighting amount with respect to the deviation amount is compared to the orthogonal direction so that even if the deviation amount is slightly large The purpose is to be gentle.

That is, when the condition (1) of [Rule 1] is expressed by an equation, When K1 (y intercept) is set to a value ± a (a: positive value) that is very close to zero, the satisfaction degree (membership value) of condition (1) occurs on the straight line (L2) ( In the range surrounded by L3), and the closer to the straight line (L1), the higher the degree of establishment. That is, as shown in FIG. 4A, the weighting amount is set to rapidly increase only when a is close to zero, that is, the straight lines (L2) and (L3) are very close to the origin side.

Furthermore, when the condition (2) of [Rule 1] is shown by an equation, Then, if K2 (y intercept) is a value close to zero ± c (c: positive value), then equation (2) holds within the range surrounded by the straight lines (L4) and (L5), and as it approaches the origin. The degree increases. That is,
As c approaches zero, that is, the straight lines (L4) and (L5) approach the origin side, the weighting amount is set to gradually increase as shown in FIG. 4 (b). Incidentally, in FIG. 4 (b), the membership value becomes zero when c is extremely large. In this way, if the color evaluation value of the relevant area is at the origin, the weighting amount is naturally maximized, and even if the color evaluation value deviates from the origin, the deviation is in the red or blue color temperature axis direction and the light source color When the value is small in the direction orthogonal to the temperature axis (L1) (the direction with extremely high saturation), the weighting amount is increased to increase the contribution to image evaluation.

Further, when the condition (1) of [Rule 2] is shown by an equation, Therefore, if K3 is set to a slightly negative value −d (d: positive value), the membership value of condition (1) is generated by enclosing it with straight lines (L1) and (L6) as shown in FIG. It will be the range.

Further, since the condition (2) is the same as the condition (2) of [Rule 1], it is the range surrounded by the straight lines (L4) and (L5).

Therefore, ri when -d becomes the peak of the mountain in Fig. 5 (a)
Increase the weighting amount as it gets closer to the value of j + bij, and gently increase the weighting amount as c gets closer to zero, that is, as the straight lines (L4) and (L5) get closer to the origin side, as shown in FIG. 5 (b). As a result, when the color of the relevant area is slightly deviated in the green direction in FIG. 10 (the linear direction indicated by rij-bij = 0 in the third quadrant), although it deviates from the color temperature axis, it looks like a fluorescent lamp. Since a greenish light source is conceivable, the weighting amount is left slightly.

[Rule 3] is a case where the brightness of the area is extremely high, and the color balance is deviated due to the saturation of the image sensor and the color is biased toward a specific color. .

In the weighting determination circuit (27), when each evaluation value is input, inference is performed according to the above rule, and the weighting amount (wij) of each area is determined in the range of 0 to 1.

On the other hand, the color evaluation values (rij) (bij) (i, j: 1-8) are input to the screen evaluation circuit (28) and are weighted by the weighting determination circuit (27).
Using the weighting amount (wij) obtained in
Based on (2), the color difference value (V−Y) of the color difference signals (RY) and (BY) for the entire screen is the screen color evaluation value (Vr).
Calculated as (Vb).

The gain control circuits (29) and (30) use the R and B amplifier circuits (4) and (5) so that the screen color evaluation values (Vr) (Vb), which are color evaluation values for the entire screen, become zero. It controls the gain. In this way, if the screen color evaluation value (Vr) (Vb) becomes zero,
White balance adjustment is complete.

As described above, fuzzy inference is performed based on some preset rules, and the representative value of the screen color information is determined. Therefore, stable setting is possible without setting detailed conditions corresponding to various conditions. It is possible to realize the controlled control.

In this embodiment, the A / D converter (22) and the integrator (2
3) luminance signal (Y), color difference signal (RY) (BY)
It is shared in order to extract the level of 3 signals in each area by digital integration, and the integrated value of each signal could only be updated in 3 field cycles, but the A / D converter and the integrator were It is needless to say that each signal level can be updated for each field if it is provided exclusively for signals.

Furthermore, the operations of the weighting determination circuit (27) and the screen evaluation circuit (28) can be processed by software using a microcomputer.

(G) Effect of the Invention As described above, according to the present invention, the fuzzy inference based on some preset rules is performed to determine the representative value of the screen color information.

Thus, stable white balance adjustment can be realized without performing detailed condition settings corresponding to various conditions.

[Brief description of drawings]

1 and 3 to 11 relate to an embodiment of the present invention,
1 is an overall circuit block diagram, FIG. 3 is an explanatory diagram of screen division, FIG. 4 is a diagram showing a membership function of rule 1, and FIG. 5 is a diagram showing a membership function of rule 2.
FIG. 6 is a diagram showing a membership function of rule 3, FIG.
The figure shows the membership function of the final inference result.
FIG. 9 is a diagram for explaining the color temperature axis of the light source, FIGS. 9 and 10 are diagrams for explaining the rules 1 and 2, and FIG. 11 is a circuit block diagram of the main part. FIG. 2 is a circuit block diagram of a conventional example. (27) ... Weighting determination circuit, (28) ... Screen evaluation circuit, (4) (5) ... R and B amplification circuit.

Claims (2)

(57) [Claims] 1. A color evaluation value detecting means for obtaining a color information signal level for each of a plurality of areas set by dividing an imaging screen as a color evaluation value of each color, and a color evaluation value of each area for each area. Screen color evaluation value calculating means for performing weighting with a weighting amount and calculating a screen color evaluation value representative of the entire screen from the weighted color evaluation value, and an amplification gain of each color information signal based on the screen color evaluation value. In order to determine the weighting amount for each area in the screen color evaluation value calculation means, the input conversion is substituted into the membership function indicating the relationship between the input variable and the membership value for each rule. Fuzzy inference is used to obtain a membership value and obtain a conclusion part from the membership value for each rule. At least a value based on the color evaluation value for each area is used as an input variable, and the weighting amount is used as a conclusion part. Color video camera which is characterized in Rukoto. 2. The color video camera according to claim 1, wherein the brightness level for each of a plurality of areas is used as an input variable for fuzzy inference.