JP2017175199A - Target range setting method for vehicle cabin camera, and vehicle cabin camera - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a target range setting technique for a vehicle cabin camera, capable of setting a target range in a short time.SOLUTION: A target range setting method for a vehicle cabin camera includes: a first image acquisition step for acquiring a first image, which includes a face part of a driver, by a camera having a built-in image processor; a second image acquisition step for acquiring, by the camera, a second image which includes a face part irradiated with illumination light; a difference acquisition step for comparing the first image with the second image to acquire a luminance difference; an identification step for identifying the face part in a camera imaging range on the basis of the difference; and a setting step for setting, based on the identified part, a target range in which the face part exists in the imaging range.SELECTED DRAWING: Figure 9

Description

  Embodiments described herein relate generally to a vehicle interior camera that captures a driver's face in order to determine whether or not he is driving a nap or looking aside.

  Conventionally, when shooting at night, use a near-infrared projector to illuminate the subject and shoot with a long exposure, and when shooting in the daytime, shoot with a short exposure without using a near-infrared projector. There is an apparatus (for example, refer to Patent Document 1).

JP 2005-295340 A

  In recent years, in order to prevent a drowsy driving or a side-by-side driving, it is determined whether the driver's face is photographed with a camera in the vehicle interior to do a drowsy driving, etc. There are devices that emit alarm sounds. In such an apparatus, a range (target range) where a face exists in the shooting range of the camera is set, and only this range is processed by the face recognition system, thereby improving the speed of determination processing such as a snooze driving. . However, when disturbance light based on sunlight or the like illuminates the driver's face and the surrounding area, it becomes difficult to specify which part of the shooting range of the camera is the face part. In particular, if the face part cannot be specified temporarily, such as when the driver looks aside, if you try to specify the face part again, you must search the entire camera shooting range. There is a problem that it takes time to reset the range in which the face exists.

  The embodiment of the present invention has been made in view of such circumstances, and an object thereof is to provide a technology for setting a target range of a vehicle interior camera that can set a target range in a short time.

  A method for setting a target range of a vehicle interior camera according to an embodiment of the present invention includes: a first image acquisition step of acquiring a first image including a driver's face portion by a camera incorporating an image processor; and illumination light. A second image acquisition step of acquiring a second image including the face portion irradiated with a light by the camera; and a difference acquisition step of acquiring a luminance difference by comparing the first image and the second image; A specifying step for specifying the face portion in the shooting range of the camera based on the difference, and a setting step for setting a target range in which the face portion in the shooting range exists based on the specified portion; including.

The figure which shows the motor vehicle equipped with the vehicle interior camera which image | photographs a driver | operator. The front view which shows the camera for vehicle interiors. The side view which shows the vehicle interior camera. The block diagram which shows the structure of the camera for vehicle interiors. Explanatory drawing which shows the production | generation aspect of a 1st image and a 2nd image. The timing chart which shows various timings. Explanatory drawing which shows the setting aspect of a target range. The flowchart which shows a driving | running condition determination process. The flowchart which shows a target range setting process.

  Hereinafter, this embodiment is described based on an accompanying drawing. Reference numeral 1 in FIG. 1 denotes a vehicle interior camera according to the present embodiment. The vehicle interior camera 1 is provided to capture an image of the face of the driver D in the vehicle interior of the automobile 2 and incorporates an image processing processor. For example, the vehicle interior camera 1 is attached to the upper surface of the base 5 of the handle 4 in the dashboard 3 in the car 2. The vehicle interior camera 1 is attached in a state where the lens 10 (see FIG. 2) faces the driver D.

  The vehicle interior camera 1 is connected to a sound control unit (not shown) for controlling a speaker (not shown) mounted on the automobile 2 via a cable (not shown). Based on the image of the face of the driver D photographed by the vehicle interior camera 1, it is determined whether the driver is dozing or looking aside, and a warning is given from the speaker when the driver is asleep or looking aside Sound is reported.

  In addition, in this embodiment, the small camera 1 for vehicle interiors which can be attached to the motor vehicle 2 by retrofit is illustrated. Further, the vehicle interior camera 1 may be incorporated as a part of the automobile 2, or the vehicle interior camera 1 may be mounted when the automobile 2 is manufactured.

  As shown in FIG. 2, the vehicle interior camera 1 includes a lens 10 provided on the front side of a small box-shaped housing. A near-infrared light irradiation unit 11 is provided around the lens 10. The near-infrared light irradiation unit 11 is a device for irradiating the face of the driver D with illumination light having a specific wavelength, and is composed of six (plural) LED lamps surrounding the lens 10. Device. In the present embodiment, illumination light composed of near infrared light can be irradiated using the near infrared light irradiation unit 11. For example, near-infrared illumination light having a wavelength in the range of 880 to 940 nm is irradiated. In the present embodiment, the near-infrared light irradiation unit 11 is configured by a plurality of LED lamps. However, the near-infrared light irradiation unit 11 is configured by one LED lamp that irradiates high-luminance illumination light. Also good.

  As shown in FIG. 3, a transmission filter 12 that transmits near-infrared light is provided on an extension line of the optical axis L of the lens 10. Further, on the back side of the transmission filter 12, an image sensor 13 that is used to capture a video such as a still image or a moving image is provided. Note that the imaging element 13 is an imaging element 13 having sensitivity in the visible or infrared wavelength region, such as a CCD or CMOS image sensor. In the present embodiment, near-infrared light transmitted through the transmission filter 12 is received by the image sensor 13.

  In the vehicle interior camera 1, a target range in which a face exists in the shooting range is set. Then, when determining such as a drowsy driving, only this target range is set as a determination target. In this way, it is possible to execute the determination with a small amount of information as compared with the case where the entire photographing range is the target of determination, and not only the processing load is reduced but also the processing speed is improved. The

  In the present embodiment, when the vehicle interior camera 1 captures an image, single frame high dynamic range imaging (HDR) is possible. The dynamic range is the range of the brightest part and the dark part that can be detected by the image sensor 13. Furthermore, high dynamic range composition is an image composition technique developed for recording and displaying a wide dynamic range.

  Note that the vehicle interior camera 1 can perform switching between two types of shooting methods: normal method shooting and HDR method shooting. In the present embodiment, when performing a driving state determination process (see FIG. 8) for determining the driver's D dozing operation or the like, an image captured by the normal method is used. Further, when performing a target range setting process (see FIG. 9) in which the face portion of the driver D is specified and the target range is set in the shooting range of the vehicle interior camera 1, an image shot by the HDR method is used. When the face part of the driver D is temporarily out of the shooting range of the vehicle interior camera 1 or when the driver D temporarily turns around, the face part cannot be specified. In addition, the face portion is specified again using the target range setting process.

  A single frame type HDR will be described with reference to FIGS. In the imaging device 13, photoelectric conversion elements configured by photodiodes that detect light and generate charges are aligned in the horizontal direction and the vertical direction. One row of photoelectric conversion elements arranged in the horizontal direction is defined as one line, and a plurality of lines are arranged in parallel in the vertical direction. An image is generated by synthesizing each pixel acquired for each line. In shooting using HDR, when shooting an image with the image sensor 13, shooting is performed with different exposure times depending on each line of the image sensor 13.

  As shown in FIG. 5, in shooting using a single frame type HDR, shooting with a first exposure time (short exposure) and shooting with a second exposure time (long exposure) longer than the first exposure time. And for each line. In the present embodiment, when performing photographing using HDR, lines for performing short-time exposure and lines for performing long-time exposure are alternately arranged in the vertical direction. And a 1st image is produced | generated by synthesize | combining each pixel acquired based on the line which performs short-time exposure. Moreover, a 2nd image is produced | generated by synthesize | combining each pixel acquired based on the line which performs long-time exposure.

  Note that in normal-type shooting, shooting is performed for all lines with the same exposure time, and one image is generated by combining the pixels acquired based on all lines. That is, the first image and the second image generated by photographing using the single frame type HDR are images whose resolution is halved compared to the image generated by the photographing of the normal method.

  As shown in FIG. 6, in shooting using a single frame type HDR, a short exposure and a long exposure are performed in one frame. In this embodiment, 30 frames can be captured per second. That is, an example of a shooting mode in which one frame is 33 milliseconds is illustrated. For example, in one frame, the short exposure and the long exposure are started at the same timing. Then, the short-time exposure ends first, and then the long-time exposure ends. That is, the end timings of the short time exposure and the long time exposure are different.

  In addition, an LED lighting timing for irradiating illumination light using the near-infrared light irradiation unit 11 (LED lamp) is provided from a short-time exposure end timing T1 to a long-time exposure end timing T2 in one frame. This illumination light irradiation starts after the end timing T1 of the short-time exposure and ends before the end timing T2 of the long-time exposure. That is, the long exposure period includes a period in which the face portion of the driver D is illuminated by the illumination light of the near infrared light irradiation unit 11.

  In addition, when the face portion of the driver D cannot be specified in one frame shooting, shooting using the HDR is continued. Further, when the face portion of the driver D can be identified, the normal mode shooting is switched.

  In this embodiment, in one frame, the long exposure time is three times the short exposure time. When specifying the face portion of the driver D, the same exposure as that of the second image photographed by the long-time exposure is performed by triple the luminance of each pixel of the first image photographed by the short-time exposure. The value is equivalent to the brightness when the picture was taken in time.

  Next, an aspect of image processing when the face portion of the driver D is specified in the imaging range of the vehicle interior camera 1 will be described with reference to FIG. First, the brightness of each pixel of the first image is tripled to a value corresponding to the brightness when the image is taken with the same exposure time as the second image. In the first image, there is a portion with high luminance due to disturbance light. The disturbance light is a portion that is brightly illuminated by light including near-infrared light such as sunlight inserted from outside the vehicle 2. Further, during the timing of capturing the first image, the illumination light using the near-infrared light irradiation unit 11 (LED lamp) is not irradiated, so the brightness of the face portion of the driver D is low.

  Further, the second image also has a portion with high luminance due to ambient light. Moreover, since the illumination light using the near-infrared light irradiation part 11 (LED lamp) is irradiated during the imaging | photography timing of a 2nd image, the brightness | luminance of the driver | operator's D face part is high. That is, the second image includes a portion whose luminance is increased by irradiation with illumination light.

  Then, each pixel of the first image and each pixel of the second image are respectively compared to acquire a luminance difference corresponding to near-infrared light, and a difference image is generated based on these differences. In this difference image, a portion having a high luminance or a portion having a low luminance in both the first image and the second image, that is, a portion having a small difference in luminance is reproduced with a low luminance. In addition, one of the first image and the second image having a high luminance or a low luminance, that is, a portion having a large luminance difference is reproduced with a high luminance.

  For example, since a portion having high luminance due to disturbance light exists in both the first image and the second image, this portion is a portion having low luminance in the difference image. On the other hand, the portion of the face of the driver D whose luminance is increased by irradiation of illumination light exists only in the second image, and this portion is a portion having high luminance in the difference image.

  Further, a binarized image is generated in which each pixel of this difference image is binarized with a predetermined threshold. In this way, the face portion is clearly identified. Thus, since the face part can be specified based on the difference in luminance, the photographing range of the face part of the driver D can be specified in a short time. Then, a predetermined area including the specified face portion is set as a target range. For example, the circumscribed rectangular coordinates (circumscribed rectangle) of the specified face part may be obtained and set as the target range, or a predetermined circular area centered on the specified face part may be set as the target range. Also good. Further, when there are a plurality of high-luminance portions in the difference image, the largest region among them may be specified as the face portion.

  Next, a driving situation determination process and a target range setting process executed by the vehicle interior camera 1 of the present embodiment will be described with reference to FIGS. 4, 8, and 9. In the description of each step in the flowchart, for example, a portion described as “Step S11” is abbreviated as “S11”.

  As shown in FIG. 4, the vehicle interior camera 1 according to the present embodiment includes a photographing module including an image processing processor that performs various processes such as a photographing control unit 20 that performs photographing control. The imaging control unit 20 controls the near-infrared light irradiation unit 11 and the image sensor 13 to capture an image.

  As shown in FIG. 8, when the vehicle interior camera 1 is activated and the driving situation determination process is started, first, a target range setting process is executed (S11). In this target range setting process, the face range of the driver D is specified in the shooting range of the vehicle interior camera 1 to set the target range.

  As shown in FIG. 9, in the target range setting process, first, imaging of an image is started using the imaging element 13 (S21: imaging step). This shooting is performed using a single frame type HDR (see FIGS. 5 and 6). Then, at the start time of one frame, short-exposure shooting is started and long-exposure shooting is started.

  Next, when the short-time exposure end timing T1 is reached, the short-time exposure photographing ends (S22). Then, detection information of each pixel acquired in each line corresponding to the short-time exposure of the image sensor 13 is transferred to the first image generation unit 21. A first image is generated based on the detection information transferred to the first image generation unit 21 (S23: first image generation step). Further, the first image acquisition unit 23 acquires the first image generated by the first image generation unit 21, and the first image is transferred to the first image storage unit 26 and stored in the first image storage unit 26. (S24: first image acquisition step).

  Next, the LED lighting timing which irradiates illumination light using the near infrared light irradiation part 11 (LED lamp) is started. And after lighting for a fixed time, LED lighting timing is complete | finished (S25: irradiation step). In this way, by irradiating near-infrared illumination light having a specific wavelength from the end of the short exposure (first exposure time) to the end of the long exposure (second exposure time) in one frame. The second image including the part of the face irradiated with the illumination light can be acquired.

  Next, when the long exposure end timing T2 is reached, the long exposure shooting ends (S26). Then, detection information of each pixel acquired in each line corresponding to the long exposure of the image sensor 13 is transferred to the second image generation unit 22. A second image is generated based on the detection information transferred to the second image generation unit 22 (S27: second image generation step). Further, the second image acquisition unit 24 acquires the second image generated by the second image generation unit 22, and the second image is transferred to the second image storage unit 27 and stored in the second image storage unit 27. (S28: Second image acquisition step).

  Next, the luminance of the first image stored in the first image storage unit 26 is doubled by the luminance doubling unit 29 (S29: doubling step). In the present embodiment, the brightness of the first image is tripled. That is, the brightness of the first image is doubled to a value corresponding to the brightness when the first image is taken with the same exposure time. The first image obtained by doubling the luminance and the second image stored in the second image storage unit 27 are transferred to the difference image acquisition unit 30. The difference image acquisition unit 30 compares each pixel of the first image and each pixel of the second image to acquire a luminance difference corresponding to near-infrared light. And the difference image acquisition part 30 produces | generates an image based on these differences, and acquires a difference image (S30: difference acquisition step). When the difference image acquisition unit 30 acquires the difference image, the stored contents of the first image storage unit 26 and the second image storage unit 27 are deleted.

  Next, the difference image acquired by the difference image acquisition unit 30 is transferred to the binarized image acquisition unit 31. In the binarized image acquisition unit 31, each pixel of the difference image is binarized with a predetermined threshold. Then, the binarized image acquisition unit 31 acquires a binarized image based on each pixel that has been binarized (S31).

  Next, the binarized image acquired by the binarized image acquiring unit 31 is transferred to the setting face part specifying unit 32. The setting face part specifying unit 32 determines whether or not a range corresponding to the face part of the driver D has been specified based on the binarized image (S32: specifying step). For example, when the area of the high luminance portion (white portion) in the binarized image is larger than a predetermined specific first threshold value or smaller than a predetermined specific second threshold value, It is determined that the range corresponding to the face portion could not be specified. In addition, when the area of the high-luminance portion (white portion) in the binarized image is smaller than the predetermined first threshold value and larger than the second threshold value, the range corresponding to the face portion Is determined to have been identified.

  When the setting face part specifying unit 32 determines that the range corresponding to the face part could not be specified, the process returns to S21 described above. On the other hand, when it is determined that the range corresponding to the face portion can be specified, the target range is set by the target range setting unit 33 (S33: setting step). The set target range is transferred to the determination face part specifying unit 34 and is a target of determination when the driver D is determined to doze. When the target range is set, the target range setting process ends, and the process returns to the driving situation determination process.

  Returning to FIG. 8, in the driving state determination process, determination of whether or not the driver D is dozing or driving is started based on the third image captured by the normal method. First, when the third image is taken, the near-infrared light irradiation unit 11 (LED lamp) is turned on to irradiate the face of the driver D with illumination light (S12). Further, a third image is taken using the image sensor 13 (S13). Then, after photographing, the near infrared light irradiation unit 11 (LED lamp) is turned off (S14).

  Next, the third image acquisition unit 25 acquires the third image captured by the image sensor 13, and the third image is transferred to the third image storage unit 28 and stored in the third image storage unit 28. (S15: Third image acquisition step). Then, the third image is transferred to the determination face part specifying unit 34. The determination face part specifying unit 34 determines whether or not the face part of the driver D has been specified based on the target range part included in the third image (S16).

  Here, when the face portion of the driver D cannot be specified, for example, when the face portion of the driver D is temporarily out of the shooting range of the vehicle interior camera 1 or when the driver D is temporarily If the user turns back, the process returns to the target range setting process (S11) described above and sets the target range again. On the other hand, when the face portion of the driver D can be specified, the target range correction unit 35 corrects the target range according to the specified face portion range (S17). Then, the corrected target range is transferred to the determination face part specifying unit 34. Note that when the range of the specified face portion is the same as the target range that has already been set, the processing for correcting the target range may be omitted.

  Next, the driving condition determination unit 36 analyzes the target range included in the third image and determines the driving condition of the driver D (S18). Then, the driving state determination unit 36 determines whether or not a dozing driving or a side-by-side driving is performed based on the face portion of the driver D included in the third image (S19: determination step). Here, when the driver does not take a nap or looks aside, the process returns to S12 described above and the above steps are repeated. On the other hand, when the driver is dozing or looking aside, the warning output unit 37 outputs a warning signal to the effect that a warning sound is output to the sound control unit (not shown) (S20: output step). And it returns to S12 mentioned above and repeats the above-mentioned step. The sound control unit (not shown) that has acquired the warning signal performs control for outputting a warning sound for a predetermined time using a speaker (not shown).

  Further, the third image used for the determination by the driving state determination unit 36 may be deleted from the third image storage unit 28. Note that the third image used for the determination by the driving state determination unit 36 may be stored in association with the shooting time without being deleted. Furthermore, only the third image when it is determined that the driver is dozing or looking aside may be stored, and the other third images may be deleted.

  As described above, in the present embodiment, it is possible to specify the face portion of the driver D in one frame shooting, and thus it is possible to set the target range in a short time. Further, since the first image and the second image can be taken simultaneously in one frame, the first image and the second image can be acquired in a short time.

  In addition, by doubling the brightness of the first image to a value corresponding to the second exposure time, the first image and the second image are compared even if the exposure time differs between the first image and the second image. Thus, it is possible to perform a process of acquiring a luminance difference.

  In the present embodiment, it is assumed that the two types of shooting methods, the normal method and the HDR method, can be switched, but a camera capable of only the HDR method may be used. That is, even when the driving situation determination process (see FIG. 8) is performed, shooting by the HDR method may be performed.

To summarize the present embodiment, the following procedure is used when the face portion of the driver D is specified by performing shooting using the HDR method.
(1) An LED is projected for a certain period of time after completion of short-time exposure image capture.
(2) The portion of the human face becomes brighter than the surroundings due to the LED projection.
(3) Long-exposure images are captured.
(4) The brightness of the short-time exposure image is multiplied by α, which is the ratio between the short-time exposure and the long-time exposure.
(5) A difference image between the image multiplied by α and the long-time exposure image of (3) is obtained.
(6) Threshold processing is performed to obtain a region having a large difference, region detection is performed, and the largest one is set as a face region.

  Further, by determining whether or not the driver D is dozing driving or looking aside based on the target range of the third image, the determination process may be performed using only the target range. Driving determination processing speed can be improved.

  Further, since the illumination light applied to the face portion of the driver D by the near infrared light irradiation unit 11 is near infrared light, the near infrared light is light that is not visible to human eyes. Even if the near-infrared light is irradiated to the face of the person D, driving is not hindered.

  In addition, the method for setting the target range of the vehicle interior camera 1 of the present embodiment is used to determine whether or not a doze driving or a side-by-side driving is performed. You may apply to the camera 1 for vehicle interiors used for determination other than a dozing driving or a side-view driving. For example, the present invention may be applied to a driver recognition system that can individually identify the face of the driver D.

  In the present embodiment, the target range is the face portion of the driver D, but for example, only the portions of both eyes of the driver D may be set as the target range. By setting only this eye portion as the target range, the determination can be executed with a smaller amount of information, and the processing load is reduced.

  In this embodiment, a predetermined warning sound is output when it is determined that the driver is dozing or looking aside. In addition to the warning sound, a predetermined warning is displayed on the display screen of the car navigation system. May be performed. Further, when it is determined that the driver is dozing or looking aside, a message to that effect may be output and the driver D's falling asleep or looking aside may be recorded.

  In the present embodiment, a shooting cycle of 30 frames per second is illustrated, but shooting may be performed at other shooting cycles. For example, an imaging cycle of 60 frames per second may be used. In the present embodiment, the shooting period is the same when shooting the first image and the second image and when shooting the third image, but the first image and the second image are shot. The shooting cycle may be different between when the third image is shot and when the third image is shot.

  In this embodiment, the long exposure time is three times the short exposure time in one frame, but the long exposure time is limited to three times the short exposure time. Instead, it may be twice, four times, or five times as long. The long exposure time may not be an integral multiple of the short exposure time. When comparing the luminance difference between the first image and the second image, the luminance of each pixel of the first image is set to a long time. What is necessary is just to double by the magnification | multiplying_factor corresponding to the time of exposure.

  In this embodiment, when the third image is taken, the near infrared light irradiation unit 11 is turned on to irradiate the face of the driver D with illumination light. When it is provided in the vehicle interior camera 1 so that the brightness around the driver D can be detected and it is determined that the brightness is sufficiently bright due to sunlight irradiation or the like, the near-infrared light is taken when the third image is taken. Turning on the light irradiation unit 11 may be omitted.

  According to the embodiment described above, the luminance difference is obtained by comparing the first image and the second image, and the face portion can be specified based on the luminance difference, so the target range can be set in a short time. can do.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, changes, and combinations can be made without departing from the scope of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 ... Vehicle interior camera, 2 ... Automobile, 3 ... Dashboard, 4 ... Handle, 5 ... Base, 10 ... Lens, 11 ... Near infrared light irradiation part, 12 ... Transmission filter, 13 ... Imaging element, 20 ... Photographing Control unit, 21 ... first image generation unit, 22 ... second image generation unit, 23 ... first image acquisition unit, 24 ... second image acquisition unit, 25 ... third image acquisition unit, 26 ... first image storage unit , 27 ... second image storage unit, 28 ... third image storage unit, 29 ... luminance doubling unit, 30 ... difference image acquisition unit, 31 ... binarized image acquisition unit, 32 ... setting face part specifying unit, 33 ... Target range setting unit 34... Judgment face part specifying unit 35... Target range correction unit 36 36 driving state determination unit 37 37 warning output unit.

Claims (7)

A first image acquisition step of acquiring a first image including a part of the driver's face by a camera incorporating an image processing processor;
A second image acquisition step of acquiring a second image including the portion of the face irradiated with illumination light by the camera;
A difference obtaining step of obtaining a luminance difference by comparing the first image and the second image;
A specifying step of specifying a part of the face in the shooting range of the camera based on the difference;
A setting step for setting a target range where the face portion in the shooting range exists based on the specified portion;
Of setting the target range of a vehicle interior camera including An imaging step of performing imaging for the first exposure time and imaging for the second exposure time longer than the first exposure time for each line of the image sensor in one frame;
A first image generation step of generating the first image based on each line corresponding to the first exposure time;
A second image generation step of generating the second image based on each line corresponding to the second exposure time;
A method for setting a target range of a vehicle interior camera according to claim 1.   The method for setting a target range of a vehicle interior camera according to claim 2, further comprising: an irradiation step of irradiating the illumination light after the first exposure time ends before the second exposure time ends in one frame.   The method for setting a target range of a vehicle interior camera according to claim 2 or 3, further comprising a doubling step of doubling the luminance of the first image to a value corresponding to the second exposure time. A third image acquisition step of acquiring a third image including the target range;
A determination step of determining whether the driver is dozing driving or looking aside based on the target range of the third image;
The method for setting a target range of a vehicle interior camera according to any one of claims 1 to 4, further comprising:   The method for setting a target range of a vehicle interior camera according to any one of claims 1 to 5, wherein the illumination light is near infrared light. A first image acquisition unit that acquires a first image including a portion of the driver's face with a camera that includes an image processing processor;
A second image acquisition unit that acquires, by the camera, a second image including the face portion irradiated with illumination light;
A difference acquisition unit that compares the first image with the second image to acquire a luminance difference;
A specifying unit that specifies a part of the face in the shooting range of the camera based on the difference;
A setting unit that sets a target range in which the face portion in the shooting range exists based on the specified portion;
A vehicle interior camera.