JPS63169879A - Automatic focus system - Google Patents

Abstract

PURPOSE:To allow data rewriting in a ROM to cope with remarkable changes in the functions and kinds of lenses by applying digital processing to a video image taken directly from the lens to improve the follow-up property. CONSTITUTION:A video signal picked up by an image pickup element 4 is subjected to analog/digital conversion by an A/D converter 6 and stored in a pattern memory 7(RAM) as video image data having gradation and the contrast of the image at the center and the circumference of the screen, that is, the contour is detected. Thus, contour values S1, S2 and the focal length la of the lens are obtained in this way and the stop position of the lens at the initial state is detected to ensure focusing the lens to an object. Moreover, if an objective person walks toward the video camera, the focal length lb of the lens to the object is estimated from the relation of equation between S1-S2 and the stored lens focal length la to attain lens focusing.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an autofocus control method for Pink Camera 2, and in particular, by determining the outline of the subject from the captured video signal, the difference between the background and the subject is determined. This relates to an autofocus method that estimates distance.

[Conventional technology]

Conventionally, this type of autofocus control system has a shaft that has an emitting part that emits infrared rays, a receiving part that receives infrared rays reflected from the object, and is linked to the rotation of the focus adjustment ring, that is, the movement of the distance of the objective lens. Alternatively, a method has been used in which the distance between the emitting section and the light receiving section is adjusted through a mechanism such as a gear, the distance of the object to be photographed is determined, and the lens is focused.

[Problem that the invention seeks to solve]

In the infrared autofocus method described above, which uses infrared distance measurement to focus the lens, if there is a transparent material such as glass between the object to be photographed and the video camera, the infrared rays will be reflected off the glass surface. The light receiving section detects the reflected infrared rays, and if the lens is focused on the glass surface, then if the object is a black body, the emitted infrared rays will be absorbed by the black body and will not come back. However, there were other problems, such as the complexity of the interlocking mechanism between the emitting part and the light receiving part, which are aligned with the focal point of the lens.

[Means for solving problems]

The autofocus method of the present invention converts the captured video signal from analog to digital in a video camera, extracts the center and several peripheral points of the image, detects the outline, and converts the image between the peripheral and central points. This system compares the average value of the image differential, which is the outline of It is.

[Effect]

In the present invention, images directly captured through a lens are digitally processed.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail based on the drawings.

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

The lens used in the autofocus method according to the present invention uses a combination lens, and the front lens is the lens that adjusts the focal length.
In this method, the focal point of the combined lens is changed by rotating the lens barrel 1 (hereinafter referred to as the lens barrel). A gear is cut in this lens barrel 1, and a focusing drive motor 3 rotates to move the lens back and forth. Tsuma υ, focus changes. Similarly, a position meter 2 is attached to the lens barrel 1, and a position meter 2 is attached to the lens barrel 1.
As the microprocessor 8 (CPU) (DI
It is configured to output digitized numerical values to 10 ports.

Reference numeral 4 denotes an image sensor that converts the input image from the lens barrel 1 from an optical signal to an electrical signal, 5 an image sensor that separates the output of this image sensor 4 into a luminance signal and a color signal, and 6 an image sensor of this image sensor 5. Analog-to-digital converter (hereinafter referred to as
7 is a screen memory (RAM) which inputs the output of this A/D converter 6, 9 is a ROM
It is.

Then, the captured video signal is converted from analog to digital by the A/D converter 6, and the points at the center and several peripheral parts of the image are extracted to detect the outline, and the outlines of the peripheral and central parts are extracted. It is a value. It is configured to compare the average value of the image differential and work in conjunction with the focusing lens of the camera lens to estimate the focal length to the object, thereby improving the followability of the camera focus to the object.

Next, the operation of the embodiment shown in FIG. 1 will be explained with reference to FIGS. 2, 3, and 4.

FIG. 2 is an explanatory diagram showing the scenery on the screen, where 11 shows a tree in the foreground, and 12 shows a person as an object. Figure 3 is an explanatory diagram showing the arrangement of contour detection pixels on the screen.
2・-13-5 is the contour detection pixel (address in the image memory)
, and 81*l)l·-S represents the contour value. FIG. 4 shows a graph of lens focal length L (horizontal axis) and contour difference (vertical axis).

First, in order to facilitate understanding of the present invention, its operating principle will be explained.

In the present invention, a video signal photographed by an image sensor 4 is converted from analog to digital by an A/D converter 6, and a screen memo I) 7 is created.
After storing (RAM) as video data with gradation,
The contrast of the image between the center and the periphery of the screen, that is, the outline is detected. Specifically, data corresponding to four arbitrary adjacent pixels "p*q*rt" are read out from the screen memory T and a solution of R.b is determined.

Then, the video data in the center of the screen is transferred to the Robert
The contrast of the image in the center is determined by differentiating the image using the operator equation of s, and the contrast of the peripheral areas on both sides is determined using the same procedure.

Now, assuming an image taken using perspective, if the person in question is located far away from the periphery of the screen, and the video camera's focus is at the nearest position, then the entire screen will be in focus. Since the contrast of the image is low (i.e. out of focus), the focus motor is rotated to lengthen the focal point of the lens, which extends the focal point of the lens.

That way, the differential value of the pixels on the entire screen will be high.
The contour is extracted. In this state, the contrast of the person in the center between the frames becomes high first.

This is because, in the perspective screen assumed, the person is located at a farther distance than the foreground in the periphery. Then, the lens focal length when the contrast of the foreground in the peripheral area is maximum is 1. and the contrast value, i.e.
The contour value 81 of the foreground (see FIG. 3) and the contour value S of the person in the center of the screen at that time.

(See FIG. 3) are interrelated, and the distance tb of the lens focal point to the object is determined by the relational expression between the lens focal length t&, which is the distance of the nearby foreground, and 5l-s. Tsumashi, tb
−t, is proportional to S, −S.

In this way, the contour value Sl+81 and the lens focal length 1. By determining the initial position of the lens and detecting the initial stop position of the lens, it is possible to reliably focus the lens on the object. Furthermore, when the target object walks toward the video camera, the contour value at the center of the screen is omitted and the contour value decreases, so the time is 5l-s as well.

The value of and the stored lens focal length 1. The distance tb of the lens focal point to the object is estimated from the relational expression and the lens is focused. In this case, the previous distance of the object tb
Based on the difference between the distance tbl and the new distance tbl, the rotational direction of the lens focusing drive motor 3, that is, the focus motor, can be detected by aligning it with the moving direction of the object. Furthermore, when the object moves from the person to the foreground, the distance Lb from the previous person
Since the contour value st j of the foreground and the contour value 8m when the person moves to the periphery of the screen can be detected at the same time, the lens focal length t1 can be estimated.

In this way, if the parameters of the differential of one pixel, that is, the contour value Sn and the focal length Ln of the lens, are programmed in advance as data in the ROM of the central processing unit for control, the contour value can be obtained, that is, the screen data Autofocus control of the video camera becomes possible by differentiating and averaging the values and detecting the position of the focusing lens.

Next, the above operation will be explained in more detail.

The image input through the lens barrel 1 is converted from an optical signal to an electrical signal by the image sensor 4, and the electrical signal is sent to the image circuit 5.
The signal is then separated into a luminance signal and a chrominance signal. This luminance signal is then digitized by the A/D converter 6 as data having gradation, and output to the RAM which is the screen memory 7. Then, the microprocessor 8 uses the Rob@rts operator formula described above to calculate the contour values of adjacent pixels, and stores the results in the screen memo IJ7.

The ROM 9 stores software for performing this control, approximate values of contour values, data for time control of the focusing drive motor 3, and parameter memory of the position of the position meter 2 and the contour difference.

Now, as shown in FIG. 2, when the scenery on the screen is photographed, the foreground tree 11 and the object person 12 are contoured according to the explanatory diagram showing the arrangement of contour detection pixels on the screen shown in FIG. Tree 11 in the foreground was detected at a value of 81m8m, and person 12
is detected at the pixel with the contour value SS. And the contour value S4
+81 will show the background of the person.

If the lens barrel 1 is now located at the frontmost position, the position meter 2 and the lens focal position are informed to the microprocessor 8, υ, the contour value st is 13g, and the differential value of Sm is Sl or 8m>8s. It can be seen that the condition is As a result, the microprocessor 8 outputs a pulse in the forward direction of the lens cylinder 1 to the lens start point alignment drive motor 3, and continues to drive the focus drive motor until the contour value 81, the differential value of the contour value 81t, that is, the contrast reaches the maximum. Issue a rotation command to 3.

And the contour value Sit or S! Since the position meter 2 value is always being traced when the differential value reaches the maximum and falls, the peak of the differential value of the contour value St or sm and the differential value of the contour value S3 of the person 12, which is the object, are The difference is taken and the value of position meter 2, i.e. distance L(
(See Figure 4) and 8l-8s υTarget person 1
Distance from 2 is 1i! L L intention can be inferred.

FIG. 4 shows a graph of the relationship between lens focal length and contour difference (81-81). As is clear from FIG. 4, the focal length t! of the tree 11 in the foreground (see FIG. 2)! If the contour value S1 or the contour difference between Sl and ag is obtained, the focus of the person 12 (see FIG. 2) is estimated, the time for energizing the focusing drive motor 3 is calculated, and the control is performed by the I10 boat 1G. Ru.

Although the present invention has been described in detail above, the scope of the present invention is not limited to this embodiment.

〔Effect of the invention〕

As explained above, according to the present invention, by digitally processing an image directly taken in through a lens, it is possible to eliminate the need for a mechanical part that moves the light emitting part in conjunction with the lens barrel, and for precise assembly. It is possible to configure an autofocus mechanism with good tracking performance without having to adjust the spacing between the light emitting part and the light receiving part, or considering variations in the light receiving and light emitting elements. Because it can be made to depend on the microcomputer software, even if the type or function of the lens changes significantly, the R
This can be dealt with by rewriting the data in the OM, so the practical effect is extremely large.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIGS. 2, 3, and 4 are explanatory diagrams showing scenes on the screen to explain the operation of FIG. 1, and outline detection pixels on the screen. FIG. 2 is an explanatory diagram showing the arrangement of the lens and a graph of lens focal length and contour difference. 1. Lens barrel, 2. Position meter,
3. Fist: Focusing drive motor, 4. Fist Φ fist image sensor,
5...Fist/image circuit, 6@...A/D converter, 7*Fist/screen memory, 8...e microprocessor, 9.
...ROM, 10-...I10 port, 11・Fist・
・Tree in the foreground, 12...Object person, 13-1~1
3-5...Contour detection pixel.

Claims (1)

[Claims] In a video camera, the captured video signal is converted from analog to digital, the center of the image and several peripheral points are extracted to detect the contour, and the average of the image differentials, which are the contours of the peripheral and central areas, is detected. This autofocus method is characterized by comparing the values and estimating the focal length of the object by linking with the focusing lens of the camera lens, thereby improving the tracking ability of the camera focus on the object.