JPH06308371A - Line-of-sight detecting device - Google Patents

Abstract

PURPOSE:To eliminate an erroneous line of sight signal and to realize operation according to the more accurate line of sight signal by judging the blink of a photographer which is a noise signal at high speed by simple constitution in the case of performing the operation of a camera by using the line of-sight signal. CONSTITUTION:When a light projection part 1 projects light toward the eyeball 90 of the photographer, a light receiving part 2 detects reflected light from the eyeball 90 of the photographer, and a control part 3 controls the projection of light by the light projection part 1 and the detection timing of the light receiving part 2. A light quantity decision part 4 detects the blink of the photographer by deciding light quantity in the same area as the light receiving part 2, and a line of sight detection part 5 detects the line of sight based on a signal from the light receiving part 2. Furthermore, the detection part 5 detects the line of sight based on a value from the decision part 4. In the case the decision part 4 detects the blink of the photographer, the detected result on the line of sight by the detection part 5 is invalidated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for detecting a line-of-sight direction of a photographer, and more particularly to a line-of-sight detection for setting camera information such as focus area and strobe on / off based on a signal of the line-of-sight direction. Regarding the device.

[0002]

2. Description of the Related Art Conventionally, various kinds of information are input to a camera by using dials, buttons, etc., and the operating environment becomes complicated as the input information increases. Therefore, for example, a technique of detecting the line-of-sight direction of a photographer looking through a finder and inputting information to the camera based on the information about the line-of-sight direction is disclosed in Japanese Patent Laid-Open No. 63-194237 and Japanese Patent Laid-Open No. 3-87818.
It is proposed by Japanese Patent Laid-Open No. 4-307506.

Further, a technique relating to a method of detecting the line-of-sight direction, etc. has been proposed in Japanese Patent Application Laid-Open No. 2-206425. Then, by detecting the line of sight of the photographer in this way,
When the camera is operated with the detected line-of-sight signal, it is important to judge whether the eye is opened or closed (blinking), which is a physiological phenomenon of human beings. In view of this point, a technique that detects the blinking of the photographer and incorporates the information into the control is disclosed in Japanese Utility Model Publication 63-.
174804, JP-A-03-39135, JP-A-2-57433 and the like. In these techniques, the line-of-sight detection device of a camera detects whether the photographer blinks or not by the brightness of the detected image, that is, the sum of the detected sensors or the total current value.

By the way, in the current blink detection, it is necessary to extract eye feature points (corneal reflected light, light collecting area), and for this purpose, a high resolution sensor such as a two-dimensional CCD is required. . In addition, it is essential to use an inexpensive PSD or PD or a CCD sensor with a rough resolution for simple line-of-sight detection. In this case, it is difficult to determine the blinking using the eye feature points as described above due to lack of resolution. Further, when blinking during detection, the amount of reflected light of the light emitting LED tends to be considerably brighter than that of light reflected from a regular eye, or conversely, it tends to be darker due to the color of the skin (eyelids).

[0005]

As described above, in order to determine the blinking of the photographer, in order to extract the feature points of the eye, for example, confirmation of the reflected light of the cornea and the amount of change in the area of the iris portion are performed. Although the determination is performed based on the above, in such a method, in order to detect the feature points of the eye, for example, a two-dimensional CCD is used.
Therefore, a high-resolution sensor is required, which causes problems in cost and mounting, and also requires a long time for condition determination.

Further, by observing the waveform of the change in the light quantity on the time axis, it is necessary to measure a predetermined time when the blink is determined, and it takes a long time to accurately determine the blink. I will end up.

The present invention has been made in view of the above problems, and an object thereof is to quickly blink a photographer with a simple structure, which becomes a noise signal when operating a camera using a line-of-sight signal. It is to provide a line-of-sight detection device that eliminates an erroneous line-of-sight signal by making a determination based on the above determination and brings about an operation with a more accurate line-of-sight signal.

[0008]

In order to achieve the above object, a line-of-sight detection apparatus according to a first aspect of the present invention includes a light projecting means for projecting light toward a photographer's eye and the photographer's eye. The light-receiving means for detecting reflected light from the light-receiving means, the control means for controlling the light-projecting means and the detection timing of the light-receiving means, and the amount of light in the same area as the light-receiving means are determined to detect the blinking of the photographer. And a light amount determining means for controlling the light amount.

The line-of-sight detection apparatus according to the second aspect is
A light projecting means for projecting light toward the photographer's eyes, a light receiving means for detecting reflected light from the photographer's eyes, and a control means for controlling the light projecting of the light projecting means and the detection timing of the light receiving means,
The eye-gaze detecting means includes a light-amount determining means that detects the blink of the photographer by determining the light amount in the same area as the light-receiving means, and a line-of-sight detecting means that detects a line of sight based on a signal from the light-receiving means. Is configured to detect the line of sight based on the value of the light amount determining unit, and invalidate the result of the line of sight detection by the line of sight detecting unit when the light amount determining unit detects the blink of the photographer.

[0010]

That is, in the visual axis detection device according to the first aspect of the present invention, when the light projecting means projects light toward the photographer's eye, the light receiving means detects reflected light from the photographer's eye, and the control means. Controls the projection timing of the light projecting means and the detection timing of the light receiving means. Then, the light amount determining means detects the blink of the photographer by determining the light amount in the same area as the light receiving means.

Further, in the line-of-sight detection apparatus according to the second aspect, when the light projecting means projects light toward the photographer's eye, the light receiving means detects reflected light from the photographer's eye, and the control means projects the projecting light. It controls the projection timing of the light means and the detection timing of the light receiving means. Then, the light amount determination means detects the blink of the photographer by determining the light amount in the same area as the light receiving means,
The sight line detecting means detects the sight line based on the signal from the light receiving means. Further, the line-of-sight detection unit detects the line-of-sight based on the value of the light-quantity determination unit, and invalidates the result of the line-of-sight detection by the line-of-sight detection unit when the light-quantity determination unit detects the blink of the photographer.

[0012]

EXAMPLES Before describing the examples of the present invention,
The principle of the gaze direction detection adopted by the embodiment of the present invention will be described. There are various methods for detecting the line-of-sight direction. Here, as a method applicable to the camera, a method of detecting using the Purkinje image and the reflection image of the fundus or the edge of the iris will be briefly described.

As shown in FIG. 20, when light 95 generally passes through the eyeball 90, reflection occurs at each interface. That is, it is reflected at the interfaces of the front surface 91 of the cornea, the rear surface 91 'of each cornea, the front surface 93' of the lens, and the rear surface 93 "of the lens. The images formed by these reflections are well known to those skilled in the art. It is generally called the Purkinje image, and the first Purkinje image from the anterior surface of the cornea 91,
They are called the second Purkinje image, the third Purkinje image, and the fourth Purkinje image. In the figure, reference numeral 92 is an iris, 93 is a lens, 94 is a retina, and 97 is a sclera.

The present invention uses the first Purkinje image to detect the line of sight, and the first Purkinje image is a virtual image of a light source formed by reflected light from the anterior surface of the cornea 91, and is also simply called a corneal reflection image. This is the strongest reflection image.

The first Purkinje image 95a which is the reflection image
21 is shown in FIG. 21, in which light is projected onto the eyeball 90,
When the reflected image is captured, the first Purkinje image 95a having a high received light output is detected. The Purkinje images other than the first Purkinje image 95a are difficult to detect because the amount of reflected light is weak and the position where a reflected image can be formed is different.

When light is projected onto the eyeball 90, the fundus image is detected as a silhouette of the peripheral edge 99 of the pupil by the reflected light 95b from the fundus. Reflected image 95 from this fundus
Although b is shown in FIG. 21 together with the first Purkinje image 95a, the gaze direction is detected using these two images.

The detected image is shown in FIG.
It changes as shown in 2. That is, when the optical axis 98 of the eyeball 90 and the light flux projected on the eye are parallel to each other, the center of the fundus image 95b, that is, the center of the pupil and the center of the first Purkinje image 95a coincide with each other, as shown in FIG. ing. And eyeball 9
When 0 is rotated, the optical axis 98 is rotated around the rotation center 90c of the eyeball 90, as shown in FIG. In that case, the center of the fundus image 95b can be received at different positions on the sensor pixel row that receives the reflected light from the eye.
Further, the center of the first Purkinje image 95a is the fundus image 95b.
The light is received at a position relatively different from the center of. This is because the center of the curved surface on the front surface of the cornea 91 is different from the center of rotation of the eyeball.

Therefore, the amount of rotation and the amount of shift of the eyeball 90 of the photographer looking into the viewfinder can be obtained from the displacement of the absolute positions of the two images with respect to the sensor pixel array and the relative displacement of the two images. Furthermore, it is possible to determine where the photographer is looking.

Next, the feature extraction processing from the line-of-sight detection image adopted by the present invention will be described. FIG. 23 is a diagram showing rotation and corneal reflection 95a and fundus reflection 95b when the center of the eyeball is fixed. In the figure, px indicates the center of gravity of the corneal reflection 95a, and ix indicates the center of gravity of the fundus reflection 95b. In the embodiment of the present invention, the feature extraction is detected using the corneal reflection image 95a or the fundus reflection image 95b.
This is because if the center of the eyeball is fixed, the rotation angle can be detected by detecting only the barycentric position ix of the fundus reflection 95b, the barycentric position px of the corneal reflection 95a, or the barycentric position including both. In addition, the fundus reflex shown in FIG.
That is, when there is a large amount of reflected light from the fundus, when the red-eye state is not present, that is, when the iris is narrowed in a bright state, or when the reflected light does not return to the light receiving system, the output of the fundus reflection 95b further decreases. When viewed from the center of the finder, FIG. 23 (a) looks at the center with a rotation angle of 0, FIG. 23 (b) looks at the left side with a negative rotation angle, and FIG. 23 (c) looks at the right side with a positive rotation angle. It will be.

Next, in FIG. 24, the eyeball shift and the corneal reflection 9
It is a figure which shows the mode of 5a and the fundus reflection 95b. In the figure, px indicates the barycentric position of the corneal reflection 95a, and ix indicates the barycentric position of the fundus reflex 95b. The rough shift amount can be detected by detecting the barycentric position ix of the fundus reflection 95b, the barycentric position px of the corneal reflection 95a alone, or the barycentric position including both. Further, the fundus reflection shown in FIG. 24 is generally in a red-eye state, that is, when there is a large amount of reflected light from the fundus, and when it is not in the red-eye state, that is, when the iris is narrowed in a bright state, or the reflected light does not return to the light receiving system. In some cases, fundus reflex 95b
Output is further reduced. Then, when viewed from the center of the finder, FIG.
24B, the shift amount is negative and the left side is viewed, and FIG. 24C, the shift amount is positive and the right side is viewed.

Therefore, by detecting the position of the center of gravity of the detected light amount distribution from FIGS. 23 and 24, it is possible to almost detect which the line of sight is looking. In addition, generally 1
Shutter sequence (2 from near 1st release)
The shift of the eye does not change significantly in the (until the nd release), and the size of the range that the photographer tries to see can be determined by detecting the relative movement. Also,
By performing reference position detection in the sequence, it is possible to detect a rough position to be viewed.

An embodiment of the present invention based on the above-mentioned principle will be described below. FIG. 1 is a diagram showing a configuration of a visual line detection device according to a first embodiment of the present invention. As shown in the figure, the control unit 3 is connected to the light projecting unit 1 and the light receiving unit 2, the light amount determining unit 4, and the line-of-sight detecting unit 5, and the light receiving unit 2 includes the control unit 3, the light amount determining unit 4, and the line-of-sight. The light amount determination unit 4 is connected to the detection unit 5, and the light amount determination unit 4 is connected to the light receiving unit 2 and the line-of-sight detection unit 5.

In such a configuration, the light projecting unit 1 projects the illumination of the photographer's eye by the light emission signal of the control unit 3, and the light receiving unit 2 responds to the light emission signal of the control unit 3 by the photographer's eye. Receives the reflected light from. Further, the control unit 3 includes a light projecting unit 1 and a light receiving unit 2,
A timing signal (light emission signal or the like) is generated in the light amount determination unit 4 and the line-of-sight detection unit 5. Then, the light quantity determination unit 4 is connected to the control unit 3
The brightness detected by the light receiving unit 2 is compared with a predetermined value with the timing signal of (1) to determine blinking, and the information is output to the line-of-sight detecting unit 5. Then, the line-of-sight detection unit 5 receives the timing signal of the control unit 3, the blink information of the light amount determination unit 4, and the light receiving unit 2.
The line-of-sight direction is detected based on the signal of.

Next, FIG. 2 is a diagram showing the structure of a visual axis detecting device according to a second embodiment which is a further implementation of the first embodiment. In this embodiment, the focus area of the horizontally arranged three-point AF is selected using the line-of-sight signal.

As shown in the figure, a multi-point distance measuring circuit (multi-AF circuit) 11 and a photographing lens 14 are connected to a central processing unit (CPU) 12, which is connected to a display circuit 13. There is. Then, the CPU 12
Is also connected to a lens 14 via a drive circuit 15, and the CPU 12 is also connected to a light projecting LED 17, and the light projecting LED 17 and the finder optical system 1 are connected.
Reference numeral 6 is connected to the line-of-sight detection optical system 18. The line-of-sight detection optical system 18 detects the position of the center of gravity of the light amount distribution with a current ratio depending on the brightness, and a line-of-sight sensor (PSD) 19
The PSD 19 outputs a signal corresponding to the detection of the center of gravity of the light quantity, and is connected to the CPU 12 via an interface circuit (I / F circuit) 20.

In such a configuration, the multi-AF circuit 11 measures the distances at a plurality of points, and the distance measurement information is sent to the CPU 12.
The lens 14 sends the focal length information (zoom value information) to the CPU 12. Further, when the line-of-sight detection system receives a command from the CPU 12, light is emitted from the light projecting LED 17 to the eyeball 90 via the line-of-sight detection optical system 18 and the finder optical system 16. Then, the reflected light from the eyeball is again received by the visual axis sensor 19 via the finder optical system 16 and the visual axis detecting optical system 18. The output signal from the line-of-sight sensor (PSD) 19 is converted by the I / F circuit 20 into a signal corresponding to the position of the center of gravity and a signal corresponding to the brightness, and then input to the CPU 12. And C
The PU 12 detects the blinking of the photographer and sets the defocus amount based on the information on the line of sight and the information on the brightness from the I / F circuit 20 and the distance measurement information on the plurality of points from the multi AF circuit 11. Further, the lens 14 is driven via the drive circuit 15 with the set defocus amount.

Next, FIG. 3 is a diagram showing a detailed structure of the visual axis detecting optical system 18. In the figure, the light emitting LED 1
A half mirror 22 is disposed on the optical path of the light projected from the light source 7 through a light projecting lens 21, and the light projected is reflected by the half mirror 22. A prism 23 is arranged on the optical path of this reflected light, and the reflected light is reflected by the reflecting surface 23a of the prism 23, and the finder 2
4 is guided to the eyeball 90 of the photographer. Then, the reflected light from the eyeball 90 of the photographer enters the prism 23 through the finder 24 through the same optical path as the incident light and is reflected by the reflecting surface 23a. Then, the reflected light passes through the half mirror 22 and is received by the line-of-sight sensor 19 via the light receiving lens 25.

On the other hand, the subject light is incident on the prism through the display liquid crystal 26 which is superimposed on the subject light and indicates a display, passes through the reflecting surface 23a thereof, and further passes through the finder 24 to the eyeball 90 of the photographer. Is led.

Hereinafter, with reference to the flowchart of FIG.
The main sequence of the second embodiment will be described. When the main sequence is started (step S10
1) First, it is determined whether or not the 1st release switch is turned on (step S102). If the 1st release switch is turned off, the process proceeds to step S115, and if the 1st release switch is turned on, initialization is performed. To do. Here, the focus area is set as the center (step S103).

Subsequently, a subroutine "line-of-sight detection 1" to be described later is executed (step S104), and the line-of-sight coordinate X (line-of-sight center of gravity) detected in this subroutine "line-of-sight detection 1" is detected.
Is stored in X1 (step S105). Then, the multi-AF is performed by the known phase difference method or active method (step S106), and the display for detecting the visual axis reference position is performed. This display is more effective if it is performed by using flushing or sound together, and it can also be used as an AF end signal (step S107).

Then, the above subroutine "visual axis detection 1" is executed again (step S108), and the visual axis coordinate X (visual axis gravity center) detected by this subroutine "visual axis detection 1" is stored in X0 (step S109). . Then, a subroutine "focus area setting" described later is executed (step S110).

Subsequently, it is determined whether or not the 1st release switch is turned on (step S111), and if the 1st release switch is turned off, step S11.
If the 1st release switch is turned on, the process proceeds to step 5 to determine the 2nd release switch (step S112). If the 2nd release switch is off, the process returns to step S111, and if the 2nd release switch is on, the lens is driven according to the defocus amount of the previously set focus area (step S113). ). Then, a camera sequence such as an exposure sequence and a winding sequence is performed (step S114), and this sequence is finished (step S115).

Next, referring to the flow chart of FIG.
The sequence of the above-mentioned subroutine "visual axis detection 1" will be described. When the subroutine "line-of-sight detection 1" is started (step S201), initialization is performed. Here, the variables i, j, m, E, and X are initialized to "0" (step S202). Then, the variable i is incremented (step S203), and the IRED (infrared LED) is projected onto the photographer's eye (step S204).
D The signal (total current value) corresponding to the light intensity is synchronized with E
It is detected as i (step S205), and the barycentric coordinate Xi is detected (step S206). Then, the signal corresponding to the brightness is subjected to addition processing (E = E + Ei) (step S
207).

When this detection is performed a predetermined number of times (k) (step S208), a signal corresponding to the average brightness is detected (E = E / k) (step S209) and the variable j is incremented (step S210). ), Blink determination is performed from each detected current. Here, when the absolute value of the difference between the detected current Ej and the average current E is larger than the predetermined value d, it is determined that the eye blinks. The determination may be made by the ratio of the detected current Ej and the average current (step S211). And | Ej-E | <
If not d, the process proceeds to step S214, and | Ej-
When E | <d, the total sum X of the centers of gravity is calculated (X = X +
Xj) (step S212), and the variable m is incremented (step S213).

When this line-of-sight calculation is performed a predetermined number of times (k) (step S214), the average value of the center of gravity is calculated (X
= X / m) (step S215), the process exits this routine and returns to the main sequence (step S216).

Next, referring to the flow chart of FIG.
The sequence of the subroutine "focus area setting" will be described. When this subroutine "focus area setting" is started (step S301), first, the amount of movement | X0-X1 | ≤k is determined. This k is a predetermined value, and if the focus area changes due to a change in zoom value, it changes according to the zoom value (step S30).
2). Then, if | X0−X1 | ≦ k, the focus area is set at the center (distance measurement data) (step S
304). Further, if | X0-X1 | ≦ k is not satisfied, it is determined that (X0-X1)> k (step S30).
3). Then, if (X0-X1)> k, the focus area is set to the right (step S305). Further, if (X0-X1)> k is not satisfied, the focus area is set to the left (step S306). In this way, the present sequence is ended and the process returns to the main sequence (step S307).

Here, FIG. 7 is a diagram showing a display state of the finder, FIG. 7A shows a case of displaying in the center of the screen by superimposing, and FIG. 7B shows a frame. The case where it is displayed in the part is shown.

Next, a third embodiment of the present invention will be described. The configuration of the third embodiment is the same as that of the second embodiment, but the sequence of the subroutine "visual axis detection" is different,
This is characteristic.

Hereinafter, with reference to the flowchart of FIG.
The sequence of the subroutine "visual axis detection 2" according to the third embodiment will be described. Subroutine "line of sight detection 2"
Is started (step S401). Initialize. Here, the variables i, Ed, and X are initialized to "0" (step S402). Then, the IRED (infrared LED) is projected onto the eyeball of the photographer (step S40).
3) In synchronization with this IRED projection, a signal (total current value) corresponding to the light amount is detected as Ep (step S40).
4) The barycentric coordinate Xp is detected (step S405).
Then, blink determination is performed based on the detected current Ep.
Here, when the detected current Ep is larger than the predetermined current Ed, it is determined that the eye blinks (step S406). And E
When p <Ed is not satisfied, the process returns to step S403 and Ep
<Ed, the variable i is incremented (step S407), and the total X of the line-of-sight center of gravity is calculated (X = X + Xp).
Yes (step S408). Further, when the detection is performed a predetermined number of times (k) (step S409), the average value of the line-of-sight center of gravity is calculated (X = X / k) (step S410), and this routine is exited to return to the main sequence (step S41).
1).

Next, a fourth embodiment of the present invention will be described. The configuration of the fourth embodiment is similar to that of the second embodiment, but is characterized in that the reference position is fetched first in the main sequence and the reliability of the line-of-sight signal is determined.

Hereinafter, with reference to the flowchart of FIG.
The main sequence of the fourth embodiment will be described. When the main sequence is started (step S50
1) First, it is determined whether or not the 1st release switch is turned on (step S502). And 1st
If the release switch is off, the process proceeds to step S523, and if the 1st release switch is on, initialization is performed. Here, the focus area is set to the center and the timer t is set to "0". (Step S503).

Subsequently, the line-of-sight input mode is determined (step S504), and if it is not the line-of-sight input mode, the process proceeds to step S516. On the other hand, in the case of the line-of-sight input mode, display for detecting the line-of-sight reference position is performed (step S505), and subsequently, a subroutine “line-of-sight detection 3” described later is executed (step S506), and this subroutine “line-of-sight detection” is executed. The line-of-sight coordinate X (line-of-sight center of gravity) detected in 3 "is stored in X0 (step S507).

Subsequently, multi-AF according to a known phase difference method or active method is performed (step S508) to evaluate the reliability of the detection signal. Here, if the detected value X is "0", it means that the blinking state has continued, and it cannot be said that the line-of-sight information is accurate (step S509). Therefore, when X = 0 is not satisfied, the process proceeds to step S517,
When X = 0, the timer starts (step S5).
10). Then, the time judgment of the timer (T is a predetermined value, T =
0 may be performed) (step S511).

When t.gtoreq.T, the subroutine "line of sight detection 3" is executed again (step S51).
2) Store the line-of-sight coordinates (line-of-sight center of gravity) X detected in this subroutine "line-of-sight detection 3" in X1 (step S
513). Furthermore, the reliability of the detection signal is evaluated (step S
514), if X = 0, the process proceeds to step S517, and if X = 0, the above-mentioned subroutine "focus area setting" is executed. This subroutine is the second
This is similar to the embodiment (step S515).

On the other hand, if the line-of-sight mode is not set in step 504, multi-AF is performed (step S51).
6) A known multi-AF is selected (for example, the closest selection) from the multi-distance measurement values (step S517). And
The drive amount of the photographing lens is calculated from the selected AF data (step S518).

Subsequently, it is determined whether or not the 1st release switch is turned on (step S519), and if the 1st release switch is turned off, step S52.
If the first release switch is turned on, the process proceeds to step 3, and the second release switch is determined (step S520). If the 2nd release switch is off, the process returns to step S519, and if the 2nd release switch is on, the lens is driven according to the defocus amount of the focus area set previously (step S521). ). Then, a camera sequence such as an exposure sequence and a winding sequence is performed (step S522), and this sequence is ended (step S523).

Next, the sequence of the subroutine "line-of-sight detection 3" will be described with reference to the flowchart of FIG. When the subroutine "line of sight detection 3" is started (step S601), first, initialization is performed. Here, the variables i, j, m, E, and X are initialized to "0" (step S602). Then, the variable i is incremented (step S603), and IRED (infrared LE)
D) is projected on the photographer's eye (step S604), and a signal (total current value) corresponding to the light amount is detected as Ei in synchronization with this IRED projection (step S605), and the barycentric coordinate Xi is detected. (Step S606).

Then, the signal corresponding to the brightness is added (E = E + Ei) (step S607). When this detection is performed a predetermined number of times (k) (step S608), a signal corresponding to the average brightness is detected (E = E / k) (step S609), the variable j is incremented (step S610), and each detection is performed. Blinking is determined from the current. Here, the magnitude of the sum of the detected current Ej and the average currents E and d is determined (step S611). Then, if Ej <E + d is not satisfied, the process proceeds to step S614, and Ej <E + d.
If it is, the total sum X (X = X + Xj) of the center of gravity is calculated (step S612), and the variable m is incremented (step S613).

When this line-of-sight calculation is performed a predetermined number of times (k) (step S614), the regular detection number m is determined (m ≧ P) (P may be a predetermined value “0”).
(Step S615). Then, when m ≧ P, the average value of the center of gravity is calculated (X = X / m) (step S61).
6) If m ≧ P, the center of gravity value is set to 0 (X = 0) (step S617), the routine exits from this routine and returns to the main sequence (step S618).

Next, a fifth embodiment of the present invention will be described. In the fifth embodiment, a sensor for measuring brightness is installed separately. The basic principle of the detection method is the same, and the method shown below is used.

That is, the line-of-sight signal as shown in FIG. 11A is differentiated as shown in FIG.
The signal is converted into an absolute value as shown in (c) and binarized at a predetermined level Vs of the signal as shown in FIG.

On the other hand, regarding the detection of the rotation angle of the eyeball 90, the above-described processing is performed on the line-of-sight detection image as shown in FIG. 12 (a), and the signal as shown in FIG. 12 (b) is obtained. obtain. 12A and 12B, the horizontal axis represents the coordinates of the pixel row. Then, in the signal shown in FIG. 12B, when the center of the signals a and b is A, the center of the signals b and c is B, and the length between the A and B is the center position interval X of the image, As shown in FIG. 12 (c), the center position interval X and the rotation angle θ of the eyeball have a proportional relationship. In this embodiment, the rotation angle is obtained from this relationship.

FIG. 13 is a diagram showing the structure of the visual axis detection device according to the fifth embodiment. In this embodiment, in the configuration of the second embodiment, the photometric sensor 3 for detecting the brightness state of the eye is further added.
3 is arranged. Then, the line-of-sight detection sensor is changed from PSD to a two-dimensional CCD. The other structure is the same as that of the second embodiment, so the same reference numerals are given and the description thereof is omitted here.

Next, FIG. 14 is a diagram showing the structure of the line-of-sight optical system 18 of the fifth embodiment. In the fifth embodiment, in the configuration of the second embodiment, the reflected light from the eyeball 90 is further branched by the half mirror 31 after the half mirror 22, and the light amount is split by the photometric sensor 33 via the condensing optical system. To detect. The other structure is the same as that of the second embodiment, so the same reference numerals are given and the description thereof is omitted here.

The main sequence of the fifth embodiment will be described below with reference to the flowchart of FIG. When the main sequence starts (step S7)
01) It is determined whether or not the 1st release switch is turned on (step S702). Then, if the 1st release switch is off, the process proceeds to step S712, and if the 1st release switch is on, initialization is performed. Here, the focus area is set to the center (step S703). Subsequently, a later-described subroutine "visual axis detection 4" is executed (step S704), and a focus area is set according to the detected coordinates (step S705). Further, the set focus area is displayed (step S706), and AF is performed in the set focus area (step S707). Then, the lens is driven based on the AF information (step S70).
8). Then, it is determined again whether or not the 1st release switch is turned on (step S709). If the 1st release switch is off, the process proceeds to step S712, and if the 1st release switch is on, it is determined whether or not the 2nd release switch is on (step S710). If the 2nd release switch is off, the process returns to step S709. If the 2nd release switch is on, a camera sequence such as an exposure sequence and a winding sequence is performed (step S711), and this sequence ends. Yes (step S712).

Next, the sequence of the above subroutine "line-of-sight detection 4" will be described with reference to the flowchart in FIG. When the subroutine "visual axis detection 4" is started (step S801), the sensor is reset (step S802), and integration is started (step S803). Then, the brightness E is detected (step S8
04), and ends the integration (step S805). Subsequently, brightness determination (E <Ep) is performed (Ep; predetermined value) (step S806). If not E <Ep, step S8
Returning to step 02, if E <Ep, a subroutine "line-of-sight processing" described below is executed (step S807), the process exits this subroutine and returns to the main routine (step S808).

Next, the sequence of the above subroutine "line-of-sight processing" will be described with reference to the flowchart in FIG. When the subroutine "line-of-sight processing" is started (step S901), the line-of-sight data is stored in the memory (step S902), and the stored line-of-sight data is differentiated (step S903). Further, this differentiated signal is binarized at a predetermined level (step S905), and the barycentric distance X between corneal reflex and fundus reflex is calculated (step S906). In this way, the present subroutine is finished and the process returns to the main routine (step S907).
The state of this line-of-sight processing is as previously shown in FIG.

Although the embodiments of the present invention have been described above, it is needless to say that the present invention is not limited to these and various improvements and changes can be made. For example, a line sensor or the like may be used as the sensor other than the PSD as long as the center of gravity can be detected. Then, the brightness determination may be A / D converted and then added by software. Furthermore, instead of corneal reflection, the ratio of white eye / black eye may be detected by a photodiode and used as eye movement. Further, the light projection may be finished at the end of the detection, or the light projection may be continuously carried out to sample the sample discretely. Incidentally, FIG. 18A shows a state of discrete light projection, and FIG. 18B shows a state of continuous light projection.

The display for setting the reference position may be temporarily configured by inserting a liquid crystal or a field diaphragm into a state as shown in FIG. Further, the detection of the center of gravity may be detected by a histogram and may be set as the most frequent value. Further, if the distance measuring points are plural, it is not necessary to be caught by three points. As for the camera sequence, the visual axis detection sequence may be omitted when visual axis detection is not necessary, for example, when the distance measuring point is within the depth. Further, it may be used for other than the distance measuring point selection (as a camera switch).

As described above in detail, in the focus detection apparatus of the present invention, the determination of the blinking of the photographer is realized with an inexpensive and simple structure, and the line-of-sight information at the time of blinking is invalidated, whereby the blinking can be more reliably performed. Information can be input by the line of sight.

[0061]

According to the present invention, when the camera is operated by using the line-of-sight signal, the eye-blink of the photographer, which is a noise signal, can be determined at high speed with a simple structure to eliminate an erroneous line-of-sight signal. It is possible to provide a line-of-sight detection device that realizes an operation with a more accurate line-of-sight signal.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a visual line detection device according to a first exemplary embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a line-of-sight detection device according to a second embodiment that further embodies the first embodiment.

FIG. 3 is a diagram showing a detailed configuration of a line-of-sight detection optical system 18.

FIG. 4 is a flowchart showing a main sequence of the second embodiment.

FIG. 5 is a flowchart for explaining a sequence of a subroutine “visual axis detection”.

FIG. 6 is a flowchart for explaining a sequence of a subroutine “focus area setting”.

FIG. 7 is a diagram showing how a finder is displayed.

FIG. 8 is a subroutine “line of sight detection” according to the third embodiment.
3 is a flowchart for explaining the sequence of FIG.

FIG. 9 is a flowchart illustrating a main sequence of a fourth embodiment.

FIG. 10 is a flowchart for explaining a sequence of a subroutine “visual axis detection”.

FIG. 11 is a diagram for explaining line-of-sight detection.

FIG. 12 is a diagram for explaining detection of a rotation angle of an eyeball 90.

FIG. 13 is a diagram showing a configuration of a visual line detection device according to a fifth exemplary embodiment.

FIG. 14 is a diagram showing a configuration of a line-of-sight optical system 18 of a fifth example.

FIG. 15 is a flowchart illustrating a main sequence of a fifth embodiment.

FIG. 16 is a flowchart for explaining a sequence of a subroutine “visual axis detection”.

FIG. 17 is a flowchart for explaining a sequence of a subroutine “line-of-sight processing”.

FIG. 18A is a diagram showing a state of discrete light projection, and FIG. 18B is a diagram showing a state of continuous light projection.

FIG. 19 is a diagram showing an improved example in which a display for setting a reference position is configured by inserting a liquid crystal or a field diaphragm.

20 is a diagram showing the structure of a human eyeball 90. FIG.

FIG. 21 is a diagram showing a state of a first Purkinje image 95a which is a virtual image of a light source formed by reflected light from the front surface of the cornea 91.

FIG. 22 is a diagram showing how a detected image changes as the eyeball rotates.

FIG. 23 is a diagram showing a state of rotation and corneal reflection 95a and fundus reflection 95b when the center of the eyeball is fixed.

FIG. 24: Eye shift and corneal reflex 95a and fundus reflex 9
It is a figure which shows the mode of 5b.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Projection part, 2 ... Light receiving part, 3 ... Control part, 4 ... Light amount determination part, 5 ... Line-of-sight detection part.

Continuation of the front page (51) Int.Cl. ⁵ Identification number Office reference number FI technical display location G03B 13/02 9120-2K 9119-2K G03B 3/00 A

Claims (2)

[Claims] 1. A light projecting unit for projecting light toward a photographer's eye, a light receiving unit for detecting reflected light from the photographer's eye, and a light projecting unit for projecting light or detecting timing of the light receiving unit. The eye-gaze detecting device comprising: a control unit that controls the light amount; and a light amount determining unit that detects the blink of the photographer by determining the light amount in the same area as the light receiving unit. 2. A light projecting means for projecting light toward a photographer's eye, a light receiving means for detecting reflected light from the photographer's eye, and a light projecting means for detecting the light projecting means and a detection timing of the light receiving means. Control means, a light amount determining means for detecting the blink of the photographer by determining the light amount in the same area as the light receiving means, and a visual line detecting means for detecting a visual line based on the signal of the light receiving means. However, the line-of-sight detection means performs line-of-sight detection based on the value of the light amount determination means, and when the light amount determination means detects the blink of the photographer, the result of the line-of-sight detection by the line-of-sight detection means is invalidated. A line-of-sight detection device.