JPH05300421A - Automatic focusing device - Google Patents

Abstract

PURPOSE:To simplify the configuration and to enable high-speed processing by performing inter-picture element differential processing concerning image signals, calculating the dispersion of outputs and performing automatic focusing with the processed data as an evaluation amount. CONSTITUTION:Image beams are converted to image signals by an image pickup device 3 composed of an optical system 1 and an photoelectric converter 2. This image signal is inputted to an inter-picture element differential processor 5, and differential data between adjacent picture elements are calculated. Next, the absolute value of the differential data between adjacent picture elements is processed by an absolute value processor 6. For the processed data outputted from the processor 6, the distributed value is calculated by an entropy calculator 7. The distributed value outputted from this device 7 is defined as an automatic focus evaluation amount 8. Namely, automatic focusing is enabled by using the change of this distributed value.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic focusing device for automatically focusing an optical system for image pickup.

[0002]

2. Description of the Related Art Generally, an image pickup device such as a video camera is
It is equipped with an automatic focusing device that performs so-called focusing, which automatically focuses on a subject. Various systems are known for such an automatic focus adjustment device, and one of them is a system for focusing using an image signal obtained from an image pickup unit.

The autofocus system in which focus control is performed by using a picked-up image signal does not require a special optical system or sensor for autofocus as compared with other systems, and the mechanism is extremely simple. Regardless, it has many advantages of being able to focus accurately. A mountain climbing servo system, which is an example of such an auto focus system, is described in the document "NHK Technical Report (Vol. 17, 1970, No. 1, Vol. 86, No. 26, pp. 26)" Automatic focus adjustment of TV camera by mountain climbing servo system "". It is disclosed.

The principle of this system utilizes the fact that the high-frequency component existing at the edge portion of the image is large when the image is in focus, and the high-frequency component becomes smaller as the defocusing is performed. ..

The basic method for performing autofocus using a picked-up image signal is disclosed in the above-mentioned document, and a more specific operation will be described with reference to the drawings.

FIG. 7 is a block diagram showing a conventional autofocus device. In the figure, reference numeral 50 denotes a video signal extraction area, that is, a gate opening / closing signal for setting a relatively narrow specific area in an image pickup screen as a sampling area. GC
1 is a gate control circuit, 51 is a gate control circuit 5
The gate circuit that passes only the luminance signal in the sampling area from 0, the focus evaluation value creation unit 60 includes a high-pass filter (HPF) 52, a detection circuit 53, an integration circuit 54, A
This is a system that includes an A / D converter 55 and outputs the evaluation amount as a digital signal.

In the focus evaluation value creating section 60, HP
F52 has a cutoff frequency of 200 kHz to 800
Among the luminance signals from the image pickup circuit 56, which has a value selected between kHz and which outputs an image pickup signal from an image pickup element (not shown) as a television system video signal, the gate circuit 5
Only high frequency components are extracted from the signal corresponding to the sampling area that has passed 1. The obtained high frequency component is integrated with the luminance signal amplitude-detected by the next detection circuit 53 for each field, and is transmitted as a signal representing the level of the high frequency component in the corresponding field of the luminance signal. The output of the integrating circuit 54 is A
The signal is converted into a digital signal by the / D converter 55 and extracted as an evaluation amount.

The operation of the above-described structure will be described below. The subject is converted by the image pickup circuit 56 into a video signal including a luminance signal having a horizontal synchronizing signal and a vertical synchronizing signal.

The luminance signal in this video signal is input to the gate circuit 51 and the sync separation circuit 57. The sync separation circuit 57 separates the vertical and horizontal sync signals from the given luminance signal, and performs gate control. It is given to the circuit 50. The gate control circuit 50 has a fixed oscillator that oscillates at a constant frequency, and a predetermined region of the luminance signal is temporally defined based on the vertical synchronizing signal, the horizontal synchronizing signal, and the fixed oscillator output, and the luminance signal is required. A gate opening / closing signal GC1 is generated so as to pass only the portion and is output to the gate circuit 51.

As a result, the gate circuit 51 to the HPF 52
Is given only the luminance signal component corresponding to the sampling area set by the gate control circuit 50,
Only the high frequency components of the corresponding area are extracted. The high frequency component extracted by the HPF 52 is amplitude-detected by the detection circuit 53 at the next stage. The detection output of the detection circuit 53, that is, the level of the high frequency component is given to the integration circuit 54, where it is integrated for each field and the A / D converter 5 is used as a focus evaluation value.
Given to 5. The A / D converter 55 outputs the evaluation amount as a digital signal.

On the basis of this evaluation amount, the hill-climbing servo type autofocus is a process in which the focus lens is moved to the reflection side when the evaluation amount becomes small and the peak of the evaluation amount is obtained by trial and error.

In the above-mentioned conventional example, the cutoff frequency of the HPF 52 for extracting the edge component of the image is not fixed but needs to be changed according to the situation. This is because in the defocused state, the edge components of the image are small and are concentrated in low frequencies, so the cutoff frequency cannot be made too high. However, although the high-frequency component increases as it approaches the in-focus point, the amount of change corresponding to the movement of the focus lens cannot be accurately detected if the cutoff frequency remains low because of the low-frequency component due to the pattern of the captured image.
Therefore, it is generally necessary to increase the cutoff frequency as the focal point is approached.

Furthermore, in many video cameras and video tape recorders with a built-in camera, the optical system has a zoom function, and the photographing field angle often changes. In such a device, the gate control circuit 5 shown in FIG.
When 0 and the size of the sampling area set by the gate circuit 51 are fixed, the high frequency component of the edge portion of the image cannot be accurately extracted in response to the change in the size of the subject due to zooming. For example, it is conceivable that the subject becomes large when the image is picked up on the telephoto side and jumps out of the sampling area. On the other hand, the subject becomes smaller on the wide-angle side and the amount of edge components of the subject in the focus evaluation amount decreases, so that the change in the evaluation amount has no correspondence with the change in the defocus during zooming, and the mountain climbing servo cannot be performed. To solve this, the size of the sampling area is made variable according to the zoom (that is, the focal length).

[0014]

Since the conventional autofocus device is constructed as described above, there are many parts to be processed by analog, and the evaluation amount cannot be accurately extracted due to circuit variations and temperature characteristics. Also, the circuit has to be switched depending on the state of the input image signal.

Further, in the conventional autofocus device,
Since the high frequency components of the entire sampling area are targeted, for example, for subjects with different distances or subjects with depth, the defocus amount and the evaluation amount do not correspond exactly, and the autofocus adjustment mechanism gets confused. There was a problem.

The present invention has been made to solve the above problems, and an object of the present invention is to obtain an automatic focusing apparatus having a simple structure and capable of high-speed processing.

[0017]

According to another aspect of the present invention, there is provided an inter-pixel difference processing means for calculating an inter-pixel difference of an image signal from an image pickup device and an absolute value of its output. And an entropy calculation means for calculating the variation in the output values,
The value output from the entropy calculating means is extracted as an autofocus evaluation amount.

According to another aspect of the present invention, there is provided an inter-pixel difference processing means for calculating the inter-pixel difference of the image signal from the image pickup device, and an absolute value processing means for calculating the absolute value of its output. It is composed of a maximum value calculating means for selecting a maximum value from the output values or a plurality of values having a large value, and the value output from the maximum value calculating means is extracted as an evaluation amount without autofocus. ..

In the automatic focusing apparatus according to the third aspect of the present invention, the image signal of the image feature extraction area which is an area for performing entropy calculation is extracted from the image signal from the image pickup device, and the image feature extraction area is extracted. The small area division processing means for dividing into a plurality of small areas, the first entropy calculation means, and the output of the first entropy calculation means are used as the characteristic amount, and the maximum value is calculated among the characteristic amounts. The maximum value calculating means, or the sum total calculating means for adding a plurality of the characteristic amounts, or the average value calculating means for calculating the average value of the respective values of the characteristic amounts, the maximum value calculating means or the sum total The value output from the calculation processing means or the average value calculation means is extracted as the autofocus evaluation amount.

In the automatic focusing apparatus according to the fourth aspect of the present invention, the image signal of the image feature extraction area which is an area for performing entropy calculation processing is extracted from the image signal from the image pickup device, and the image feature extraction area is extracted. Is divided into a plurality of small areas, a first entropy calculating means, and an output of the first entropy calculating means as feature quantities, and a small area having a large value among the feature quantities is selected. It is composed of a small area selecting means for selecting one or a plurality of areas, a block division processing means for further subdividing the small areas, and a second entropy calculating means for each block, which is output from the second entropy calculating means. This value is extracted as the autofocus evaluation amount.

In the automatic focusing apparatus according to claim 5 of the present invention, the small area division processing means, the orthogonal transformation means for performing orthogonal transformation within the small area, and the maximum value of the transformation coefficient obtained by the orthogonal transformation. Maximum value calculating means for calculating
Alternatively, it is composed of a variance value calculating means for calculating the variance value of the transform coefficient obtained by the orthogonal transformation, and the maximum value calculating means or the value output from the variance value calculating means is extracted as an autofocus evaluation amount. It is a thing.

[0022]

In the automatic focusing apparatus according to the first aspect of the present invention, the pixel signal difference processing is performed on the image signal output from the image pickup apparatus, and the degree of variation in the output value is calculated,
Automatic focus adjustment is performed by using the processing result as an evaluation amount corresponding to the defocus amount.

In the entropy calculation means according to the second aspect of the present invention, the evaluation amount is calculated by using the maximum value calculation means for the variation in the inter-pixel difference value.

In the automatic focus adjusting device according to the third aspect, the image signal from the image pickup device is divided into small regions, the evaluation amount is calculated for each small region using the entropy calculating means, and the automatic focus adjustment is performed based on the evaluation amount. I am trying.

In the automatic focusing apparatus according to the fourth aspect, the small area selecting means selects a small area having a large feature amount, and the second entropy calculation processing means selects the evaluation amount in the selected small area. It is calculated and the automatic focus adjustment is performed based on it.

In the automatic focusing apparatus according to the fifth aspect, the image signal is transformed into the frequency domain component by the orthogonal transformation means,
The degree of variation in the orthogonal transform coefficient indicating the size of the region spectrum is calculated, and the calculation result is used as the evaluation amount corresponding to the defocus amount to perform automatic focus adjustment.

[0027]

EXAMPLES Example 1. FIG. 1 is a block diagram showing an example in which an inter-pixel difference processing device is used in the automatic focus adjustment device of the present invention. The image light is converted into an image signal by an image pickup device 3 including an image pickup optical system 1 and a photoelectric converter 2, and the image signal passes through an inter-pixel difference processing device 5 and an absolute value processing device 6 to calculate entropy. The value output from the device 7 is the autofocus evaluation amount 8.

Video signal 4 output from the image pickup device 3
Is a digital signal. The inter-pixel difference processing device is a process for extracting a feature of an image, and by obtaining a difference between adjacent pixels, it is possible to extract an amount of change in luminance between adjacent pixels. The probability density distribution of the difference data between adjacent pixels shows a Laplace distribution generally used in statistics, and can be expressed by the following equation (1) where the standard deviation is σ.

[0029]

[Equation 1]

In the vicinity of the in-focus point, the difference in the difference data between adjacent pixels is large, and in the out-of-focus state, it is small and concentrates in the vicinity of 0.

Since the difference data between adjacent pixels requires only the amount of change in future processing, the absolute value processing device 6
Performs absolute value processing.

The processed data output from the absolute value processing device 6 is processed by the entropy calculating device 7. The entropy processing device 7 is for calculating the variation in the difference data between adjacent pixels, and calculates the variance value used in statistics. As described above with respect to the difference data between adjacent pixels, the variation of the data becomes large near the focal point, and therefore the variation can be evaluated by the numerical value by performing the variance value calculation process.

Therefore, when the dispersion value is calculated at a certain position of the focus lens and then the dispersion value is similarly calculated after moving the focus lens, if the lens is moved toward the in-focus point. , The latter value becomes larger,
Since the latter becomes smaller if it is moved away from the in-focus point, automatic focus adjustment can be performed using the change in the dispersion value.

The output from the entropy calculation device 7 is used as the evaluation amount 8, and the calculated output for determining the focus lens drive amount in the automatic focus adjustment device is used as the evaluation value of autofocus.

Further, the calculation time can be shortened by performing the thinning-out process of the adjacent pixel difference data before the entropy calculation process. In the entropy calculation process, the calculation shown in the following equation (2) is performed on all the data between adjacent pixels.
Obviously, if the number of data is thinned to 1/2, the amount of calculation will also be 1/2. However, if the thinning-out amount is increased, the accuracy of the evaluation value becomes worse. Therefore, it is necessary to thin out the adjacent pixel difference data within a range in which the required accuracy can be maintained.

[0036]

[Equation 2]

Further, the data length of the adjacent pixel difference data can be shortened by performing the gradation process on the adjacent pixel difference data before the entropy calculation process. Further, by performing the gradation processing, it is possible to smooth the noise contained in the gradation width, which has the effect of reducing the noise in the processed data.

In the above embodiment, the difference between adjacent pixels is used. However, this is less accurate, but the difference value may be obtained every other pixel or every two pixels. This method has a weak correlation with the defocus amount, but has an advantage of being strong against noise. In addition, there is an advantage that the calculation amount can be reduced. Furthermore, the statistical property that the difference between pixels shows a Laplace distribution has the same property not only in the horizontal direction of the image but also in the vertical direction, so either or both of them are calculated and used as the evaluation amount. Good.

Example 2. FIG. 2 is a block diagram of an automatic focus adjusting device using the maximum value calculating device 9. The image signal output from the image pickup device 3 is subjected to the inter-pixel difference processing by the inter-pixel difference processing device 5, and is then subjected to the absolute value processing by the absolute value processing device 6. For the processing data subjected to the absolute value processing, the maximum value of the processing data is calculated by the maximum value calculation device 9, and the calculated maximum value is used as the evaluation amount 10.

The image pickup device 3, the inter-pixel difference processing device 5, and the absolute value processing device 6 are provided, for example, in the first embodiment.
Is the same as that described in.

The maximum value calculation device 9 pays attention to the fact that the inter-pixel difference data described in the first embodiment shows a Laplace distribution and the variation becomes large near the focal point, and instead of calculating the variance, Similar results can be obtained by using the maximum value calculating means for selecting the largest value among the pixel difference data. The distributed value calculation in the first embodiment requires a calculation time because multiplication and addition are performed for each data, but the maximum value calculation is advantageous because it only requires subtraction.

Further, when there is a singular point such that the subject has a high brightness, the difference between pixels shows a large inclination even in the defocused state, so the processing data at that point is selected as the evaluation value. there is a possibility. Therefore,
By calculating a plurality of processing data from the one with the largest inter-pixel difference data and using the sum of the values as the evaluation amount, the processing data of only the singular point is selected, and the correlation between the evaluation amount and the defocus amount. Even if the relationship is broken, erroneous processed data will not be used as the evaluation amount. Also, regarding noise,
Similar effects can be obtained.

In a device for moving images such as a video tape recorder with a built-in camera, the time required for focusing becomes a problem in terms of performance. That is, the defocus amount and the direction in which the focus lens is moved are determined based on the above evaluation amount. However, when it takes time to calculate the evaluation amount, the subject is moving when the calculation is completed, and the shooting situation is It is changing (as a result, the true evaluation amount is changing). In such a case, it means that autofocus does not work well for a moving subject. Therefore, Example 2
It is essential to reduce the amount of calculation as shown in.

Example 3. FIG. 3 is a block diagram showing an example in which a small area division processing device is used in the automatic focus adjustment device of the present invention. The image pickup device 3 is the same as that in the first embodiment. The image signal 4 output from the image pickup device 3 is divided into small regions by the small region division processing device 11, the inter-pixel difference processing device 5 performs inter-pixel difference processing, and the absolute value processing device 6 performs absolute processing. Value processing is performed.

The data calculated in each area after the absolute value processing is calculated by the entropy calculating device 7,
The maximum value calculation 12 is performed on the entropy calculation data of each area, and the output is set as the evaluation amount 15.

The small area division processing device 11 performs an entropy calculation processing on the entire area (entire screen) of the image signal, for example, an image of a half area of the entire area at the center of the image. A feature extraction area is defined, and the area is divided into a plurality of blocks, for example, 30 small areas.

After the division into the small areas, the degree of dispersion is calculated by the entropy calculating device 7 in each area, whereby the characteristic amount 30 indicating the complexity of the image in the small area is calculated.

When the entropy calculation processing is performed in the entire area, that is, when one feature amount is extracted from one area, the characteristic of the image appears, but there is no ability to detect a change in detail. However, in the small area division processing, after performing the small area division processing, if the subsequent calculation processing is performed in each divided area, the feature amount of each divided area can be known, and all the characteristics By performing the process of comprehensively judging from the amount, different from the above,
It is possible to extract the evaluation amount considering the details of the image.

Further, since the feature amount is extracted according to the distance of the subject included in each region by dividing the region into the above-mentioned small regions, the subject having different distance can be seen by observing the change in the feature amount of each region. It is also possible to handle subjects with depth.

A large value is calculated for the characteristic amount 30 when it is in focus, so the maximum value is calculated from the characteristic amount 30 by the maximum value calculating device 12, and the value is set as the evaluation amount 15. Produces similar results.

Further, although the feature amount 30 is not related to the defocus amount because the size of the processed data of the region including a singular point such as a high brightness subject does not change even if defocused, the feature amount 30 is destroyed. In the other areas, the original feature amount of the image is extracted. If the above maximum value calculation 1
In 2, when a region including a singular point is selected,
It will be extracted as an incorrect feature amount.

Therefore, by processing the sum of all or a plurality of the above feature quantities by the sum calculation processing unit 13, the feature quantity of the singular point is buried in the other feature quantities, and the evaluation quantity 16 is stored. The impact will be less.

Further, by the processing by the average value calculating device 14 for obtaining the average of the feature quantities, the influence on the evaluation quantity 17 of the feature quantity of the singular point is reduced as in the above. Further, by averaging, the output data length (the number of bits) of the evaluation amount 17 is reduced as compared with the total sum calculation process.

Although the maximum value calculating device 12, the sum total calculating device 13 and the average value calculating device 14 are shown in parallel in FIG. 3, it is sufficient that at least one of them is provided.

Further, in FIG. 3, the small area division processing device 11 is placed in front of the inter-pixel difference processing device 5 and the absolute value processing device 6, but this may be placed later or in between. The good thing is clear.

Example 4. FIG. 4 is a block diagram showing an example in which a block division processing device is used in the automatic focus adjustment device of the present invention. Image signal 4 output from the imaging device 3
Is subjected to small area division processing by the small area division processing device 11, and is processed by the inter-pixel difference processing device 5 and the absolute value processing device 6. Next, the value calculated by the first entropy calculation device 18 is set as the feature amount 23, and the small region selection processing device 19 performs the small region selection processing based on the feature amount 23.

The selected small area is divided into smaller blocks by the block division processing device 20. The processing data of the block-divided small area includes an entropy calculation device 28 and an evaluation amount calculation processing device 29. The value processed and output by the second entropy calculation device 21 is used as the evaluation amount 22.

The feature amount 23 output by the first entropy calculation device 18 is a value for selecting a small region in the next small region selection processing device 19.
The first entropy calculation device 18 extracts a feature amount for selecting a region suitable for calculating the evaluation amount 22 from the divided small regions. Therefore, a simple algorithm that can be processed at high speed is suitable.

The first entropy calculation device 18 calculates the maximum value, and the following small area selection processing device 1 is used.
In 9, the small area having the largest difference value between adjacent pixels is selected.

The small area selected by the small area selection processing device 19 is divided into smaller blocks, and the second entropy calculating device 21 next calculates the value for each block, so that all the images are selected. As a result, the evaluation amount can be calculated only for the characteristic part, and more complicated feature extraction processing can be efficiently executed.

The first entropy calculating device 1
Although the maximum value 8 is calculated, the same result can be obtained by calculating the difference between the variance value and the average value and the maximum value. Further, the number of data having a value equal to or larger than a certain threshold may be obtained.

The block division processing device 20 divides the small area into blocks of 5 × 5 pixels, for example, and the entropy calculation device 28 in the second entropy calculation device 21 calculates the variance value. is there. The variance value calculation process is highly accurate and effective as the entropy calculation process, but the calculation amount is large and the processing time is long. However, if the region is limited and the variance value calculation process is performed therein as in the present embodiment, it is possible to significantly reduce the time for calculating the evaluation amount.

When the evaluation amount is calculated in the small area,
If there are two or more subjects with different distances in the area, the features extracted from each subject will be synergized,
Even when looking at the evaluation amount in this area, it is not possible to perform the hill climbing control of the focus. However, by performing the block division with a block of 5 × 5 pixels and performing the entropy calculation process in the reduced area, the probability that two subjects are included in the block is reduced. The entropy calculation result of each block in the small area is calculated by the evaluation amount calculation device 29 to calculate the sum of a plurality or all of the entropy calculation results so that the evaluation amount indicates the peak of hill climbing at the focus position of each subject. Become.

When the photographer holds the image in the hand like a camera-integrated video tape recorder, the small area is first selected by the feature amount 23. That is, the so-called "swing" of the image pickup device 3 during the image pickup. It is effective against camera shake. That is,
In a system in which the evaluation amount is calculated only in 5 × 5 pixel block units and the defocus amount is determined based on the evaluation amount, the evaluation amount and the block from which it can be obtained change momentarily due to the hand shake as described above. Servo autofocus is not possible. On the other hand, if the evaluation amount is monitored in units of a small area larger than the 5 × 5 pixel block, when a shake with a small amplitude occurs, the image only moves in the small area and the evaluation amount changes. Because it does not, you can achieve hill climbing servo auto focus normally. Further, even in the case of an image that moves between small areas due to a large shake, in a system that enables high-speed calculation as in the above embodiment, compared to the main frequency component (10 Hz or less) of the shake,
Since the small area selection and the evaluation amount calculation can be performed sufficiently fast (1/30 Hz or less), the evaluation amount in the small area does not change abruptly and the mountain climbing can be performed normally.

The entropy calculation unit 28 in the second entropy calculation unit 21 in the above embodiment calculates the variance value, but this is the maximum value calculation or average value calculation as shown in the second and third embodiments. But it can be achieved.

Further, the evaluation amount calculation device 29 in the second entropy calculation device 21 in the above-mentioned embodiment is arranged to obtain the sum of a plurality of or all the values calculated for each block, but this is the average value. It may be calculated.

Example 5. FIG. 5 is a block diagram showing an example in which an orthogonal transformation device is used in the automatic focus adjustment device of the present invention. The image signal 4 output from the image pickup device 3 is divided into small regions by the small region division processing device 11, and orthogonal transformation processing is performed by the orthogonal transformation processing device 24. Next, the data calculated by the variance value calculation device 25 that calculates the degree of variation in the data of the small area subjected to the orthogonal transformation process is extracted as the evaluation value 27. The variance value calculation device 25 performs the maximum value calculation device or entropy calculation processing.

The small area division processing device divides into small areas of 8 × 8 pixels. The orthogonal transform processing device 2
4 is to extract the evaluation amount by paying attention to the frequency component of the image, whereas the above-mentioned inter-pixel difference processing has extracted the evaluation amount based on the inclination of the edge of the image.

Looking at the frequency components of the image, a complex image,
Alternatively, the focused image contains many high frequency components, and the variation of the conversion coefficient is large. On the other hand, a simple image or a blurred image does not contain high frequency components so much, the absolute value of the conversion coefficient is small, and the variation is small.

The orthogonal transform processing device 24 uses the DCT (Di
screte Cosine Transform (discrete cosine transform), and the dispersion value calculation device 25 processes the data output from the orthogonal transform processing device 24 excluding the value of the upper left DC component in the block. To do. 8 x 8
FIG. 6 shows the correspondence between the DCT coefficient and the frequency for the small area of the pixel. The definition of the DCT coefficient is as follows.

[0071]

[Equation 3]

The value output from the variance value calculating device 25 becomes the evaluation amount 27, but the evaluation amount 27 is small in the defocus state and gradually increases as the focus approaches, so The change characteristic with respect to the lens position in which the hill climbing servo described in Examples 1 to 4 is possible is shown.

When the above orthogonal transformation process is used, the feature quantity is extracted in the frequency domain, and therefore, it is less affected by the change in illuminance.

Since the variance value calculating device 25 calculates the degree of variation of the orthogonal transformation coefficient, any amount having a correlation with the variation may be used, and the block shown in FIG. For the frequency components other than the DC component on the upper left of the above, the sum of the coefficients having a value (or below) a certain threshold value or the number of coefficients may be output.

Further, FIG. 1 to FIG. 1 of the block diagram of the above embodiment.
Each processing device in FIG. 5 can be executed by high-speed hardware using a digital signal processor (DSP) or the like, but along the process flow of the block diagrams shown in FIGS. Software processing is also possible in the microcomputer. In this case, it is indispensable to utilize the high-speed calculation method described in Embodiments 2 to 4 in order to deal with a moving subject and camera shake.

[0076]

In the automatic focus adjusting device according to the first aspect of the present invention, the image signal output from the image pickup device is subjected to inter-pixel difference processing, the degree of variation in the output value is calculated, and the processed data is stored. Since the automatic focus adjustment is performed as the evaluation amount corresponding to the defocus amount, the configuration is simpler than the conventional one and the processing can be performed at high speed.

In the entropy calculating means according to the second aspect of the present invention, the maximum value detecting means is used to reduce the amount of data to be processed and to calculate the evaluation amount of automatic focus adjustment at high speed. There is an effect that a high-speed focusing operation can be performed even for a fast-moving subject or a sudden change in shooting conditions.

In the automatic focus adjusting device according to the third aspect of the present invention, the image signal from the image pickup device is divided into small areas by the area division processing, and the degree of data variation in the inter-pixel difference processing is calculated in each small area. As a result, there is an effect that the evaluation amount with respect to the defocus amount can be accurately extracted.

In the automatic focusing apparatus according to the fourth aspect of the present invention, the small area selection processing means selects a small area having a large feature amount, and the second entropy calculation processing means within the selected small area. Since the evaluation amount is calculated, there is an effect that the evaluation amount can be extracted at high speed and with high accuracy, and a stable evaluation amount can be obtained even when the image is shaken.

In the automatic focusing apparatus according to the fifth aspect of the present invention, the image signal is transformed into the frequency domain component by the orthogonal transformation means, the variation degree of the orthogonal transformation coefficient indicating the magnitude of the domain spectrum is calculated, and the By using the calculation result as the evaluation amount corresponding to the defocus amount, the configuration is simpler than in the past, and stable and high-speed processing can be performed even when the illuminance of the subject changes.

In the above embodiment, the inter-pixel difference processing means, the absolute value processing means, the entropy calculating means, the maximum value calculating means, the sum total calculating means, the small area selecting means, the block dividing means, the orthogonal transforming means and the variance value. Although the calculation means and the like are used as the automatic focus adjustment device, since they can calculate the complexity of the image, they can be used, for example, for the detection of the redundancy in the high efficiency coding system (image compression). Further, it is obvious that it can be used as an evaluation amount detection device that represents the characteristics of an image when adaptive processing such as image processing is performed.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example in which an inter-pixel difference processing device is used in an automatic focus processing device of the present invention.

FIG. 2 is a block diagram showing an example in which a maximum value calculation device device is used in the automatic focus processing device of the present invention.

FIG. 3 is a block diagram showing an example in which a region division processing device is used in the automatic focus processing device of the present invention.

FIG. 4 is a block diagram showing an example in which a region selection processing device is used in the automatic focus processing device of the present invention.

FIG. 5 is a block diagram showing an example in which an orthogonal transform processing device is used in the automatic focus processing device of the present invention.

FIG. 6 is a diagram showing a relationship between orthogonal transform coefficients and frequencies.

FIG. 7 is a block diagram showing a conventional autofocus device.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Optical system lens 2 Photoelectric converter 3 Imaging device 4 Image signal 5 Inter-pixel difference processing device 6 Absolute value processing device 7 Entropy calculation device 9, 12 Maximum value calculation device 11 Small area division processing device 13 Summation calculation device 14 Average value calculation Device 18 First entropy calculation device 19 Small area selection processing device 20 Block division processing device 21 Second entropy calculation device 23 Feature amount 24 Orthogonal transformation processing device 25 Variance value calculation device 8, 10, 15, 16, 17, 22 , 26, 27 Evaluation amount

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] July 14, 1992

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be corrected] Claim 1

[Correction method] Change

[Correction content]

[Procedure Amendment 2]

[Document name to be amended] Statement

[Name of item to be corrected] Claim 2

[Correction method] Change

[Correction content]

[Procedure 3]

[Document name to be amended] Statement

[Correction target item name] 0017

[Correction method] Change

[Correction content]

[0017]

An automatic focusing device according to claim 1 of the present invention According to an aspect of the pixel difference processing means for calculating a pixel difference of the image signal from the imaging device, variation in the value of the output of its The entropy calculation means for calculating the value is output, and the value output from the entropy calculation means is extracted as an autofocus evaluation amount.

[Procedure amendment 4]

[Document name to be amended] Statement

[Correction target item name] 0031

[Correction method] Change

[Correction content]

The difference data between adjacent pixels can be calculated from the formula (1) as
It is symmetrical about the origin with respect to the degree x. Moreover, since it requires only a change amount in future processing, the absolute value processing unit 6 performs absolute value processing. However, this absolute value processing is
Is not mandatory in
It is also possible to remove it because it can be processed as it is.
Yes.

Claims (5)

[Claims] 1. An image pickup optical system, and an image pickup apparatus for obtaining an image signal from a photoelectric converter, an inter-pixel difference processing means for calculating a difference between pixels of an image signal from the image pickup apparatus,
The absolute value processing means for calculating the absolute value of the output of the inter-pixel difference processing means and the entropy calculation means for calculating the degree of dispersion of the output values of the absolute value processing means are provided, and the output of the entropy calculation means is used as the evaluation amount. An automatic focus adjusting device characterized by controlling an optical system for automatic focus adjustment based on it. 2. The entropy calculating means calculates the largest value among the output values after the absolute value processing, or selects a plurality of values from the largest value,
The automatic focus adjusting device according to claim 1, further comprising maximum value calculating means for extracting the value as an evaluation amount. 3. The image signal obtained from the image pickup device is divided into a plurality of small areas in an image feature extraction area for performing the entropy calculation processing, and the first entropy calculation processing is performed in each of the small areas. The output is used as a feature amount, and a maximum value calculating means for calculating the maximum value among the feature amounts, a sum calculation processing means for adding a plurality of the feature amounts, or an average value of the feature amounts is calculated. An automatic focus adjusting device comprising any one of average value calculating means, wherein the value calculated by each means is used as an evaluation amount to control an optical system for automatic focus adjustment based on the evaluation amount. 4. After the first entropy calculation processing, small area selection processing means for selecting one or a plurality of small areas having a large value in the feature amount, and further area for the selected small areas. Block division processing means for subdividing,
The automatic focus adjustment device according to claim 3, wherein the second entropy calculation processing means is performed in each block, and the output of the second entropy calculation processing means is used as an evaluation amount. 5. In the image pickup apparatus for obtaining an image signal from the optical system for image pickup and the photoelectric converter, the image feature which is an area for performing image processing from the image signal obtained from the image pickup apparatus. The image signal of the extraction area is extracted, and the small area division processing means for dividing the image feature extraction area into a plurality of small areas, the orthogonal transformation means for performing orthogonal transformation in the small area, and the orthogonal transformation are obtained. An automatic focusing system comprising a dispersion value calculating means for calculating the degree of dispersion of the AC component of the conversion coefficient, and controlling the optical system for automatic focus adjustment based on the output of the dispersion value calculating means as an evaluation amount. Adjustment device.