JPH08160283A - Automatic focusing device - Google Patents

Abstract

PURPOSE: To obtain a focusing point with high accuracy regardless of the contents of a picked-up image. CONSTITUTION: A video signal picked up by an image picture device through a focusing lens 1 becomes the video signal of television by means of a processing circuit 3 and supplied to an image feature extracting part 20. Two video signals of the current frame and one frame before are obtained by means of a delay circuit 24 through a filter circuit 22. An image correlation detecting circuit 26 divides the images of two video signals into plural blocks and obtains the distribution of the sum of frame difference absolute values of signals of two corresponding blocks. Form this distribution, an image feature extracting circuit 28 imparts position information 8 indicating a region containing a few of high frequency components to a circuit 4, makes the video signal pass through and adjusts the lens 1 so as to maximize the high frequency components by means of a BPF circuit 5, detecting circuit 6 and a motor 7. A part such as the sky and a lateral stripe pattern is removed and focus adjustment is accurately performed.

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 mounted on a camera such as a video camera.

[0002]

2. Description of the Related Art As an automatic focus adjusting device used in video equipment such as a video camera, a high frequency component is extracted from a video signal obtained from an image pickup device such as a CCD so that the high frequency component becomes maximum. A so-called hill-climbing system is known in which a photographing lens is driven to perform focusing. That is, the video signal level changes sharply at the edge of the subject image, and the high frequency component of the video signal increases,
It can be considered that the closer to the focused state, the closer the high frequency component increases. Such an automatic focus adjustment method does not require a special optical member for focus adjustment, and has an advantage that it can accurately focus regardless of the distance regardless of the distance.

Next, an automatic focusing system of this kind will be described with reference to FIG.
Will be explained. The focus lens 1 performs focusing by moving in the optical axis direction by the focus motor 7. The light passing through the lens 1 is imaged on the image pickup surface of the image pickup device 2, photoelectrically converted into an electric signal, and output as a video signal. This video signal is input to the process circuit 3 of the camera and converted into a video signal of a standard television system, and a band pass filter (hereinafter referred to as BPF).
And the high frequency component in the video signal is extracted. This high frequency component is input to the detection circuit 6 and a signal corresponding to the absolute amount of the high frequency component is extracted.

The extracted signal of the detection circuit 6 is shown in FIG.
As shown in, the maximum value is obtained at a certain position, which varies depending on the amount of extension of the focus lens 1. This is because when the amount of extension of the focus lens 1 changes, the focus state of the optical image projected on the image pickup surface of the image pickup element 2 changes, and this is the in-focus state, that is, the projected optical image when there is so-called focus. Sharpness becomes maximum, and when converted into a video signal by the image sensor 2, the absolute amount of high-frequency components contained therein becomes maximum. Next, the focus motor 7 moves the position of the focus lens 1 in such a direction that the output signal of the detection circuit 6 becomes large, and stops the focus lens 1 at the position where the output of the detection circuit 6 becomes maximum. To do.

[0005]

However, in the conventional example, there is a problem in that the output level of the output signal of the detection circuit 6 changes depending on the type and contrast of the subject, and the focus cannot be accurately controlled depending on the subject. there were. That is, in the case of the conventional example, as shown in FIG. 3A, a normal subject has less noise and the output signal level of the detection circuit 6 is high, so that the focus motor 7 can be controlled easily, but When the sky is reflected in a large area, as shown in FIG. 3B, the high frequency component of the video signal is reduced and the noise component is increased.
Since the output signal level of the detection circuit 6 becomes low and the peripheral portion shows a characteristic that is raised by noise, and therefore the position of the peak becomes unclear, hunting may be repeated in the vicinity of the focus.

The present invention has been made in order to solve the above-mentioned problems, and provides a quick focus without hunting even when photographing a low-contrast subject such as the sky or a subject at low illuminance. It is an object of the present invention to provide an automatic focus adjustment device that can be adjusted.

[0007]

According to a first aspect of the invention, an image feature extracting means for extracting features of an image of a video signal obtained by imaging through a focus lens, and the image feature extracting means for extracting the features. And a focus adjusting means for adjusting the focus lens to a focal point according to the characteristics of the image.

According to the second aspect of the present invention, the image characteristic means for detecting a region having a small amount of high frequency components in the image of the video signal and outputting the position information of the region, and the signal of the region indicated by the position information from the video signal. A focus adjusting means is provided for detecting a high frequency component from the signal from which is removed, and controlling the position of the focus lens so that the high frequency component has a maximum value.

[0009]

According to the first aspect of the present invention, the feature of the imaged image is extracted, and the focus adjustment of the focus lens is controlled according to the feature, so that appropriate adjustment is always performed according to the content of the image. The focus is obtained accurately.

According to the second aspect of the present invention, a region of the picked-up image having a small amount of high frequency components is detected, and the high frequency component of the video signal from which the signal in the region is removed is detected so that the high frequency component becomes maximum. Since the focus lens is adjusted to, the part with a low high frequency component such as the sky of the subject is excluded from the focus adjustment target, and the focus is adjusted only on the part with a high frequency component. The focus can always be obtained with high precision, and is optimal for focus adjustment according to the hill climbing direction.

[0011]

【Example】

[First Embodiment] FIG. 1 is a block diagram showing a first embodiment of the present invention. In FIG. 1, the focus lens 1
Is moved in the optical axis direction by the motor 7 for driving the focus lens to perform focusing. The light passing through the lens 1 is imaged on the image pickup surface of the image pickup device 2, photoelectrically converted into an electric signal, and output as a video signal.
This video signal is input to the process circuit 3 of the camera,
The video signal is converted into a video signal of a standard television system, and is input to a bandpass filter (hereinafter referred to as BPF) 5 via the gate circuit 4. On the other hand, the video signal of the standard television system is input to the image feature extraction unit 20,
The position information 8 indicating an area unsuitable for focus adjustment included in the video signal is output and input to the gate circuit 4 as a gate signal. The gate circuit 4, the BPF circuit 5, the detection circuit 6 and the motor 7 constitute a focus adjusting means.

Next, the operation of the image feature extraction section 20 for extracting the feature of the image from the input video signal will be described. The filter circuit 22 performs a spatial BPF process of removing a DC component and an unnecessary high frequency component from the input video signal in order to prevent a malfunction due to a change in luminance of the input video signal and reduce an influence of noise. In order to reduce the information amount of the signal, the spatial BPF processed signal is
It may be valued. The video signal subjected to the spatial BPF processing is directly input to the image correlation detection circuit 26 and is delayed by one frame (or one field) by the delay means. The delayed video signal is output to the image correlation detection circuit 26.
Is input to

The image correlation detection circuit 26 divides the images of the two input video signals into a plurality of blocks as shown in FIG. 2A, and in each block, pixel data of the previous frame and pixel data of the current frame are divided. And the difference absolute value sum is calculated, and the information of the distribution of the difference absolute value sum is output as the correlation value distribution. The image feature extraction circuit 28 obtains the feature of the input video signal from the distribution information of the correlation value which is the output signal of the image correlation detection circuit 26.

For example, the distribution of the correlation values of the blocks of the input video signal when the subject is empty is a distribution without peaks as shown in FIG. 2 (c). On the other hand, when the subject is a mountain or a house, the distribution of the correlation value of the block of the video signal is shown in FIG.
The distribution has a sharp peak as shown in (b). Therefore, by detecting whether or not there is a sharp peak in the distribution of the correlation value, which is the output signal of the image correlation detection circuit 26, and finding the distribution of the blocks having no peak, it is possible to know at which position of the subject the sky is present. Whether there is a sharp peak in the distribution of correlation values, the maximum value of the distribution, the difference between the maximum value and the average value,
Alternatively, it can be determined by obtaining a difference between the maximum value and a value near the maximum value.

Further, subjects such as window blinds and patterns with horizontal stripes are subjects that are difficult to focus on. In other words, it is difficult to construct a vertical bandpass filter in the current automatic focus adjustment device due to cost problems (increase in the number of parts), and the focusing signal is output only from the output of the horizontal bandpass filter. Is getting Therefore, since there is no edge in the horizontal direction in a subject such as a window blind or a pattern with horizontal stripes, the level of the output signal of the bandpass filter is low and focus adjustment becomes difficult. However, even in an object such as a blind or a pattern with horizontal stripes on the window, the output signal of the image correlation detection circuit 26 has a distribution in which peaks are connected in the horizontal direction, and thus such a distribution appears. By obtaining the distribution of blocks, it is possible to know the position where the subject is not good.

In this way, the position of a region having a small amount of high frequency components necessary for focus detection such as an edge included in a video signal such as a sky or a portion having a horizontal stripe pattern is detected, and the position information 8 is gated. The signal is output to the gate circuit 4 as shown in FIG. The gate circuit 4 controls whether to pass or block the video signal from the image sensor 2 according to the input position information 8. That is, the video signal in a region having a small amount of high frequency components such as a sky or a portion having a horizontal stripe pattern is blocked, and only the video signal in a region having a large amount of high frequency components is BP.
Output to F5. The BPF 5 extracts the high frequency component in the video signal and outputs it to the detection circuit 9. The detection circuit 6 outputs a signal corresponding to the absolute amount of the high frequency component to the motor 7 for driving the focus lens. The motor 7 moves the position of the focus lens 1 in a direction in which the output signal of the detection circuit 6 becomes large, and stops the focus lens 1 at the position where the output of the detection circuit 6 becomes maximum.

The state at this time will be described with reference to FIG. In the case of a normal subject, since the high frequency component contained in the input video signal is large, the relationship between the extension amount of the focus lens 1 and the output of the detection circuit 6 is as shown in FIG. At this time, since the peak has a sharp shape, the position of the maximum value of the peak is clear and focus adjustment becomes easy. Further, when a lot of sky is reflected in the input video signal, the relationship between the extension amount of the focus lens 1 and the output of the detection circuit 6 is as shown in FIG. Since there are many, the detection output of the peripheral part has risen. At the same time, since the high frequency component is small even in the focused portion, the detection output level remains low. Therefore, the peripheral portion is lifted up and the peak is lowered, so that the focus position is not clear and the focus adjustment becomes difficult.

According to this embodiment, the position information 8 indicating the position where the sky is detected, which is detected by the screen feature extraction unit 20, is shown by the oblique line in which the sky included in the input video signal is reflected as shown in FIG. By cutting (cutting) the portion to remove noise in the empty portion, the level of the peripheral portion of the detection output is lowered as shown in FIG. 3 (c). As a result, the peak of the detection output becomes sharp and focus adjustment becomes easy.

[Second Embodiment] In the first embodiment described above, the data of the representative point of the previous frame or the pixel data of the previous frame and the pixel data corresponding to each of the regions of the current frame are provided for each area. The feature of the input video signal in each area is extracted from the distribution of the sum of the frame difference absolute values of and, and the feature of the input video signal in each region is also extracted by the method described below. Can be extracted. FIG. 5 shows a second embodiment.
In FIG. 5, 30 is an image feature extraction unit, 32 is a two-dimensional bandpass filter for extracting a predetermined frequency component,
Reference numeral 34 is a binarization circuit for binarizing the output signal of the two-dimensional bandpass filter 32, 36 is a block division circuit for dividing the binarized signal into a predetermined region of N pixels × M pixels, Reference numeral 38 denotes a noise detection circuit that detects the amount of noise included in each of the divided areas.
The other portions correspond to the same reference numerals in FIG.

Next, the operation will be described. The input signal from the process circuit 3 is input to the image feature extraction unit 30,
A predetermined frequency component is extracted by the two-dimensional bandpass filter 32 in order to remove a change in the brightness signal due to a change in brightness and a high frequency component. In the binarization circuit 34, D
Binarization processing is performed on the video signal from which the C component and the high frequency component have been removed, at the point where it crosses the 0 level. By performing such processing, the binary image pattern does not change even if the signal strength of the input video signal changes. Next, in the block division circuit 36, the binarized image pattern is divided into regions of N pixels × M pixels, and the data of the binarized image pattern in that region is supplied to the noise amount detection circuit 38 for each region. Output.

The noise amount detection circuit 38 counts the number of times the pixels of the binarized image pattern change from 0 level to 1 level and the number of changes from 1 level to 0 level, and compares the number of times with a predetermined value. However, if it is larger than the specified value,
The area is judged to be an unsuitable area because there is a lot of noise, and when it is smaller than a predetermined value, the area is judged to be an appropriate area. In order to count the number of changes of the binarized image pattern, a counter having a signal changing from 0 level to 1 level as a clock may be provided in the noise amount detection circuit 38, that is, scanning in the horizontal and vertical directions respectively. Then, the output value of the counter in each direction is compared with a predetermined value.

In the case of a normal subject, as the binary image pattern, as shown in FIG. 6A, a 0 level pattern and a 1 level pattern are clearly output, so that the horizontal level of the counter in the noise amount detection circuit 38 is increased. Direction output value is 1
Assuming that the number of times the level is changed from 0 level to 1 level by scanning a line is kh, a value close to a value of (kh × N) is output. In FIG. 6A, kh is about 3. Also, the vertical output value of the counter is kv, where the number of changes from 0 level to 1 level in one line scan is kv
A value close to the value of (kv × M) is output.

On the other hand, in an area where a subject such as the sky is reflected, there are many noise components as shown in FIG. 6B, so the horizontal output value of the counter in the noise amount detection circuit 36 is (kh × N). A value greater than the value of is output, and
The output value in the vertical direction of the counter is also larger than the value of (kv × M). Therefore, for example, the output value of the counter is 1.2 × (kh × N), 1.2 × (kv × M),
Alternatively, it can be determined whether or not there is a sky in the area based on whether or not it is larger than the total value. Position information of the subject that is difficult to perform automatic focus adjustment such as the sky 8
Is input to the gate circuit 4, and the portion of the subject which is not good is removed from the signal for performing the automatic focus adjustment.

[Third Embodiment] In the second embodiment,
By providing the coring means before the binarization circuit 34, the configuration of the noise amount detection circuit 38 can be simplified. Figure 7
The third embodiment is shown in FIG. In FIG. 7, reference numeral 31 is an image feature extraction unit, and 33 is a binarization circuit 34 that removes a minute noise component from the output signal of the two-dimensional bandpass filter 32.
It is a coring processing circuit for outputting to. The other portions correspond to the same reference numerals in FIG.

The operation is the same as in the second embodiment, except that
The output signal of the two-dimensional band-pass filter 32 is subjected to coring processing for removing minute noise by the coring processing circuit 33, and then binarized. Therefore, the influence of noise on the binarized pattern can be eliminated. That is, as shown in FIG. 8B, since the specific frequency component is small in a subject in which the sky is reflected, the binarization pattern is 0.
There are many levels or one level states.

Therefore, in the case of an ordinary subject as shown in FIG. 8A, the horizontal output value of the counter in the noise amount detection circuit 38 is a value close to a value of (kh × N). . Further, the output amount of the counter in the vertical direction is a value close to a value of (kh × M). On the other hand, in an area where a subject such as the sky is shown, the horizontal output value of the counter in the noise amount detection circuit 36 becomes a value close to 0 due to the coring process. Also, the output value of the counter in the vertical direction is close to 0. Therefore, for example, the main value of the counter is 0.5 × (kh × N),
Whether or not the sky is R can be determined by 0.5 × (kv × M), or whether it is smaller than the total value.

[0027]

As described above, according to the invention of claim 1, the feature of the picked-up image is extracted, and the focus adjustment of the focus lens is controlled according to the feature. Regardless of the content of the image, appropriate adjustment is always performed, and there is an effect that the in-focus point can be accurately obtained.

According to the second aspect of the present invention, a region of the picked-up image having a small amount of high frequency components is detected, and the high frequency component of the video signal from which the signal in the region is removed is detected. Since the focus lens is adjusted so as to satisfy the above condition, there is an effect that even an object such as a sky or horizontal stripes, which is poor in focus adjustment, can be focused with good brightness. Especially, it is most suitable for the focus adjustment by the mountain climbing method.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is an explanatory diagram illustrating a difference in distribution of correlation values depending on a subject.

FIG. 3 is a characteristic diagram illustrating a relationship between a lens extension amount and a detection output.

FIG. 4 is an explanatory diagram illustrating an operation of the gate circuit of FIG.

FIG. 5 is a block diagram showing a second embodiment of the present invention.

FIG. 6 is an explanatory diagram illustrating a state of a binarized pattern.

FIG. 7 is a block diagram showing a third embodiment of the present invention.

FIG. 8 is an explanatory diagram illustrating a state of a binarized pattern according to a third embodiment.

FIG. 9 is a block diagram showing a conventional automatic focus adjustment device.

[Explanation of symbols]

 1 Focus Lens 2 Image Sensor 4 Gate Circuit 5 Band Pass Filter (BPF) 6 Detection Circuit 7 Focus Lens Driving Motor 20, 30, 31 Image Feature Extraction Section 24 Delay Circuit 26 Image Correlation Detection Circuit 28 Image Feature Extraction Circuit 33 Coring Processing circuit 34 Binarization circuit 36 Block division circuit 38 Noise amount detection circuit

Claims (8)

[Claims] 1. An image feature extracting means for extracting image features of an image signal obtained by being imaged through a focus lens, and the focus lens according to the image features extracted by the image feature extracting means. An automatic focus adjusting device having a focus adjusting means for adjusting the focus. 2. The image feature extracting means is configured to detect an area having a low frequency component in an image and output position information of the area, and the focus adjusting means extracts the position information from a video signal. The automatic focusing apparatus according to claim 1, wherein a high-frequency component of a signal obtained by removing a signal in the indicated area is detected, and the position of the focus lens is controlled so that the high-frequency component has a maximum value. 3. The image feature extracting means divides two images of a video signal of the current frame or current field and a video signal of one frame before or one field before into a plurality of regions, respectively, and the two images correspond to each other. 3. The automatic focus adjustment device according to claim 1, wherein the automatic focus adjustment device is configured to obtain a sum of absolute differences between signals in two areas and generate position information based on a distribution of the absolute differences. 4. The image feature extracting means outputs the video signal as 2
The image is divided into a plurality of regions after the binarization, the noise amount of the signal for each region is detected, and the position information indicating the region having a small noise amount is output. Auto focus device. 5. The automatic focusing apparatus according to claim 4, wherein the image feature extracting means is configured to perform binarization after removing minute noise of the video signal. 6. The image feature extracting means includes a maximum value in the distribution of the difference absolute values, a value of a difference between the maximum value and an average value,
The value of the difference between the maximum value and the value near the maximum value, or the maximum value,
The automatic focusing apparatus according to claim 3, wherein the position information is detected by calculating from at least one of a minimum value and an average value. 7. The screen feature extracting means is configured to obtain position information based on a distribution of 0 level and 1 level for each region in the binarized signal. The described automatic focus adjustment device. 8. The position information indicating the area in which the number of times that the binarized signal changes from 0 level to 1 level or from 1 level to 0 level is larger or smaller than a predetermined threshold value by the screen feature extraction means. 6. The automatic focusing device according to claim 4, wherein the automatic focusing device is configured to output.