JPH07313460A - Glance detecting device of camera - Google Patents

Abstract

PURPOSE:To detect the glance direction of a photographer with high reliability through simple constitution. CONSTITUTION:A projection part 1 projects 1st and 2nd pieces of luminous flux which have specific angles of incidence on the eyeballs of the photographer and a sensor 4 photodetects reflected light from the eyeballs by the projection part 1 after the light is imaged. On the basis of the signal from the sensor 4 corresponding to the light projection of the projection part 1, the gaze point of the photographer is estimated while the reliability is evaluated. The 1st pieces of projection luminous flux is reflected by the eyeground and the 2nd piece of projection luminous flux is not reflected by the eyeground. If a Purkinje image shifts by more than specified, the glance information detected by time division is ignored.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a visual axis detecting device for a camera, which gives an instruction to the camera based on the visual axis information of the photographer.

[0002]

2. Description of the Related Art Conventionally, various kinds of information are input to a camera by using, for example, a dial or a button, and the operating environment becomes complicated as the input information increases. For example, many techniques for detecting the line-of-sight direction of a photographer looking into the viewfinder and inputting information to the camera based on the line-of-sight information are disclosed in Japanese Patent Laid-Open No. 63-194237, Japanese Patent Laid-Open No. 3-87818, and the like. .

Regarding the detection method using any one of the cornea, the iris, the pupil, the sclera (white eye), etc., JP-A-1-160537, JP-A-2-206425, etc. Have been disclosed by many.

Further, regarding the determination of the distance measurement information using the sight line detection position, a large number are disclosed in Japanese Patent Laid-Open Nos. 4-307506 and 5-88075.
Further, also in the signal processing of the sensor, it is common to perform the differential processing of the signal at the time of lighting / non-lighting in synchronization with the light emitting LED.

Then, the signals detected by the light projections from different positions (light projection at the red-eye generating position and light projection at the position where the red-eye does not occur) are subjected to difference processing and extracted from the face of the red-eye signal to extract the pupil. Japanese Patent Application Laid-Open No. 2-13867
No. 3 gazette is disclosed.

[0006]

However, the line-of-sight system that has already been put into practical use has a high cost because of the use of the area sensor, and there is a problem in mountability. Furthermore, since the image processing is performed from the two-dimensional image to accurately detect the direction, it takes time and a processing device is required for high-speed processing.

Further, when the line-of-sight information is used for the automatic determination of the focus area, the AF algorithm has already been able to considerably optimize the selection, and it is quite effective to add the rough line-of-sight direction as information. Therefore,
A rough line-of-sight detection system with simple processing is required.

Further, the area sensor (two-dimensional image)
However, the following problems occur. That is, it is difficult to determine whether the first detected image is really an eye image,
Secondly, it is difficult to determine the eye movement in time division.

The present invention has been made in view of the above problems, and an object thereof is to ignore the line-of-sight information when the Purkinje image deviates by a predetermined amount or more from the line-of-sight information detected by time division. It is to enable highly reliable detection of the line-of-sight direction of a photographer with a simple configuration.

[0010]

In order to achieve the above object, the visual axis detection device for a camera according to the first aspect of the present invention uses a first and a second luminous flux having a predetermined incident angle, respectively. The first and second light projecting means for projecting the light onto the eyeball, and the photoelectric conversion means for receiving the reflected light from the eyeball by the first and second light projecting means and thereafter receiving the light. It is characterized in that the gazing point of the photographer is estimated while evaluating the reliability based on the signals from the photoelectric conversion means corresponding to the first and second light projecting means.

The visual axis detecting device for a camera according to the second aspect is characterized in that the luminous flux by the first light projecting means causes fundus reflection and the luminous flux by the second light projecting means does not cause fundus reflection. To do.

Further, the visual axis detection device for a camera according to the third aspect is arranged at the first position and the optical path splitting means included in the eyepiece optical system for determining the composition, and the image is taken through the optical path splitting means. A first light projecting means for projecting light of a predetermined wavelength onto the eyeball of the person and a second light projecting device which is placed at a second position and projects light of a predetermined wavelength onto the eyeball of the photographer through the optical path dividing means. It comprises a light means and a photoelectric conversion means for receiving the reflected light from the first and second light projecting means after forming an image through the optical path dividing means, and of the outputs from the photoelectric conversion means. , And estimating the gazing point of the photographer while evaluating the reliability based on the signal corresponding to the first light source means and the signal from the second light source means.

[0013]

That is, in the visual axis detection device for a camera according to the first aspect of the present invention, the first and second light projecting means respectively emit the first and second light fluxes having a predetermined incident angle to the photographer's eyeball. And the photoelectric conversion means receives the light reflected from the eyeball by the first and second light projecting means after forming an image. And
The gazing point of the photographer is estimated while evaluating the reliability based on the signals from the photoelectric conversion means corresponding to the first and second light projecting means.

In the visual axis detecting device for a camera according to the second aspect, the luminous flux by the first light projecting means causes fundus reflection, and the luminous flux by the second light projecting means does not cause fundus reflection.

Furthermore, in the visual axis detection device for a camera according to the third aspect, the optical path dividing means is included in the eyepiece optical system for determining the composition, and the first light projecting means is placed at the first position. The light of a predetermined wavelength is projected onto the eyeball of the photographer via the optical path splitting means, and the second light projecting means is placed at the second position.
Light of a predetermined wavelength is projected onto the eyeball of the photographer through the optical path splitting means. Further, the photoelectric conversion means receives the reflected light from the first and second light projecting means after forming an image through the optical path dividing means. Then, the gazing point of the photographer is estimated based on the signal corresponding to the first light source means and the signal from the second light source means out of the outputs from the photoelectric conversion means while evaluating the reliability.

[0016]

EXAMPLES Before describing the examples of the present invention,
The principle of line-of-sight detection adopted in the present invention will be described.
There are various methods for detecting the line-of-sight direction. Here, as a method applicable to a camera, a corneal reflection image called a first Purkinje image and a fundus oculi which are already well known to those skilled in the art. A method of detecting using the reflection image or the edge of the rainbow collection will be briefly described.
Since the structural description is already known, it is omitted here.

First, FIG. 2 shows the manner in which the reflected light from the eye of the luminous flux projected from substantially the optical axis is received and output. When light is projected onto an eyeball 90 (not shown) and a reflected image thereof is captured, the first Purkinje image having a high received light output has a small amount of reflected light and a position where a reflected image can be formed is different, which makes it difficult to detect.

Then, the reflected light 95b from the fundus when the light is projected onto the eyeball 90 causes the fundus image to be detected as the silhouette of the rainbow sampling edge 94 which is the peripheral image of the pupil. The reflected image 95b from the fundus is shown in FIG. 2 together with the first Purkinje image 95a, and the gaze direction is detected using these two images. Incidentally, reference numeral 91 is a rainbow harvest, 92 is a sclera (white eye), 9
3 is a pupil, and 94 is a rainbow sampling edge.

Next, FIG. 3 shows how the detected image changes due to the rotation of the eyeball. When the optical axis 98 of the eyeball 90 and the light flux projected on the eye are parallel, as shown in FIG. 3A, 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. There is. When the eyeball 90 rotates, the optical axis 98 rotates around the rotation center 90c of the eyeball 90 as shown in FIG.

In this 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 receives light at a position relatively different from the center of the fundus image 95b. This is because the center of the curved surface on the front surface of the cornea 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 finder can be obtained from the displacement of the absolute positions of these 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. In the present invention, detection is performed using at least one of the corneal reflection image 95a and the fundus reflection image 95b from the line-of-sight detection image.

Next, FIG. 4 shows the state of rotation and corneal reflection 95a and fundus reflection 95b when the center of the eyeball is fixed and described. If the center of the eyeball is fixed, the fundus reflex 9
The rotation angle can be detected by detecting the barycentric position ix of 5b, the barycentric position px of the corneal reflection 95a, or the barycentric position including both. 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.

The fundus reflection shown in FIG. 4 is a so-called "red-eye state" in which a large amount of light is reflected from the fundus. The output of the fundus reflex 95b is further reduced when the rainbow is narrowed in a bright state or when the reflected light does not return to the light receiving system and the red eye state is not achieved. When viewed from the center of the finder, FIG. 4A shows the center at a rotation angle of 0 (reference), and FIG.
Indicates a left side with a negative rotation angle and a right side with a positive rotation angle in FIG.

Next, in FIG. 5, the eyeball shift and the corneal reflection 95
The state of a and the fundus reflection 95b will be described below. Fundus reflex 9
The approximate shift amount can be detected by detecting the barycentric position ix of 5b, the barycentric position px of the corneal reflection 95a, or the barycentric position including both.

In the figure, px indicates the position of the center of gravity of the corneal reflection 95a, and ix indicates the position of the center of gravity of the fundus reflection 95b. Further, similar to FIG. 4 described above, the fundus reflex is generally in the red-eye state, and when it is not in the red-eye condition, the fundus reflex 95
The output of b is further reduced. When viewed from the center of the finder, FIG. 5A shows the center with shift 0 (reference).
In FIG. 5B, the shift amount is negative and the left side is seen. In FIG. 5C, the shift amount is positive and the right side is seen. Generally, in one shutter sequence, that is, from the vicinity of the 1st release to the 2nd release, the eye shift does not greatly change, and the relative movement and the reference position are detected to determine the direction in which the photographer intends to see. I can know. Further, the reference position detection may be performed within the sequence.

An embodiment of the present invention based on the above-mentioned principle will be described below. FIG. 1 shows the configuration of a visual axis detection device for a camera according to a first embodiment of the present invention, which will be described. As shown in FIG. 1, the light projecting unit 2 has a light source at a position where a red-eye image and a non-red-eye image of the eye are generated in the sensor 4, and is connected to the control unit 1 and the optical system 5. It is connected to the control unit 1 and the sensor 4, and the sensor 4 is further connected to the optical system 5.

In such a configuration, the light projecting section 2 time-divisionally projects the light projecting the red-eye image and the light projecting no red-eye according to the signal from the control section 1 through the optical system 5. Do it. The sensor 4 receives the reflected light from the eye of the projected light flux and detects the detected signal via the I / F unit 3 to the control unit 1.
Send to. The I / F unit 3 performs a sequence process such as integration control and reading of the sensor 4. The control unit 1 detects the eye movement determination by time division, the reliability determination based on the integration time, and the line-of-sight direction from the detected signal from the pupil barycentric position and the corneal reflection position of the difference processing.

Hereinafter, with reference to the flowchart of FIG.
A visual line detection sequence by the visual line detection device for a camera according to the first embodiment will be described. In this sequence, first the initialization is performed. That is, the flags Fp, Ft, Fk are each set to "0" (step S1). Then, light emission for generating red eyes (light emission 0) is performed (step S2), and a subroutine "signal detection" described later is executed (step S3). Subsequently, it is determined whether or not the flag Ft = 0 regarding the integration time is set (step S4). If Ft = 0 is not established, the process proceeds to step S12.

In step S4, when Ft = 0, the detected integration time ts is stored in t0 (step S5),
Light emission (light emission 1) that does not generate red eyes is performed (step S
6) The subroutine "signal detection" described later is executed (step S108).

Then, a flag Ft = 0 relating to the integration time.
Is determined (step S8), and if Ft = 0 is not satisfied, the process proceeds to step S12. When Ft = 0, the detection integration time ts is stored in t1 (step S
9), signal processing for determining the reliability of the detection signal and calculating the line-of-sight direction is performed (step S10).

Further, it is judged whether the flag Fk = 0 regarding reliability (step S11). If Fk = 0 is not satisfied, the process proceeds to step S12, and the flag Fp for judging overall reliability is set to Fp = 1. . Then, when Fk = 0 in step S11, the present sequence is ended (step S13).

Next, the sequence of the subroutine "signal detection" will be described with reference to FIG. When the sequence of the subroutine "signal detection" is started, first, initialization is performed. That is, the flag Ft for integration is set to 0, and integration reset, timer reset, and start are performed (step S21). Then, integration is started (step S22),
A limiter determination of the integration time is performed (step S23).

In this step S23, t> Tmax
If (Tmax is a predetermined value of the limiter time of integration), the process proceeds to step S29, and if t> Tmax, the peak monitor level of the sensor 4 is determined (step S24). If the peak determination is not made, the process returns to step S23.

In step S24, when the peak is determined, the integration (integration time ts) is ended (step S2
5), the integration time ts is determined (step S26).
If ts <Tmin, the process proceeds to step S29, the flag Ft is set to 1 (step S29), and the present sequence ends (step S3).
0).

On the other hand, if ts <Tmin is not satisfied, signal reading and processing (AGC, A / D, etc.) are performed (step S27), and the peak value coordinate Pmax is detected from the signal thus read (step S28). ), And this sequence is completed (step S30).

Next, the sequence of the above-mentioned subroutine "signal processing" is shown in FIG. 8 and will be described. When the signal processing is started, first, the detected data D0 and D1 are weighted in a predetermined manner to create data D (D = a * D0-b * (t0 / t
1) × D1; a, b predetermined variables) (step S3)
1). Subsequently, the center of gravity position xd is detected from the data D (step S32), and the MAX value position (MAX
When there are a plurality of values, the barycentric position of the MAX data group) x0 and the MAX value M0 are detected (step S33), and the data D1 is detected.
Then, the MAX value position (the barycentric position of the MAX data group when there are a plurality of MAX values) x1 and the MAX value M1 are detected (step S34). Then, the amount of deviation between x0 and x1 is evaluated (the amount of eye movement is evaluated) (step S35).

If | x0-x1 | <ε is not satisfied, the process proceeds to step S42. Then, | x0-x1 |
If <ε, the average value Av0 is obtained from the data D0 (step S36). Then, the MAX value M of the data D0
The difference between 0 and the average value Av0 is evaluated (k is a predetermined value) (step S37). If not M0-Av0> k, the process proceeds to step S42. If M0-Av0> k, the average value is calculated from the data D1. Av1 is calculated (step S38).

Then, the difference between the MAX value M1 of the data D1 and the average value Av1 is evaluated (k is a predetermined value) (step S3).
9), if M1-Av1> k is not satisfied, step S42
When M1-Av1> k, the pupil diameter dp is detected from the detection signal D (step S40). A detection block is determined from the red eye center of gravity position xd and the Purkinje image position x0 thus detected (step S41), and the present sequence is exited (step S43).

In the above step S42, the flag Fk related to the total reliability is set to 1, and this sequence is exited (step S43). In the step S35, the eye movement is determined by the detection signals D0 and D1, and the step S35 is performed.
At 37 and S39, blinking is determined.

Further, when the finder optical system has a noise signal source such as internal reflection, the noise signal D detected by previously shielding the eyepiece side of the finder and lighting each LED.
After detecting d0 and Dd1, Dd = a × Dd0-b
× Dd1; a, b It is preferable to store a difference between the detection signal D and the noise signal Dd in the signal D, using a predetermined variable (integral time is the same ta: normalization processing is performed with respect to actual integral time).

Next, FIG. 9 shows the structure of a visual axis detecting device for a camera according to a second embodiment of the present invention, and will be described. The second
In this embodiment, one-dimensional multi-A is detected by one-dimensional line-of-sight detection.
The selection of the F focus area is realized.

As shown in FIG. 9, the LED circuit 13 that projects a light beam to the eye is connected to the finder / line-of-sight optical system 16 and the CPU 11, and the line sensor 15 that receives the reflected light beam from the eye is a finder / line-of-sight. Optical system 16 and I / F
The I / F circuit 14 is connected to the circuit 14 and the I / F circuit 14 is connected to the line sensor 1.
5 and the CPU 11. The CPU 11 is connected to the LED circuit 13 and the I / F circuit 14 related to the visual axis detection, and is also connected to the multi AF circuit 12 and the lens drive circuit 17.

In such a configuration, the LED circuit 13
Receives a control signal from the CPU 11 and emits light from an LED arranged at a position to generate red-eye (generates reflected light from the fundus at the sensor position) and non-red-eye, and through a luminous flux finder / line-of-sight optical system 16. Send it to your eyes. The line sensor 15 receives the reflected light flux of the projected light flux from the eye through the finder / line-of-sight optical system 16 and outputs a detection signal as I / O.
Send to the F circuit 14. The I / F circuit 14 processes the signal from the line sensor and sends it to the CPU 11 as digital information. The multi-AF circuit 12 measures the distances at a plurality of points and sends the distance measurement information to the CP.
Send to U11. The lens driving circuit 17 drives the lens according to the control signal from the CPU 11. The CPU 11 controls the line-of-sight detection system (LED circuit 13, I / F circuit 14, line sensor 15), detects the line-of-sight direction from the obtained signal, selects the distance measurement information from the multi-AF circuit 12, and drives the lens. The amount is calculated and the lens drive circuit 17 is controlled.

FIG. 10 shows the detailed structure of the finder / line-of-sight optical system. An optical arrangement is shown in FIG. An accumulating line sensor is used as a detection sensor to detect light reflected from the eye. LED for the light source
0 and LED1 are used, LED0 is arranged so that a red-eye image is generated on the detection line sensor, and LED1 is arranged with respect to the projection optical axis of LED0 so as not to generate a red-eye image on the detection line sensor. It is arranged so that the light is projected at an angle of 2 to 5 °. The light projected from the LEDs 0 and 1 is projected onto the eye through the beam splitter, and the reflected light from the eye is guided again to the line-of-sight optical system and the line sensor through the beam splitter.

10 (b) and 10 (c) show LED0 and LED.
The appearance of the image of the eye when 1 is projected is shown. Figure 10 (b)
As shown in Fig. 10, the Purkinje image and the red-eye image can be detected by the light emitted from the LED0, and as shown in Fig. 10 (c),
There is a Purkinje image due to the light projected from the LED 1, but a red-eye image cannot be detected. The positions of the Purkinje image formed by these LEDs 0 and LED1 are arranged so that they can be located at the same position in the x-axis direction (the light source is arranged on the detection sensor at a position orthogonal to the detection direction of the Purkinje image). .

Next, FIG. 11 shows signals on the line sensor shown in FIGS. 10 (b) and 10 (c). The image in FIG. 11 (c) is shown in FIG.
It shows that the signal of FIG. 1 (a) and the signal of FIG. 11 (b) are subjected to the difference processing, and it can be seen that only the red-eye image is extracted. From the image shown in FIG. 11C, the position of the center of gravity of the red eye (the center of gravity of the pupil) and the MAX value (the position of the Purkinje image) are detected, and the line-of-sight direction can be determined.

The sequence of the visual axis detection device for a camera according to the second embodiment will be described below with reference to the flowchart of FIG. When the camera sequence starts, the sequence is initialized. Here, the focus area is set to the center (step S101). Then 1
The st release is determined (step S102), 1s
If the t-release switch is off, step S108
If the 1st release switch is on
The line-of-sight detection is performed in synchronization with the st release (FIGS. 6 to 8).
Reference) (step S103). Subsequently, a subroutine "distance measurement" described later is executed to perform distance measurement of multi-AF and selection of distance measurement information (step S104).

Further, the determination of the 1st release switch is performed again (step S105). If the 1st release switch is off, the process proceeds to step S108, and if the 1st release switch is on, the 2nd release is determined (step S106). And 2
If the nd release switch is off, step S10
Returning to step 5, when the 2nd release switch is ON, a photographing sequence (lens drive, exposure determination, exposure) is performed (step S107). In this way, this sequence ends (step S108).

Here, the sequence of the subroutine "distance measurement" executed in step S104 will be described with reference to the flowchart of FIG. When you start ranging
Multi-distance measurement is performed (step S111), and the flag Fp related to total reliability is evaluated (step S112).
If Fp = 0 is not set, AF data is set by a known evaluation AF algorithm that does not use line-of-sight data (step S114). Then, when Fp = 0, the AF data is set by the information of the distance measurement block determined in the line-of-sight direction (step S113). Further, the lens drive amount is calculated from the AF data (step S115), and this sequence is exited (step S116).

Here, the structure of the line sensor is shown in FIG. 14 and will be described. FIG. 14A shows a block diagram of the line sensor, and BS is composed of a photoelectric conversion element. The DBS is a basic block for dark current correction by shielding the BS. Further, the shift register drives and controls each BS by a control signal from the KCPU. Each BS is integration-controlled by the control signals (T) and (C) of the KCPU and the monitor signal (M) output from the BS itself. The integration control performs peak value detection (when one of the BSs reaches a predetermined level, a signal for ending the integration is output) and time control. The signal (M) is recorded in the KCPU and held as the signal (T) until the next integration is started, and C holds the signal. Further, the image of the eye is sent to the A / D circuit of the CPU 21 after the dark current difference is erased and amplified according to the signal of the shift register. Then, the integration time control is determined in the brightest state (when red eye occurs), and the integration time when red eye does not occur is determined by the control signal (T) as the integration time when red eye occurs.

Further, FIG. 14B is a diagram showing an internal circuit of the BS. It is photoelectrically converted by PD and is accumulated in the capacitor C through the gate G02 for integration control, and the gate G0
1 resets PD and C together with G02. Gate G
03 performs read amplification, and the gate G04 reads a signal according to the signal of the shift register. Then, the diode D01 detects the peak value.

The integration time control sequence will be described below with reference to the flowchart of FIG. When the integration sequence is started, the sensor and the timer are reset (step S201), and the light emitting LED0 is determined (step S2).
02). Then, in the case of LED0 light projection, the timer is started at the same time when the integration is started (step S203).

Subsequently, the timer t is determined (step S204). If t> td (td is a predetermined limit value), the process proceeds to step S208. If t> td is not satisfied, the integration monitor (peak value is determined). The detection) signal is determined (step S205).

If the monitor signal is OFF, the process returns to step S204 to continue the integration. Monitor signal ON
In the case of, the integration is ended (step S206), the integration time tc is recorded by the CPU (step S207), the signal is read (step S213), and the process returns (step S214).

In step S202, LED0
If is not emitted, the timer is started at the same time as the integration is started (step S209). Then, the integration monitor (peak value detection: monitor determination level is step S2).
Signal is set higher than 07) (step S)
210), if the monitor signal is ON, step S208
When the monitor signal is OFF, the timer t is determined (step S211). If t> tc is not satisfied (step S211), the process returns to step S210 to continue the integration.

On the other hand, if t> tc in step S211, the integration is terminated and the process proceeds to step S213 (steps S211, 212). Step S208
Then, if the signal is low in reliability (the integration time is longer than expected: the eye does not exist, the integration end is shorter than expected: there is strong reflection other than the eye), the signal is reset, and the process proceeds to step S213. (Step S208).

If the integration time is judged and the integration time is not within the predetermined time, it may be judged that the reliability is low and the data may be invalidated. Next, the third aspect of the present invention
An example will be described.

The configuration of the visual axis detection device for a camera according to the third embodiment is the same as that of the second embodiment, and the calibration by the repeated signal sampling and display of a predetermined number of times when the reliability is low in the sequence of the second embodiment. It is characterized by the addition of options.

The main sequence will be described below with reference to the flowcharts of FIGS. When the camera sequence starts, the sequence is initialized. Here, the focus area is set to the center (step S301).

Then, the 1st release is determined (step S302), and if the 1st release switch is off, the process proceeds to step S315, and if the 1st release switch is on, the above-mentioned subroutine "distance measurement" is executed. Then, the distance measurement of the multi AF and the selection of the distance measurement information are performed (step S303). Then, the subroutine "line of sight detection"
Is executed and the detection block S0 is stored (step S3
04).

Unless there is a 1st or 2nd interrupt, the flag Fp relating to the overall reliability is 0 (step S30).
7) or until the predetermined number of times (M times) ends, the sight line detection (detection block S0) is repeated (steps S306 to S30).
9).

Then, in step S307, Fp = 0
In this case, the subroutine "visual axis detection (detection block S1) / correction" described later is executed (step S310).
That is, as shown in the flow chart of FIG. 17, in the subroutine “line-of-sight detection / correction” sequence, initialization display (see FIG. 19) is displayed in the finder (step S321). This becomes more effective when combined with flushing and sound. Subsequently, the line of sight is detected. Here, the detection block is S1 (step S322). Then, the block S0 detected with this line of sight is corrected in S1 to determine the final line of sight block (step S32).
3) Then, the present sequence is exited (step S324).

Then, a selection is made from a plurality of distance measurement information (step S311). That is, as shown in the flowchart of FIG. 18, in the subroutine "distance measurement determination", the flag Fp relating to the overall reliability is evaluated (step S33).
1) If Fp = 0 is not set, the AF data is set by a known evaluation AF algorithm that does not use the line-of-sight data (step S333), and if Fp = 0, the range-finding block determined in the line-of-sight direction is selected. The AF data is set by the information (step S332). Then, the lens drive amount is calculated from the AF data (step S334), and the present sequence is exited (step S335).

Since the subsequent processing is substantially the same as that of the above-mentioned second embodiment, its explanation is omitted here (step S3).
12-S315). Note that FIG. 19 shows a display state for confirming the initial position in the finder.

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 modifications can be made. For example, the line-of-sight detection method may use an area sensor. Further, although the reliability determination is performed by blinking and moving the eyes, it is more preferable to combine the determination of contact between the camera and the face (the eye is pressed against the camera finder) and the determination of glasses.

Further, the line-of-sight detection may be performed in parallel with the sequence of cameras such as AF and AE. Further, if the calibration is performed for each individual, the detection accuracy is further improved. Also in the AF method, a passive type two-dimensional wide-field AF using an area sensor may be used, or a smaller number of discrete multi-AFs may be used. Furthermore, regarding the integral control of the sensor, FIG.
As shown in, the hardware may be controlled so that the integration times are the same.

As described in detail above, according to the present invention, the present invention can provide a highly reliable visual axis detection device for a camera with a simple structure. According to the above embodiment of the present invention, the following configurations can be obtained. (1) Detection means for receiving a reflected light beam from the photographer's eyes and outputting a signal, a first light projecting means arranged at a position for generating fundus reflection when viewed from the detection means, and the detection means The second light projecting means placed at a position where the fundus reflection does not occur when viewed from above, and the line-of-sight direction of the photographer is detected by a signal from the detecting means in accordance with the first and second light projecting means, and detected. A line-of-sight detection device for a camera, comprising: a signal processing unit that determines the reliability of the received signal. (2) In the apparatus described in (1) above, the first and second light projecting means are light source means each having a wavelength region longer than 650 nm. (3) In the device described in (1), the light paths of the first and second light projecting means and the optical path of the reflected light beam share a part thereof with the optical axis of the eyepiece optical system of the camera. Is characterized by. (4) In the device described in (1), the luminous flux from the first light projecting unit and the luminous flux from the second light projecting unit are 2
It is characterized by having an angle difference of 8 to 8 degrees. (5) In the apparatus described in (1) above, the signal processing means corrects the illuminance of both when performing the difference processing between the image signal by the first light projecting means and the image signal by the second light projecting means. To do. (6) In the apparatus described in (1) above, the signal processing means is characterized in that the reliability is determined based on the positional deviation between the corneal image by the first light projecting means and the cornea image by the second light projecting means. . (7) In the device described in (1) above, the signal processing means determines reliability by using an accumulation time of the detection means. (8) In the apparatus described in (1) above, the signal processing means determines reliability according to a difference between the maximum value and the average value of the detected image signals. (9) In the apparatus described in (1) above, the signal detecting means includes a position of the maximum value of the image signal by the first light projecting means, an image signal by the first light projecting means, and an image by the second light projecting means. It is characterized in that the line-of-sight direction is detected by the position at which the center of gravity of the differential signal from the signal is given. (10) In the device described in (1) above, the first light projecting means and the second light projecting means are at the same position in the detection coordinate direction. (11) A signal detection device that detects reflected light from the photographer's eyes, a first light projecting device that causes fundus reflection when viewed from the signal detection device, and a fundus reflection that does not occur when viewed from the signal detection device A second light projecting device, a display device for displaying a line-of-sight direction at a predetermined position in the finder by a predetermined taming for line-of-sight detection, and the first and second display devices when the display device is not illuminated.
The photographer uses the signal detected by the signal detecting device according to the light projecting device and the signal detected by the signal detecting device according to the first and second light projecting devices when the display device is turned on. A visual line detection device for a camera, wherein the visual line direction of the camera is detected, and the reliability of the detected visual line direction signal is determined.

[0068]

According to the present invention, when the Purkinje image is deviated by a predetermined amount or more from the line-of-sight information detected by time division, the line-of-sight information is ignored, so that a highly reliable photographer can be provided with a simple structure. It is possible to provide a visual line detection device for a camera that can detect the visual line direction.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a light reception output of reflected light from an eye of a light beam projected from a substantially optical axis.

FIG. 3 is a diagram showing how a detected image changes due to rotation of an eyeball.

FIG. 4 is a diagram showing rotation and corneal reflection 95a and fundus reflection 95b when the center of the eyeball is fixed.

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

FIG. 6 is a flowchart showing a visual line detection sequence by the visual line detection device for a camera according to the first embodiment.

FIG. 7 is a flowchart showing a sequence of a subroutine “signal detection”.

FIG. 8 is a flowchart showing a sequence of a subroutine “signal processing”.

FIG. 9 is a diagram showing a configuration of a visual line detection device for a camera according to a second embodiment of the present invention.

FIG. 10 is a diagram showing a detailed configuration of a finder / line-of-sight optical system.

FIG. 11 is a diagram showing signals on the line sensor corresponding to FIGS. 10B and 10C.

FIG. 12 is a flowchart showing a sequence of a visual line detection device for a camera according to a second embodiment.

FIG. 13 is a flowchart showing a sequence of a subroutine “distance measurement”.

FIG. 14 is a diagram showing a configuration of a line sensor.

FIG. 15 is a flowchart for explaining a sequence of integration time control.

FIG. 16 is a flowchart showing a sequence of the visual line detection device of the third embodiment camera.

FIG. 17 is a flowchart showing a sequence of a subroutine “visual axis detection / correction”.

FIG. 18 is a flowchart showing a sequence of a subroutine “distance measurement determination”.

FIG. 19 is a diagram showing a display state of confirmation of an initial position in the finder.

[Explanation of symbols]

1 ... Control unit, 2 ... Projection unit, 3 ... I / F unit, 4 ... Sensor,
5 ... Optical system.

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location G03B 13/02 17/00 Q

Claims (3)

[Claims] 1. A first and second light projecting means for projecting first and second light fluxes, each having a predetermined incident angle, onto an eyeball of a photographer; and the first and second light projecting means. A photoelectric conversion unit that receives reflected light from the eyeball and then receives the reflected light; and based on a signal from the photoelectric conversion unit corresponding to the first and second light projecting units, a photographer's note A line-of-sight detection device for a camera, which estimates a viewpoint while evaluating reliability. 2. The line-of-sight detection device for a camera according to claim 1, wherein the luminous flux by the first light projecting means causes fundus reflection, and the luminous flux by the second light projecting means does not cause fundus reflection. . 3. An optical path splitting means included in an eyepiece optical system for determining a composition, and a first optical path splitting means which is placed at a first position and which projects light of a predetermined wavelength onto the eyeball of the photographer through the optical path splitting means. A first projecting means, a second projecting means which is placed at a second position and projects light of a predetermined wavelength onto the eyeball of the photographer through the optical path splitting means, and the first and second projecting means. Photoelectric conversion means for receiving the reflected light from the light means after forming an image through the optical path dividing means, and a signal corresponding to the first light source means in the output from the photoelectric conversion means. A line-of-sight detection device for a camera, which estimates a gazing point of a photographer while evaluating reliability based on a signal from the second light source means.