JPH04116636A - Exposure determining method for copying device - Google Patents

Abstract

PURPOSE:To reduce the variance of exposure to raise the acceptance ratio by determining the exposure with the fuzzy inference to obtain correlations strong against the exposure even in the case of the occurrence of positional deviation. CONSTITUTION:A color film image is divided to many parts to measure the light, and the certainty of image data obtained by light measurement is obtained by the membership function of image data to the position change of the image picture in a film carrier aperture part. Certainties are synthesized to obtain a weight, and image data at the light measuring point is weighted to determine the exposure. That is, the fuzzy inference is applied to determine the exposure. Since the weight is determined based on a maximum certainty in this manner when the light measuring point exists in a preliminarily determined area, strong correlations are obtained between the exposure in a correct position and that in a deviated position regardless of deviation of a film or the like from the correct position, and the variance of exposure is reduced.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a method for determining the exposure amount of a copying device, and more particularly to a method of determining the exposure amount of a copying device such as a photo printing device that prints and copies a film image onto photographic paper. Regarding.

[Problems to be solved by the conventional technology and invention; Conventionally, the film screen is automatically stopped at the printing position, the film screen is divided into many parts, and the exposure is measured automatically, and the exposure amount is automatically determined by analyzing the photometric data. Automatic photo printers that determine and print images are now in widespread use. However, since the pass rate of prints obtained with this automatic photo printing apparatus is about 70 to 98%, it is necessary to correct the exposure amount and reprint (reprint) rejected prints. This reprinting is particularly frequent in printing apparatuses that require high quality such as enlarged prints.

When reprinting, the previous print conditions are corrected and printed, but the print results may not be as correct as the corrections. The main reason for this is that the set position of the film screen during the first print and the set position of the film screen during the second print are slightly different, resulting in differences in the obtained photometric data or image features, and the exposure This is a guarantee that the quantity determination is not reproducible. In particular, during the first print, the film screen is automatically set to the printing position, and during the second print, the operator selects the film frame and sets it manually, so the deviation in the film setting position becomes large. There is a tendency. This problem also occurs when reprinting (reordering).

The present invention has been made to solve the above problems,
It is an object of the present invention to provide an exposure amount determination method for a copying apparatus that can reduce variations in the determined exposure amount due to variations in the film set position.

[Means to solve the problem]

In order to achieve the above object, the present invention measures the photometry of a color film image by dividing it into a large number of parts, and determines the membership of the image data with respect to the positional change of the film image at the film carrier opening, regarding the image data obtained by photometry. The reliability of the image data is determined by a function, the reliability is combined to determine a weight, and the exposure amount is determined using the image data of the photometric point and the weight.

3. Effects: In the present invention, a color film image is divided into a large number of parts and photometered, and the image data obtained by photometry is determined based on the membership function of the image data with respect to the change in the position of the film image at the film carrier opening. Find the degree. This certainty becomes maximum when the photometric point exists in a predetermined area, and increases as it approaches the predetermined area. Then, the certainty factors are combined to obtain a weight t, and the image data of the photometric point is weighted to determine the exposure amount. The present invention applies fuzzy reasoning to exposure determination, and weights are determined based on the confidence that is maximized when the photometric point is in a predetermined area. Even if there is a shift, a strong correlation can be obtained between the exposure amount at the correct position and the exposure amount when the position is shifted, and variations in the exposure amount can be reduced.

[Effects of the Invention] As explained above, according to the present invention, since fuzzy inference is used, a strong correlation can be obtained with the exposure amount even if a positional shift occurs, reducing variations in the exposure amount and increasing the pass rate. The effect is that it can be made higher.

〔Example〕

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

In the first embodiment, the present invention is applied to a color photographic printing apparatus, and FIG. 2 schematically shows a color photographic printing apparatus to which the present invention is applicable.

A mirror box 18 and a lamp house 10 equipped with a halogen lamp are arranged below the color negative film 20 loaded in a negative carrier and transported to a printing section. A light control filter 60 is arranged between the mirror box 18 and the lamp house 10. Light control filter 6
0 is composed of three color filters: a Y (yellow) filter, an M (magenta) filter, and a C (cyan) filter, as is well known.

A lens 22, a black shutter 24, and a color paper 26 are arranged in this order above the negative film 20, and the light from the lamp house 10 illuminates the light control filter 60.
, the light beams transmitted through the mirror box 18 and the negative film 20 are imaged onto the color paper 26 by the lens 22.

A sensor 28 is arranged in a direction oblique to the optical axis of the imaging optical system and at a position where the image density of the negative film 20 can be measured photometrically. The sensor 28 is composed of a two-dimensional image sensor composed of a CCD (charge-coupled device), a line sensor, etc., and divides the negative image into a large number of pixels Sn along the scanning line SL, as shown in FIG. to measure the light.

In this case, photometry for each pixel is performed for the three primary colors B, G, and R. This sensor 28 detects position data and photometric data of each pixel.

The sensor 28 is connected to an exposure control example circuit 40 that controls the exposure amount by calculating an exposure amount control value and controlling the dimming filter 60 via the driver 30. The exposure control circuit 40 includes a read-only memory (ROM) storing the exposure control routine program shown in FIG. 1, membership functions, etc. described below, a random access memory (RAM), and a central processing unit (
It consists of a microcomputer with a CPU.

Next, the fuzzy inference rules of the first embodiment will be explained. First, in this embodiment, the following feature quantities V+, V2, V3, and V4 are employed as feature quantities obtained from photometric data. Note that these feature quantities V1, V, , V3, V41tR
, G, and 83 colors, but the following explanation will be made without distinguishing between them.

Vl: Area ratio of the low-density part in the image to the total screen area Average density of the area including the minimum density V3: Skin color density or highlight area density with a specific color characterized by skin color - screen average density ＼, D+, D+-1 is the density of the adjacent pixel Sη,
N: Well, it is the total number of pixels.

As a fuzzy inference rule), the following five rules (1) to (5) are used in the form 1f-theT'1~.

(1) If the feature value V1 is small and the feature value ~" is small, use the exposure calculation formula F1. ), increase the confidence level, that is, the weight.

(2) If the feature value v1 is small and the feature value ˜2 is large, use the exposure amount calculation formula F2 to increase the confidence level for the final exposure amount of the exposure amount obtained using the exposure amount calculation formula F2.

(3) If the feature amount V1 is large and the feature amount V3 is large, the degree of confidence in the final exposure amount of the exposure amount obtained using the exposure amount calculation formula F3 is increased.

(4) If the feature amount v1 is large, the feature amount v3 is small, and the feature amount V is small, increase the confidence level for the final exposure amount of the exposure amount obtained using the exposure amount calculation formula F4.

(5) If the feature amount v1 is large, the feature amount v3 is small, and the feature amount v4 is large, increase the degree of confidence in the final exposure amount of the exposure amount obtained using the exposure amount calculation formula F5.

The linguistic values such as large and small are the feature quantities V including the judgment values (however, i-1, 2, 3, 4
) is quantified by the membership function of This member function is determined as follows. First, the position of the photometry area relative to the screen is shifted to the left and right in stages, and the fourth
Characteristics when the photometry area indicated by the broken line in figure (B) is in the correct position
'll V Or (predetermined value), the difference in the feature quantity ■1 ' when the photometry area is shifted △V, (-V, '-V
. 1) Create a histogram. In addition, Fig. 4 (A)
, (C) show the photometry areas shifted to the right and left from the normal position, respectively. This histogram becomes as shown in FIG. 5(B). The part of this histogram with the highest frequency is the part where there was no particularly slight change even though the photometry area was shifted (difference Δ10). Therefore, the membership function is determined so that the membership tab value of the special iFk@V corresponding to the point where the frequency of the histogram is maximum, that is, the -political change is 0.5, and the determination value is 1. In addition,
Feature ff corresponding to the point where the frequency of the histogram is maximum
1V, (to the -judgment value) main tab value is 00
It is possible to set it to a value between ˜1.0. Also,
The feature value V1 whose menstub value is 1.0 and OIO is
The width may be made to match the value at which the frequency of the histogram is minimum or around this value, or may be made to match the width of the value where the frequency is between the maximum and minimum. Furthermore, the member function may be determined by other methods. For example, it is also possible to create a film in which the same subject is photographed at slightly different positions, and to determine from the variation in the used feature amount or the distribution of the variation.

When the membership function is determined using the above method, the feature quantity ■1
For the above rule (1), the membership function of
For Figure (A), Rule (2), Figure 6 (B), Rule (
6(C) for rule (3), FIG. 6(D) for rule (4), and FIG. 6(E) for rule (5). In the membership function of the feature quantity V+, the membership value is set to 1 (for example, 30, the value is 100 times the density value, and the same applies hereinafter) to V1- and 0.5, respectively.
Its position is indicated by a dotted line. In addition, the member 7NJ ship function of the feature V2 is shown in Fig. 6 (F
), and for rule (2), as shown in Figure 6 (G),
The membership value is determined to be 0.5 for each v2K 2 (for example, 48). Furthermore, the membership function for the feature V3 is shown in FIG. 6(H) for rule (3), FIG. 6(I) for rule (4), and FIG. 6(I) for rule (4).
), as shown in FIG. 6(J), each V3- has a membership value of 3 (for example, 76) and a membership value of 0. 5. The membership function for the feature V4 is as shown in Figure 6 (K) for rule (4) and Figure 6 (L) for rule [5].
= 4 (for example, 14), the ship value is determined to be 0.5. These member sub functions are
This determines the determination condition, and indicates that as the membership value increases, the degree of agreement with the determination condition of the feature amount v1 increases.

Note that the above exposure amount calculation formula F1 satisfies V, <K, and V2 <
At K2 (pattern with low contrast, mainly landscape), the exposure calculation formula F2 is V<K + and v
When 2≧2 (pattern centered on landscape with contrast), exposure calculation formula F3 is 1 for VI≧ and V3≧
3 (pattern with a background part with a monotonous structure (standard scene, backlight, etc.)), the exposure amount calculation formula F is v1 ≥ 1
And when V3<K3 and V4<K4 (patterns with contrasting background parts), and when the exposure calculation formula % formula % (patterns with main parts in highlights such as strobe lights and night scenes), the optimum exposure amount is It is determined that it will be obtained.

Then, the exposure amount calculation formula F, (i=1 to 5) is given, for example, as follows using the feature amount.

F+ =ko -I-+D-,,+2D+ai
n +L DM+ k 4 D++ + k s ・
S ten... (1) However, ko, ni+ ・
... is a constant, D□ is the maximum concentration, Dmin is the minimum concentration,
DM is the screen average density, D)I is the skin color density, and S is the area ratio of the blank area.

The membership functions and exposure amount calculation formulas determined as described above are stored in advance in the ROM of the exposure control circuit 40.

Note that the membership function may be determined using a constant ratio width region of a histogram representing the distribution of the feature amount V1 and the amount of variation in the feature amount obtained by repeatedly photometrically measuring an image while changing conditions.

Next, the exposure control routine of this embodiment for controlling the exposure amount using the above membership function will be explained with reference to FIG. In step 100, the photometric data detected by the sensor 28 is taken in, and in step 104, the above-described feature amounts V1 to V4 are calculated using the photometric data. In the next step 106, the matching degree corresponding to the feature value V I""" V 4 is calculated from the membership function shown in FIG. 6 according to the fuzzy inference rules (1) to (5) explained above, i.e. The degree of coincidence with the determination condition is calculated for the feature values V1 to V4, and in step 108, the product w1 of the degree of coincidence is calculated for each of rules (1) to (5).
1, that is, the degree of matching for each of the above five patterns is calculated. Since the degree of coincidence with this pattern represents the degree of certainty of the exposure amount obtained from each exposure amount calculation formula with respect to the final exposure amount, it will be explained below as the degree of certainty. In the example shown in FIG. 6, for rule (1), the degree of coincidence with feature quantity V1, that is, the membership value, is 0.7 (represented by a solid line, the same applies hereinafter), and the degree of coincidence with feature quantity v2 is 0. .4, so the confidence level is 0.7X'0.4=0.28.

For rule (2), the degree of matching for the feature V1 is 0.
゜7, since the degree of matching for feature quantity v2 is 0.6,
The confidence level w2 becomes 0.7XO, 6=0.42. Rules (
For 3), the matching degree for the feature amount V1 is 0.3,
Since the matching degree for the feature amount V3 is 0, the confidence level w3
becomes 0.3XO=0. For rule (4), the degree of coincidence with the feature quantity v1 is 0°3, the degree of coincidence with the feature quantity v3 is 1.0, and the degree of coincidence with the feature quantity v4 is 0.9, so the confidence level W4 is 0. .3X1. OXo, 9=0.2
It becomes 7.

Then, for rule (5), if the degree of coincidence for feature quantity v1 is 0.3, the degree of coincidence for feature quantity V3 is 1°0, and the degree of coincidence for feature quantity v4 is 0.1, the confidence level is W
5 is 0.3X1. OXo, 1=0°03.

In the next step 110, the reflection degree is normalized according to the following formula.

By performing normalization processing in this way, ΣW11.0 is obtained.

In the next step 112, the exposure amount f1 is calculated from the exposure amount calculation formula F1.
The final exposure amount X is obtained by adding a weight corresponding to , and integrating the value as shown in equation (3).

X-ΣW, fl...(3) Then, by driving the driver 30 with the exposure control value determined by this final exposure amount X, the light control filter 6G
Change the position by 1"": Change the exposure by 1.

Above, we explained an example of calculating confidence by calculating the product of coincidence degrees (fuzzy inference by algebraic product), but as explained below, logical It is also possible to use fuzzy inference by That is, the sixth
In the example shown in Figure 7, which uses the same membership functions as in the figure, rule (1)! ': For the feature value V1, the matching degree, that is, the membership value is 0. 7. Since the degree of coincidence with respect to feature v2 is 0.4, the degree of certainty is 0.
Set it to 0, 4, which is the minimum value within 70.4. Rule (2)
, the matching degree for the feature amount V1 is 0. 7. The degree of matching for the feature amount V2 is 0. 6, so the confidence level w2 is 0. 7.0. Set it to 0°6, which is the minimum value among 6. For rule (3), the degree of coincidence with the feature quantity V is 0.3, and the degree of coincidence with the feature quantity V3 is O, so the confidence level W3 is 0. Set it to 0, which is the minimum value within 3.0. For rule (4), the degree of coincidence with the feature quantity V1 is 0.3, the degree of coincidence with the feature quantity V3 is 1.0, and the degree of coincidence with the feature quantity v4 is 0.9, so the confidence degree W4
is the minimum value of 043.1.0.0.9,
Make it 3. Then, for rule (5), the feature value V1
Since the matching degree for the feature amount V3 is 0.3, the matching degree for the feature amount V3 is 1.0, and the matching degree for the feature amount V is 0.1, the confidence level W5 is 0.3.1. Set it to 0 and 1, which are the minimum values among Olo and 1.

The distribution of exposure changes in this embodiment (fuzzy inference by algebraic product, fuzzy inference by logical product) due to film set position shift when controlling the exposure dose as described above using the same original image is compared to the conventional example ( Figure 8 shows a comparison with the distribution for the current model.

In FIG. 8, ΔDK indicates a change in exposure due to misalignment of the film set position, and a change of 0.5 corresponds to a 10% change in exposure. As can be understood from the figure, in fuzzy inference by algebraic product and fuzzy inference by logical product, Δ
Parts with small DK (00~0.5, -〇, 5~0.0
), the distribution is higher than the current one, and it can be seen that even if the film set position is shifted, the change in exposure amount is small. In addition, in the fuzzy inference based on algebraic product in this example, the exposure amount changes with respect to film set position deviation is smaller than in the fuzzy inference based on logical product. Better results are being obtained.

Next, set the exposure amount at the normal position and the film by 2.5 mm from the normal position.
FIG. 9 shows a comparison between the present example and the conventional example, and the correlation with the exposure amount when the light is increased by n+m. FIG. 9(1) shows the conventional correlation, and FIG. 9(2) shows the correlation of this embodiment.

Further, the vertical axis in FIGS. 9(1) and 9(2) shows the exposure amount at the normal position, and the horizontal axis shows the exposure amount after shifting from the normal position. The conventional correlation was 0.869, but the correlation in this example was 0.
920, and there was a strong correlation between the exposure amount at the normal position and the exposure amount after the shift.

As explained above, according to this embodiment, even if the film set position is shifted, there is little change in the exposure amount, so the difference in exposure amount for similar images is reduced, and a series of similar images are reproduced as uniform colors and densities. It becomes two things that are done.

Next, the fuzzy inference side of the second embodiment will be explained.

First, in this embodiment, as shown in FIG. 10, the screen 50 of the color film is divided into a center region R1 and a peripheral region R3, which are predetermined regions. An overlapping region R2 is provided between R3 and R3. Furthermore, the following distances x and y are used as position data.

X: Horizontal distance of pixels on the screen 50 with the left side of the screen 50 as the origin y: Vertical distance of the pixels on the screen 50 with the top side of the screen 50 as the origin On the fuzzy inference side, 1f-then~ Use the following rules in the form:

(1) If the horizontal distance X is medium and the vertical distance y
If is medium, the pixel is in the central region.

(2) If the horizontal distance X is large and the vertical distance y is large, the pixel exists in the peripheral region.

(3) If the horizontal distance X is large and the vertical distance y is small, the pixel exists in the peripheral region.

(4) If the horizontal region X is small and the vertical distance y is large, the pixel exists in the periphery.

(5) If the horizontal distance X is small and the vertical distance y is small, the pixel exists in the periphery.

The above-mentioned medium, large, and small language values are quantified by the Men/Nl ship function of the distance x, y, which will be explained below. As shown in FIG. 10, the membership function for distance Membership value is 1.0 to 0
When the distance X is a value corresponding to the central region R1, the membership value becomes 1, and when the distance X is a value corresponding to the overlapping region R2, the membership value becomes There is a membership function M2 that varies from 1.0 to 0. Further, the membership function for the distance y has a membership value of 1.0 when the distance y is a value corresponding to the peripheral region R3, and a membership value of 1.0 when the distance y is a value corresponding to the overlapping region R2. The membership function changes from 0 to 0, L, and the membership value is 1 when the distance y is a value corresponding to the central region R,
There is a membership function L2 in which the membership value changes from 1.0 to 0 when the value is 0 and corresponds to the overlap region R2. Note that since the amount of vertical shift of the image frames is small in - boat fishing, the slope of the membership function with respect to the distance y with respect to the overlapping area is larger than the slope of the membership function with respect to the distance X with respect to the overlapping area.

The above membership function is determined in advance by the RO of the exposure control circuit 40.
It is stored in M.

Next, the exposure control routine of this embodiment for controlling the exposure amount using the above membership function will be explained with reference to FIG. The count value l is set to the initial value (0), and in step 100, the photometric data detected by the sensor 28 is taken in, and in step 104, the photometric data D is
Get the distance x, y of pixel 1. In step 106,
Determine the membership value based on the membership function described above. To explain this membership tab value for the pixel marked x in FIG. 10, as shown in FIG. 11, the membership value for the distance X is 0.3 for the membership function M, and 0.7 for the membership function M2. , and the membership value for the distance 11itty is 1.0 for the membership function L2.

In the next step 108, the minimum membership value for the central region R1 (in the above example, the membership value is 0) is determined.
．． 7 and 1.0, so 0°7), and the minimum membership value for the peripheral region R2 (in the above example, the membership value is 0.3 and 0, so 0) is determined. In the next step 110, the weights for the minimum membership values are calculated, and the larger weight is determined by Wl (
i=1.2.3, . . . nSn is the total number of pixels).

Then, in step 112, the count value is incremented by 1, and in step 114, it is determined whether or not 1≧n, thereby determining whether the weight Wl for all pixels on one screen has been calculated, and the weight Wl for all pixels is determined. If the weights of all pixels have not been determined, the steps described above are repeated, and when the weights of all pixels have been determined, the process advances to step 116. In step 116, the value obtained by multiplying each of the photometric data D1 by Wl is used as new photometric data D1.
Then, in step 118, the exposure amount is calculated using this new photometric data D1.

As described above, it is possible to reduce the variation in the exposure amount determined due to the variation in the set position on the negative carrier, and it is possible to stably determine the exposure amount with good reproducibility.

[Brief explanation of drawings]

FIG. 1 is a flowchart showing the exposure control routine of the first embodiment of the present invention, FIG. 2 is a schematic diagram of a color photographic printing apparatus to which the present invention can be applied, and FIG. 3 is a photometry meter using a sensor that divides the screen into planes. The diagrams shown in Figure 4 (A), (B), and (C) are
A diagram showing the positional deviation of the photometric area with respect to the screen, FIG. 5 is a pictorial diagram showing the relationship between the histogram and the membership function, FIGS. 6 and 7 are diagrams showing the membership function, etc. of the first embodiment, FIG. 8 is a diagram showing the distribution of changes in exposure amount due to misalignment of the film set position, FIG. The figure is a diagram showing the relationship between regions and membership functions in the second embodiment of the present invention, Figure 11 is a diagram explaining how to calculate weights, and Figure 12 is a diagram showing the exposure amount calculation in the second embodiment. 2 is a flowchart showing a routine. 28...Sensor, 30...Driver 40...Exposure control circuit.

Claims (3)

[Claims] (1) A color film image is divided into a large number of parts and photometered, and the reliability of the image data obtained by photometry is determined by a membership function of the image data with respect to a change in the position of the film image at the film carrier opening. A method for determining an exposure amount for a copying apparatus, comprising: determining a weight by combining the certainty factors, and determining an exposure amount by weighting image data of a photometric point. (2) The exposure amount determination method for a copying apparatus according to claim 1, wherein the image data is a feature amount used for determining the exposure amount of a color film image. (3) The exposure amount determining method for a copying apparatus according to claim (1), wherein the image data is the density of divided points on the film screen.