JP2017084157A - Warning device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a warning device which estimates a seat which is highly likely to be a driver's seat, even if the driver's seat of another vehicle cannot be identified, to issue a warning at a proper timing.SOLUTION: A warning device 100 includes: a driver's seat identifying unit 23 which executes first image recognition processing on a front image formed by a front camera 10 arranged in a vehicle 1 and capturing an oncoming vehicle 2, to identify a position of a driver's seat of the oncoming vehicle 2 in the front image; a driver's seat estimating unit 24 which executes second image recognition processing different from the first image recognition processing on the front image when the driver's seat identifying unit 23 cannot identify the position of the driver's seat, to estimate the position of the driver's seat in the front image; and an attention state determination unit 25 which executes third image recognition processing on an area including the driver's seat in the front image, to determine whether a driver of the oncoming vehicle 2 is in a careless driving state or not.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an alarm device for a vehicle.

  Conventionally, a camera mounted on the host vehicle is used to determine whether the driver of the other vehicle is in a forward careless state using an image obtained by shooting the other vehicle. An alarm device that issues an alarm to a driver of a vehicle has been developed (see, for example, Patent Document 1).

JP 2014-178971 A

  Generally, a front seat of a vehicle includes a plurality of seats, and any one of these seats is a driver seat. The warning device that issues a warning to the driver of another vehicle specifies which seat of the front seat is the driver's seat when multiple passengers are seated in the front seat of the other vehicle. It is required to identify the person. This is because, for example, when the passenger in the passenger seat is in a forward careless state, the driver is prevented from issuing an alarm erroneously even though the driver is not in the forward careless state.

  In the vehicle collision warning device disclosed in Patent Document 1, first, the image recognition unit detects the steering position of another vehicle using a known image recognition method such as pattern matching based on the captured image data input from the imaging device. To do. By detecting the position of the steering wheel first, when there is an occupant not only in the driver but also in the passenger seat in the other vehicle, it is specified that the occupant existing on the side where the steering is located is the driver.

  However, the steering wheel (hereinafter referred to as “handle”) of the other vehicle may not be shown in the image obtained by photographing the other vehicle with the camera mounted on the host vehicle. For example, when a camera is provided in the bumper part of the host vehicle, the handle of the other vehicle may be hidden behind the front part of the other vehicle in an image taken by the camera. Even if the camera mounting position is increased, if the vehicle height of the other vehicle is higher than the vehicle height of the host vehicle, the handle of the other vehicle may be hidden behind the front part of the other vehicle. Furthermore, when an in-vehicle information device or an ornament is installed on the dashboard of another vehicle, the handle of the other vehicle may be hidden behind the in-vehicle information device or the ornament.

  The vehicle collision warning device disclosed in Patent Document 1 has a problem in that the position of the driver's seat of the other vehicle cannot be specified and the driver cannot be specified when the handle of the other vehicle is not shown in the captured image. In this case, there is a problem that an alarm is issued at an improper timing, such as when an occupant in the front passenger seat is in a state of carelessness in front of the passenger.

  The present invention has been made to solve the above-described problems. Even when the driver's seat of another vehicle cannot be specified, the seat having a high probability of being the driver's seat is estimated and an appropriate timing is obtained. An object of the present invention is to provide an alarm device that can issue an alarm.

  The alarm device according to the present invention includes a driver seat identifying unit that performs a first image recognition process on an image obtained by photographing a vehicle with a camera provided in the host vehicle, and identifies the position of the driver seat of the other vehicle in the image; If the position of the driver's seat cannot be specified by the seat specifying unit, a driver's seat estimating unit that executes a second image recognition process different from the first image recognition process for the image and estimates the position of the driver's seat in the image And a caution state determination unit that executes a third image recognition process on a region including the driver's seat in the image to determine whether the driver of the other vehicle is in a forward careless state, and a caution state determination unit And a warning output unit that issues a warning to the driver using an output device provided in the host vehicle when it is determined that the driver is in a forward careless state.

  Since the alarm device of the present invention is configured as described above, even if the driver's seat of the other vehicle cannot be specified by the first image recognition process, a highly probable seat that is the driver's seat by the second image recognition process. It is possible to estimate and issue an alarm at an appropriate timing.

It is a block diagram which shows the state which provided the alarm device which concerns on Embodiment 1 of this invention in the own vehicle. It is a block diagram which shows the hardware constitutions of the control part which concerns on Embodiment 1 of this invention. It is a block diagram which shows the other hardware constitutions of the control part which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows the own vehicle and oncoming vehicle which concern on Embodiment 1 of this invention. It is a flowchart which shows operation | movement of the driver's seat specific | specification part which concerns on Embodiment 1 of this invention. It is a flowchart which shows operation | movement of the driver's seat estimation part which concerns on Embodiment 1 of this invention. It is a flowchart which shows operation | movement of the caution state determination part and alarm output part which concern on Embodiment 1 of this invention. It is a block diagram which shows the other state which provided the alarm device which concerns on Embodiment 1 of this invention in the own vehicle. It is a block diagram which shows the state which provided the alarm device which concerns on Embodiment 1 of this invention in the server. It is a block diagram which shows the state which provided the alarm device which concerns on Embodiment 2 of this invention in the own vehicle. It is explanatory drawing which shows the own vehicle and subsequent vehicle which concern on Embodiment 2 of this invention.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a state in which an alarm device according to Embodiment 1 of the present invention is provided in a host vehicle. With reference to FIG. 1, the alarm device 100 of Embodiment 1 is demonstrated.

  The storage device 30 is, for example, a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), an EEPROM (Electrically Erasable Programmable Read Only Memory), or a hard disk such as an EEPROM. Yes. In the storage device 30, an emblem database 31, a vehicle type database 32, and a traffic classification database 33 are stored in advance.

  The emblem database 31 includes image data of a vehicle manufacturer's emblem (hereinafter referred to as “maker emblem”), information indicating whether or not the vehicle manufacturer indicated by the maker emblem manufactures a right-hand drive vehicle, It is the database which linked | related with the information which shows whether the vehicle manufacturer which a maker emblem manufactures the vehicle of a left steering wheel.

  The vehicle type database 32 associates information indicating the vehicle type of the vehicle, information indicating whether or not the right-hand drive type exists in the vehicle type, and information indicating whether or not the left-hand drive type exists in the vehicle type. Database.

  The traffic classification database 33 is a database in which the coordinate values on the map are associated with information indicating the vehicle traffic classification in an area including the coordinate values. There are so-called “left-hand traffic” and “right-hand traffic” in the vehicle traffic classification. Generally, right-hand drive vehicles are popular in left-hand traffic areas, and left-hand drive vehicles are popular in right-hand traffic areas.

  A GPS (Global Positioning System) receiver 40 receives a GPS signal transmitted by a GPS satellite (not shown). The own vehicle position calculation unit 21 calculates a coordinate value indicating the current position of the own vehicle 1 using the GPS signal received by the GPS receiver 40. The own vehicle position calculation unit 21 outputs the calculated coordinate value to the driver seat estimation unit 24.

  The front camera 10 is, for example, a camera provided in the front bumper portion of the host vehicle 1, the front half of the vehicle body, the vehicle interior, or the like. The front camera 10 captures the front of the host vehicle 1 at predetermined time intervals according to instructions from the oncoming vehicle detection unit 22. The front camera 10 outputs data of a photographed image (hereinafter referred to as “front image”) to the oncoming vehicle detection unit 22. The time interval at which the front camera 10 captures the front image may be a predetermined time interval (for example, every second) or may be a time interval that varies depending on the traveling speed of the host vehicle 1.

  The oncoming vehicle detection unit 22 instructs the front camera 10 to capture a front image. The oncoming vehicle detection unit 22 detects an oncoming vehicle that runs in a lane adjacent to the traveling lane of the host vehicle 1 using data input from the front camera 10. Specifically, the oncoming vehicle detection unit 22 performs image recognition processing such as so-called “pattern recognition” on the front image indicated by the input data, and the oncoming vehicle detection unit 22 faces the vehicle 1 within a predetermined distance. It detects the car. The predetermined distance may be a predetermined distance (for example, 20 meters), or may be a different distance depending on the traveling speed of the host vehicle 1. When the oncoming vehicle is detected, the oncoming vehicle detection unit 22 outputs data of a front image showing the oncoming vehicle to the driver seat identification unit 23.

  The driver seat specifying unit 23 uses the data input from the oncoming vehicle detection unit 22 to specify the position of the driver seat of the oncoming vehicle in the front image indicated by the data. Here, “specifying” the position of the driver's seat means specifying the seat that is the driver's seat almost certainly among the front seats of the oncoming vehicle, and specifying the position of this seat. That is, the front image indicated by the data is a picture of the oncoming vehicle taken from the front, and the vehicle interior is shown through the windshield of the oncoming vehicle. The driver seat identification unit 23 performs image recognition processing such as pattern recognition on the front image, and detects a passenger sitting on the front seat of the oncoming vehicle. When the number of detected passengers is one, the driver seat specifying unit 23 specifies that the seat on which the passenger sits is the driver seat among the front seats of the oncoming vehicle.

  In addition, the driver's seat specifying unit 23 performs image recognition processing such as pattern recognition on the front image, and detects the handle of the oncoming vehicle imaged in the front image. When the steering wheel can be detected, the driver seat specifying unit 23 specifies that the seat provided with the steering wheel is the driver seat among the front seats of the oncoming vehicle.

  Hereinafter, the image recognition processing performed by the driver seat specifying unit 23 is collectively referred to as “first image recognition processing”. When the position of the driver's seat can be specified by the first image recognition process, the driver's seat specifying unit 23 outputs data of the front image and information indicating the position of the driver's seat in the front image to the attention state determining unit 25. Is. Further, when the position of the driver's seat cannot be specified by the first image recognition process, the driver's seat specifying unit 23 outputs the data of the front image to the driver's seat estimating unit 24.

  The driver seat estimation unit 24 uses the data input from the driver seat identification unit 23 to estimate the position of the driver's seat of the oncoming vehicle in the front image indicated by the data. Here, “estimating” the position of the driver's seat is to estimate a seat that has a high probability of being the driver's seat when the driver's seat in the front seats of the oncoming vehicle cannot be identified with certainty. It means to estimate the position of this seat. That is, the driver seat estimation unit 24 performs image recognition processing such as pattern recognition on the front image, and determines whether each of the plurality of passengers sitting on the front seat of the oncoming vehicle is a child or an adult. Determine and detect adults included in the passenger. Whether each passenger is a child or an adult is determined by, for example, the physique of the passenger, the position of the head in the height direction, and the presence or absence of a child seat. When the detected number of adults is one, the driver seat estimation unit 24 estimates that the seat on which the adult sits is the driver seat among the front seats of the oncoming vehicle.

  In addition, the driver seat estimation unit 24 performs image recognition processing such as pattern recognition on the front image, and detects a fender lamp of an oncoming vehicle that is captured in the front image. Generally, the fender lamp is disposed at an end portion farther from the driver's seat among the left end portion and the right end portion of the vehicle front end portion. When the fender lamp can be detected, the driver seat estimation unit 24 estimates that the seat farther from the fender lamp among the front seats of the oncoming vehicle is the driver seat.

  In addition, the driver seat estimation unit 24 performs image recognition processing such as pattern recognition on the front image, and detects the maker emblem of the oncoming vehicle that is captured in the front image. If the driver seat estimation unit 24 can detect the manufacturer emblem, the driver seat estimation unit 24 acquires information associated with the manufacturer emblem from the emblem database 31 stored in the storage device 30. When the vehicle manufacturer indicated by the detected maker emblem manufactures only the right-hand drive vehicle, the driver seat estimation unit 24 estimates that the right seat among the front seats of the oncoming vehicle is the driver seat. When the vehicle manufacturer indicated by the detected manufacturer emblem manufactures only the left-hand drive vehicle, the driver seat estimation unit 24 estimates that the left seat among the front seats of the oncoming vehicle is the driver seat.

  In addition, the driver seat estimation unit 24 performs image recognition processing such as pattern recognition on the front image, and detects the appearance of the oncoming vehicle imaged in the front image. Here, the “appearance” of the oncoming vehicle is the shape, pattern or color of the oncoming vehicle, or a combination thereof. The driver seat estimation unit 24 identifies the vehicle type of the oncoming vehicle based on the detected appearance. When the driver seat estimation unit 24 can identify the vehicle type of the oncoming vehicle, the driver seat estimation unit 24 acquires information associated with the vehicle type from the vehicle type database 32 stored in the storage device 30. The driver seat estimation unit 24 estimates that the right seat among the front seats of the oncoming vehicle is the driver seat when only the right steering wheel type exists in the detected vehicle type. The driver seat estimation unit 24 estimates that the left seat among the front seats of the oncoming vehicle is the driver seat when only the left steering wheel type exists in the detected vehicle type.

  Hereinafter, the image recognition processing by the driver seat estimation unit 24 is collectively referred to as “second image recognition processing”. When the position of the driver seat can be estimated by the second image recognition process, the driver seat estimation unit 24 outputs data of the front image and information indicating the position of the driver seat in the front image to the attention state determination unit 25. Is.

  Moreover, the driver seat estimation part 24 estimates the position of a driver's seat with reference to the traffic classification database 33, when the position of a driver's seat cannot be estimated by the 2nd image recognition process. That is, the driver seat estimation unit 24 acquires a coordinate value indicating the current position of the host vehicle 1 from the host vehicle position calculation unit 21. The driver seat estimation unit 24 refers to the traffic classification database 33 stored in the storage device 30 and determines the traffic classification of the vehicle in the area including the current position of the host vehicle 1. When the traffic classification is left-hand traffic, the driver seat estimation unit 24 regards the oncoming vehicle as a right-hand drive vehicle and estimates that the right seat among the front seats of the oncoming vehicle is the driver seat. When the traffic classification is right-hand traffic, the driver seat estimation unit 24 regards the oncoming vehicle as a left-hand drive vehicle, and estimates that the left seat among the front seats of the oncoming vehicle is the driver seat. Even when the driver's seat estimation unit 24 estimates the position of the driver's seat with reference to the traffic classification database 33, the driver's seat estimation unit 24 sends the data of the front image and the information indicating the position of the driver's seat in the front image to the attention state determination unit 25. Output.

  The attention state determination unit 25 uses the data of the front image and the information indicating the position of the driver seat input from the driver seat specification unit 23 or the driver seat estimation unit 24, and the driver of the oncoming vehicle is in a state of carelessness ahead. Whether or not. Specifically, the attention state determination unit 25 executes image recognition processing such as pattern recognition on a region including the position of the driver's seat indicated by the input information in the front image indicated by the input data. At least the head of the driver of the oncoming vehicle is shown in this area.

  As a result of the image recognition process, for example, when the driver's face is facing a direction different from the front of the oncoming vehicle, or when the driver's line of sight is facing a direction different from the traveling direction of the oncoming vehicle, There is a high probability that the driver of the car is doing so-called “side-view driving”. The caution state determination unit 25 determines that the driver of the oncoming vehicle is in a forward careless state when the driver of the oncoming vehicle is in these states.

  In addition, as a result of image recognition processing for a plurality of front images that are temporally continuous, for example, when the driver of an oncoming vehicle continuously closes her eyes, or when her head is continuously bowed There is a high probability that the driver of the oncoming vehicle is doing so-called “sleeping driving”. In addition, when the frequency of blinking by the driver of the oncoming vehicle is high, or when the eyes are insufficiently opened, there is a high probability that the driver will fall asleep afterwards. The caution state determination unit 25 determines that the driver of the oncoming vehicle is in a forward careless state when the driver of the oncoming vehicle is in these states.

  In addition, as a result of the image recognition process, when the driver of the oncoming vehicle is making a call using a portable terminal such as a smartphone that is capable of voice communication, the attention state determination unit 25 indicates that the driver of the oncoming vehicle is in a state of carelessness ahead. It is determined that

  Hereinafter, the image recognition processing by the attention state determination unit 25 is collectively referred to as “third image recognition processing”. The caution state determination unit 25 instructs the alarm output unit 26 to output an alarm when it is determined by the third image recognition process that the driver of the oncoming vehicle is in a forward careless state.

  The warning output unit 26 issues a warning to the driver of the oncoming vehicle using the output device 50 provided in the host vehicle 1 according to an instruction from the attention state determination unit 25. Specifically, for example, the alarm output unit 26 alerts the driver of the oncoming vehicle by sounding the horn 51 of the host vehicle 1 or blinking the headlamp 52 or the hazard lamp 53.

  The driver seat identification unit 23, the driver seat estimation unit 24, the caution state determination unit 25, and the alarm output unit 26 constitute an alarm device 100. The own vehicle position calculation unit 21, the oncoming vehicle detection unit 22, and the alarm device 100 constitute a control unit 20.

Next, the hardware configuration of the control unit 20 will be described with reference to FIGS. 2 and 3.
FIG. 2 shows an example of the hardware configuration of the control unit 20. As shown in FIG. 2, the control unit 20 includes a dedicated processing circuit 71. The processing circuit 71 is, for example, a system LSI (Large-Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof. In addition, the function of each part of the own vehicle position calculation part 21, the oncoming vehicle detection part 22, the driver seat specification part 23, the driver seat estimation part 24, the caution state determination part 25, and the alarm output part 26 shown in FIG. It may be realized, or the functions of the respective units may be collectively realized by a processing circuit. The processing circuit 71 may be a processing circuit of an ECU (Electronic Control Unit) mounted on the host vehicle 1.

  FIG. 3 shows another example of the hardware configuration of the control unit 20. As shown in FIG. 3, the control unit 20 includes a memory 72 and a processor 73. In the memory 72, a program for realizing the functions of the own vehicle position calculation unit 21, the oncoming vehicle detection unit 22, the driver seat identification unit 23, the driver seat estimation unit 24, the caution state determination unit 25, and the alarm output unit 26 illustrated in FIG. Is remembered. The processor 73 reads out and executes a program stored in the memory 72. The processor 73 is, for example, a CPU (Central Processing Unit). The memory 72 is, for example, a RAM (Random Access Memory), a ROM, a flash memory, a semiconductor memory such as an EPROM or an EEPROM, a hard disk drive, or an optical disk. As the memory 72 and the processor 73, a memory and a processor of an in-vehicle information device such as a car navigation device mounted on the host vehicle 1 can be used. Alternatively, a memory and a processor of a portable information terminal such as a PND (Portable Navigation Device) or a smartphone brought into the host vehicle 1 can be used.

Next, the operation of the alarm device 100 will be described with reference to FIGS.
FIG. 4 shows an image of the host vehicle 1 and an oncoming vehicle 2 running in a lane adjacent to the travel lane of the host vehicle 1. In the example of FIG. 4, the front camera 10 having a shooting angle of view θ of about 90 ° is provided in the front bumper portion of the host vehicle 1. When the oncoming vehicle 2 appears in the front image captured by the front camera 10 and the oncoming vehicle detection unit 22 detects the oncoming vehicle 2, the oncoming vehicle detection unit 22 outputs the data of the front image to the driver seat identification unit 23.

  FIG. 5 is a flowchart showing the operation of the driver seat identification unit 23. When the front image data is input from the oncoming vehicle detection unit 22, the driver seat identification unit 23 starts the process of step ST1.

  First, in step ST <b> 1, the driver seat identification unit 23 detects a passenger sitting in the front seat of the oncoming vehicle 2 with respect to the front image indicated by the data input from the oncoming vehicle detection unit 22. Execute the process.

  Next, in step ST2, the driver seat identifying unit 23 determines whether or not the position of the driver seat of the oncoming vehicle 2 in the front image can be identified based on the result of the first image recognition process in step ST1. That is, the driver seat identification unit 23 determines that the position of the driver seat can be identified when the number of passengers detected in the first image recognition process in step ST1 is one. On the other hand, when the passenger cannot be detected in the first image recognition process in step ST1, or when the detected number of passengers is plural, the driver seat specifying unit 23 cannot specify the position of the driver seat. Is determined.

  If the position of the driver's seat can be specified (step ST2 “YES”), then, in step ST3, the driver's seat specifying unit 23 determines the oncoming vehicle 2 in the front image based on the result of the first image recognition process in step ST1. Locate the driver's seat. That is, the driver's seat specifying unit 23 specifies that the seat on which the passenger is sitting detected in step ST1 among the front seats of the oncoming vehicle 2 is the driver's seat. The driver seat identification unit 23 outputs the data of the front image and information indicating the identified driver seat position to the attention state determination unit 25.

  If the position of the driver's seat cannot be specified (step ST2 “NO”), then, in step ST4, the driver's seat specifying unit 23 performs the oncoming vehicle 2 with respect to the front image indicated by the data input from the oncoming vehicle detection unit 22. The first image recognition process for detecting the handle of is performed.

  Next, in step ST5, the driver seat identifying unit 23 determines whether the position of the driver seat of the oncoming vehicle 2 in the front image can be identified based on the result of the first image recognition process in step ST4. That is, the driver seat identification unit 23 determines that the position of the driver seat can be identified when the steering wheel can be detected in the first image recognition process of step ST4. On the other hand, when the steering wheel cannot be detected in the first image recognition process in step ST4, the driver seat specifying unit 23 determines that the position of the driver seat cannot be specified.

  If the position of the driver's seat can be specified (step ST5 “YES”), then in step ST3, the driver's seat specifying unit 23 determines the oncoming vehicle 2 in the front image based on the result of the first image recognition process in step ST5. Locate the driver's seat. That is, the driver's seat specifying unit 23 specifies that the seat provided with the handle detected in step ST4 among the front seats of the oncoming vehicle 2 is the driver's seat. The driver seat identification unit 23 outputs the data of the front image and information indicating the identified driver seat position to the attention state determination unit 25.

  If the position of the driver's seat cannot be specified (step ST5 “NO”), then in step ST6, the driver's seat specifying unit 23 does not specify the position of the driver's seat of the oncoming vehicle 2 in the front image, and the data of the front image. Is output to the driver seat estimation unit 24.

  FIG. 6 is a flowchart showing the operation of the driver seat estimation unit 24. When the front image data is input from the driver seat identification unit 23, the driver seat estimation unit 24 starts the process of step ST11.

  First, in step ST11, the driver seat estimation unit 24 compares each of a plurality of passengers sitting in the front seat of the oncoming vehicle 2 with respect to the front image indicated by the data input from the driver seat identification unit 23. It is determined whether there is an adult or not, and a second image recognition process for detecting an adult included in the passenger is executed.

  Next, in step ST12, the driver seat estimation unit 24 determines whether or not the position of the driver seat of the oncoming vehicle 2 in the front image can be estimated based on the result of the second image recognition process in step ST11. That is, the driver seat estimation unit 24 determines that the position of the driver seat can be estimated when the number of adults detected in the second image recognition process in step ST11 is one. On the other hand, if an adult cannot be detected in the second image recognition process in step ST11, or if there are a plurality of detected adults, the driver seat estimation unit 24 determines that the position of the driver seat cannot be estimated. To do.

  If the position of the driver's seat can be estimated (step ST12 “YES”), then, in step ST13, the driver's seat estimation unit 24 determines the oncoming vehicle 2 in the front image based on the result of the second image recognition process in step ST11. Estimate the position of the driver's seat. That is, the driver seat estimation unit 24 estimates that the seat on which the adult sits detected in step ST11 among the front seats of the oncoming vehicle 2 is the driver seat. The driver seat estimation unit 24 outputs the data of the front image and information indicating the estimated driver seat position to the attention state determination unit 25.

  If the position of the driver's seat cannot be estimated (step ST12 “NO”), then, in step ST14, the driver's seat estimating unit 24 performs the oncoming vehicle 2 with respect to the front image indicated by the data input from the driver's seat specifying unit 23. The second image recognition process for detecting the fender lamp is executed.

  Next, in step ST15, the driver seat estimation unit 24 determines whether the position of the driver seat of the oncoming vehicle 2 in the front image can be estimated based on the result of the second image recognition process in step ST14. That is, the driver seat estimation unit 24 determines that the position of the driver seat can be estimated when the fender lamp can be detected in the second image recognition process of step ST14. On the other hand, if the fender lamp cannot be detected in the second image recognition process in step ST14, the driver seat estimation unit 24 determines that the driver seat position cannot be estimated.

  If the position of the driver's seat can be estimated (step ST15 “YES”), then, in step ST13, the driver's seat estimation unit 24 determines the oncoming vehicle 2 in the front image based on the result of the second image recognition process in step ST14. Estimate the position of the driver's seat. That is, the driver seat estimation unit 24 estimates that, among the front seats of the oncoming vehicle 2, a seat farther from the fender lamp detected in step ST4 is the driver seat. The driver seat estimation unit 24 outputs the data of the front image and information indicating the estimated driver seat position to the attention state determination unit 25.

  If the position of the driver's seat cannot be estimated (step ST15 “NO”), then in step ST16, the driver's seat estimation unit 24 performs the oncoming vehicle 2 with respect to the front image indicated by the data input from the driver's seat specifying unit 23. The second image recognition process for detecting the manufacturer emblem is executed. When the manufacturer emblem can be detected, the driver seat estimation unit 24 acquires information associated with the detected manufacturer emblem from the emblem database 31 stored in the storage device 30.

  Next, in step ST17, the driver seat estimation unit 24 determines whether the position of the driver seat of the oncoming vehicle 2 in the front image can be estimated based on the result of the second image recognition process in step ST16. That is, the driver seat estimation unit 24 can detect the maker emblem in the second image recognition process in step ST16, and the vehicle manufacturer indicated by the maker emblem is a right-hand drive vehicle or a left-hand drive vehicle. When only one of them is manufactured, it is determined that the position of the driver's seat can be estimated. On the other hand, when the maker emblem cannot be detected in the second image recognition process of step ST16, or when the vehicle manufacturer indicated by the detected maker emblem manufactures both a right-hand drive vehicle and a left-hand drive vehicle. The driver seat estimation unit 24 determines that the position of the driver seat cannot be estimated.

  If the position of the driver's seat can be estimated (step ST17 “YES”), then in step ST13, the driver's seat estimation unit 24 determines the oncoming vehicle 2 in the front image based on the result of the second image recognition process in step ST16. Estimate the position of the driver's seat. That is, when the vehicle manufacturer indicated by the detected manufacturer emblem manufactures only the right-hand drive vehicle, the driver seat estimation unit 24 determines that the right seat among the front seats of the oncoming vehicle 2 is the driver seat. presume. Further, when the vehicle manufacturer indicated by the detected maker emblem manufactures only the left-hand drive vehicle, the driver seat estimation unit 24 determines that the left seat among the front seats of the oncoming vehicle 2 is the driver seat. presume. The driver seat estimation unit 24 outputs the data of the front image and information indicating the estimated driver seat position to the attention state determination unit 25.

  If the position of the driver's seat cannot be estimated (step ST17 “NO”), then, in step ST18, the driver's seat estimating unit 24 performs the oncoming vehicle 2 with respect to the front image indicated by the data input from the driver's seat specifying unit 23. A second image recognition process for detecting the appearance of the image is executed. The driver seat estimation unit 24 identifies the vehicle type of the oncoming vehicle 2 based on the detected appearance. When the vehicle type of the oncoming vehicle 2 can be specified, the driver seat estimation unit 24 acquires information associated with the specified vehicle type from the vehicle type database 32 stored in the storage device 30.

  Next, in step ST19, the driver seat estimation unit 24 determines whether the position of the driver seat of the oncoming vehicle 2 in the front image can be estimated based on the result of the second image recognition process in step ST18. That is, if the driver seat estimation unit 24 can identify the vehicle type of the oncoming vehicle 2 in step ST18 and only one of the right handle type and the left handle type exists in the vehicle type, It is determined that the position of the driver's seat can be estimated. On the other hand, if the vehicle type of the oncoming vehicle 2 cannot be specified in step ST18, or if both the right-hand drive type and the left-hand drive type exist in the specified vehicle type, the driver seat estimation unit 24 It is determined that the position cannot be estimated.

  If the position of the driver's seat can be estimated (step ST19 “YES”), then, in step ST13, the driver's seat estimation unit 24 determines the oncoming vehicle 2 in the front image based on the result of the second image recognition process in step ST18. Estimate the position of the driver's seat. That is, the driver seat estimation unit 24 estimates that the right seat among the front seats of the oncoming vehicle 2 is the driver seat when only the right steering wheel type exists in the specified vehicle type. Further, the driver seat estimation unit 24 estimates that the left seat among the front seats of the oncoming vehicle 2 is the driver seat when only the type of the left steering wheel exists in the specified vehicle type. The driver seat estimation unit 24 outputs the data of the front image and information indicating the estimated driver seat position to the attention state determination unit 25.

  If the position of the driver's seat cannot be estimated (step ST19 “NO”), then, at step ST20, the driver's seat estimating unit 24 refers to the traffic classification database 33 and the position of the driver's seat of the oncoming vehicle 2 in the front image. Is estimated. That is, the driver seat estimation unit 24 acquires a coordinate value indicating the current position of the host vehicle 1 from the host vehicle position calculation unit 21. The driver seat estimation unit 24 refers to the traffic classification database 33 stored in the storage device 30 and determines the traffic classification of the vehicle in the area including the current position of the host vehicle 1. When the traffic classification is left-hand traffic, the driver seat estimation unit 24 regards the oncoming vehicle 2 as a right-hand drive vehicle, and estimates that the right seat among the front seats of the oncoming vehicle 2 is the driver seat. . When the traffic classification is right-hand traffic, the driver seat estimation unit 24 regards the oncoming vehicle 2 as a left-hand drive vehicle and estimates that the left seat of the front seats of the oncoming vehicle 2 is the driver seat. . The driver seat estimation unit 24 outputs the data of the front image and information indicating the estimated driver seat position to the attention state determination unit 25.

  FIG. 7 is a flowchart showing operations of the attention state determination unit 25 and the alarm output unit 26. When the attention state determination unit 25 receives the data of the front image and the information indicating the position of the driver seat from the driver seat identification unit 23 or the driver seat estimation unit 24, the attention state determination unit 25 starts the process of step ST31.

  First, in step ST31, the attention state determination unit 25 is input from the driver seat specification unit 23 or the driver seat estimation unit 24 among the front images indicated by the data input from the driver seat specification unit 23 or the driver seat estimation unit 24. The third image recognition process is executed on the region including the position of the driver's seat indicated by the information.

  Next, in step ST32, the caution state determination unit 25 determines whether or not the driver of the oncoming vehicle 2 is in a forward careless state as a result of the third image recognition process in step ST31. Specifically, for example, when the driver's face is facing a different direction from the front of the oncoming vehicle 2 as a result of the third image recognition process, the driver's line of sight is facing a direction different from the traveling direction of the oncoming vehicle 2. The driver's eyes are continuously closed, the driver is continuously lying down, the driver's blinking frequency is high, and the driver's eyes are not fully open In some cases, when the driver is talking using the mobile phone terminal, the caution state determination unit 25 determines that the driver of the oncoming vehicle 2 is in a forward careless state. On the other hand, as a result of the third image recognition process, when the driver of the oncoming vehicle 2 is not in these states, the caution state determination unit 25 determines that the driver of the oncoming vehicle 2 is not in the forward careless state.

  When the driver of the oncoming vehicle 2 is in a forward careless state (step ST32 “YES”), the attention state determination unit 25 instructs the alarm output unit 26 to output an alarm. Next, in step ST <b> 33, the warning output unit 26 issues a warning to the driver of the oncoming vehicle 2 using the output device 50. Specifically, the alarm output unit 26 alerts the driver of the oncoming vehicle 2 by sounding the horn 51 of the host vehicle 1 or blinking the headlamp 52 or the hazard lamp 53.

  When the driver of the oncoming vehicle 2 is not in the forward careless state (step ST32 “NO”), the attention state determination unit 25 does not instruct the alarm output unit 26 to output an alarm. Thereby, the process of step ST33 is skipped and the alarm device 100 ends the process.

  As described above, in the alarm device 100 according to the first embodiment, first, the driver's seat specifying unit 23 executes the first image recognition process, specifies the driver's seat among the front seats of the oncoming vehicle 2, and the front image. The position of the driver's seat at is identified. When the driver seat identification unit 23 cannot identify the position of the driver seat, the driver seat estimation unit 24 executes a second image recognition process that is a detection target different from the first image recognition process, and A seat with high probability that is a driver seat among the seats is estimated, and the position of the driver seat in the front image is estimated. Furthermore, when the position of the driver's seat cannot be estimated by the second image recognition processing, the driver's seat estimating unit 24 refers to the traffic classification database 33 and the probability that the driver's seat is the driver seat among the front seats of the oncoming vehicle 2. A high seat is estimated, and the position of the driver's seat in the front image is estimated. That is, even if the driver's seat of the oncoming vehicle 2 cannot be specified by the first image recognition process, a seat with a high probability of being the driver's seat is estimated by the second image recognition process and the warning is issued at an appropriate timing. be able to.

  The first image recognition process only needs to be able to specify the seat that is the driver seat in the front seat of the oncoming vehicle 2 with certainty and specify the position of the driver seat in the front image. It is not limited to the processing shown in steps ST1 and ST4.

  Further, the second image recognition process may be any process that can estimate the position of the driver seat in the front image by estimating a seat that is highly likely to be a driver seat among the front seats of the oncoming vehicle 2, and FIG. It is not limited to the processing shown in steps ST11, ST14, ST16, ST18.

  For example, the second image recognition process may detect a room mirror provided in the passenger compartment of the oncoming vehicle 2. In this case, the driver's seat estimation unit 24 determines the installation angle of the detected rearview mirror, and estimates that the seat to which the mirror surface of the rearview mirror is directed is the driver's seat among the front seats of the oncoming vehicle 2.

  Alternatively, the second image recognition process may detect an information terminal such as a PND installed on the dashboard of the oncoming vehicle 2. In this case, the driver's seat estimation unit 24 determines the installation angle of the detected information terminal, and estimates that the seat to which the display screen of the information terminal is directed is the driver's seat among the front seats of the oncoming vehicle 2. Alternatively, the driver seat estimation unit 24 determines the installation position of the detected information terminal, and estimates that the seat closer to the information terminal among the front seats of the oncoming vehicle 2 is the driver seat.

  Alternatively, the second image recognition process may detect a HUD (Head-Up Display) projection device installed on the dashboard of the oncoming vehicle 2. In this case, the driver seat estimation unit 24 determines the installation position of the detected projection device, and estimates that the seat closer to the projection device among the front seats of the oncoming vehicle 2 is the driver seat.

  Further, the execution order of the processes of steps ST11 and ST12, the processes of steps ST14 and ST15, the processes of steps ST16 and ST17, and the processes of steps ST18 and ST19 is not limited to the order shown in FIG. , It may be executed by changing the order. Further, the driver seat estimation unit 24 may execute only a part of these processes.

  Further, the driver seat estimation unit 24 executes the processes of steps ST11 and ST12, the processes of steps ST14 and ST15, the processes of steps ST16 and ST17, and the processes of steps ST18 and ST19 in parallel. Also good. The driver seat estimation unit 24 may adopt a result of a process with a higher priority when the priority of each process is set in advance and the estimation result is different between the processes.

  Moreover, the structure which removed the GPS receiver 40, the own vehicle position calculation part 21, and the traffic classification database 33 which are shown in FIG. 1 may be sufficient. In this case, whether or not the driver's seat estimation unit 24 regards the oncoming vehicle 2 as a right-hand drive vehicle or a left-hand drive vehicle in step ST20 is set in advance. For example, when the destination of the host vehicle 1 is a left-hand traffic area, it is preferable to set it to the right steering wheel, and when it is a right-hand traffic area, it is preferable to set the left steering wheel.

  Further, the output device 50 is not limited to the horn 51, the headlamp 52, and the hazard lamp 53 as long as it can issue a warning to the driver of the oncoming vehicle 2 to alert the driver. For example, a sounding device different from the horn 51 may be attached to the own vehicle 1 and the alarm output unit 26 may sound this sounding device. Further, a light source different from the headlamp 52 and the hazard lamp 53 may be attached to the host vehicle 1, and the alarm output unit 26 may turn on the light source. Moreover, even if it attaches a display apparatus to the own vehicle 1, the display screen of this display apparatus faces the front of the own vehicle 1, and the alarm output unit 26 displays a text or an illustration to issue an alarm. good. When vehicle-to-vehicle communication is possible between the host vehicle 1 and the oncoming vehicle 2, a transmitter for vehicle-to-vehicle communication provided on the host vehicle 1 is used as the output device 50, and a receiver provided on the oncoming vehicle 2 is used. An alarm signal may be transmitted, and an alarm may be issued by voice from a speaker provided in the passenger compartment of the oncoming vehicle 2.

  Further, the oncoming vehicle detection unit 22 may detect the oncoming vehicle 2 using an ultrasonic sensor or a photoelectric sensor provided in the exterior portion of the host vehicle 1 instead of the front camera 10. In this case, the oncoming vehicle detection unit 22 instructs the front camera 10 to start capturing a front image when the oncoming vehicle 2 is detected using these sensors. Alternatively, the oncoming vehicle detection unit 22 may detect the oncoming vehicle 2 using both the front camera 10 and the sensor.

  In addition, as shown in FIG. 8, the storage device 30 is provided in the server 3 outside the host vehicle 1 and the wireless communication device 60 is provided in the host vehicle 1 to perform wireless communication between the host vehicle 1 and the server 3. good. The driver seat estimation unit 24 transmits an acquisition request for information stored in the storage device 30 to the server 3 using the transmitter 61 of the wireless communication device 60. The driver seat estimation unit 24 receives information transmitted by the server 3 using the receiver 62 of the wireless communication device 60.

  In addition, as shown in FIG. 9, the storage device 30 and the alarm device 100 are provided in the server 3 outside the host vehicle 1, and the wireless communication device 60 is provided in the host vehicle 1 to perform wireless communication between the host vehicle 1 and the server 3. It can be done. The oncoming vehicle detection unit 22 uses the transmitter 61 of the wireless communication device 60 to transmit data on the front image of the oncoming vehicle 2 to the driver seat identification unit 23 of the server 3. The own vehicle position calculation unit 21 transmits a coordinate value indicating the current position of the own vehicle 1 to the driver seat estimation unit 24 of the server 3 using the transmitter 61 of the wireless communication device 60. The alarm output unit 26 of the server 3 transmits a signal instructing output of the alarm to the wireless communication device 60. The wireless communication device 60 causes the output device 50 to output an alarm in response to the instruction signal received by the receiver 62.

  As described above, the alarm device 100 according to the first embodiment performs the first image recognition process on the front image in which the front camera 10 provided in the host vehicle 1 captures the oncoming vehicle 2, and the oncoming vehicle 2 in the front image. And a second image recognition process different from the first image recognition process for the front image when the driver's seat position cannot be specified by the driver's seat specifying unit 23. The driver seat estimator 24 for estimating the position of the driver seat in the front image and the third image recognition process for the region including the driver seat in the front image are executed, and the driver of the oncoming vehicle 2 When the attention state determination unit 25 determines whether or not the vehicle is in a forward careless state, and the attention state determination unit 25 determines that the driver is in a forward careless state, the output device 50 provided in the host vehicle 1 is Use to driver And an alarm output unit 26 that emits broadcast. Even if the driver's seat of the oncoming vehicle 2 cannot be specified by the first image recognition process, a highly probable seat that is the driver's seat can be estimated by the second image recognition process, and an alarm can be issued at an appropriate timing. .

Embodiment 2. FIG.
FIG. 10 is a block diagram illustrating a state in which the alarm device according to Embodiment 2 of the present invention is provided in the host vehicle. With reference to FIG. 10, the alarm device 100 of Embodiment 2 is demonstrated. In addition, the same code | symbol is attached | subjected to the structural member similar to Embodiment 1, and description is abbreviate | omitted.

  The rear camera 10a is, for example, a camera provided in the rear bumper portion of the host vehicle 1, the rear half of the vehicle body, the vehicle interior, or the like. The rear camera 10a captures the rear of the host vehicle 1 at predetermined time intervals according to an instruction from the following vehicle detection unit 22a. The rear camera 10a outputs data of a captured image (hereinafter referred to as “rear image”) to the subsequent vehicle detection unit 22a.

  The succeeding vehicle detection unit 22a instructs the rear camera 10a to capture a rear image. Further, the succeeding vehicle detection unit 22a detects a succeeding vehicle that travels in the same lane as the traveling lane of the host vehicle 1 using data input from the rear camera 10a. Specifically, the following vehicle detection unit 22a performs image recognition processing such as pattern recognition on the front image indicated by the input data, and detects a following vehicle whose distance from the host vehicle 1 is within a predetermined distance. To do. When the succeeding vehicle detection unit 22a detects the following vehicle, the succeeding vehicle detection unit 22a outputs data of a rear image in which the following vehicle is captured to the driver seat identification unit 23.

  The output device 50 includes a tail lamp 54a and a brake lamp 55a in addition to the horn 51 and the hazard lamp 53 similar to those in the first embodiment. The warning output unit 26 is configured to issue a warning to the driver of the following vehicle by sounding the horn 51 or blinking the tail lamp 54a, the brake lamp 55a, or the hazard lamp 53. In this way, the alarm device 100 according to the second embodiment is configured.

  The hardware configuration of the control unit 20 shown in FIG. 10 is the same as that described with reference to FIG. 2 or FIG.

  FIG. 11 shows an image of the host vehicle 1 and the following vehicle 4 running in the same lane as the travel lane of the host vehicle 1. In the example of FIG. 11, a rear camera 10 a having a shooting angle of view θ of about 90 ° is provided in the rear bumper portion of the host vehicle 1. When the subsequent vehicle 4 appears in the rear image captured by the rear camera 10a and the subsequent vehicle detection unit 22a detects the subsequent vehicle 4, the subsequent vehicle detection unit 22a outputs data of the rear image to the driver seat identification unit 23.

  Since the operation of the driver seat identification unit 23 shown in FIG. 10 is the same as that described with reference to FIG. 5 in the first embodiment, illustration and description thereof are omitted. The operation of the driver seat estimation unit 24 shown in FIG. 10 is the same as that described in Embodiment 1 with reference to FIG. Since the operations of the attention state determination unit 25 and the alarm output unit 26 shown in FIG. 10 are the same as those described with reference to FIG. 7 in the first embodiment, illustration and description thereof are omitted.

  That is, in the alarm device 100 according to the second embodiment, first, the driver seat identification unit 23 executes the first image recognition process to identify the driver seat among the front seats of the succeeding vehicle 4 and drive in the front image. Identify the seat location. If the driver seat identification unit 23 cannot identify the position of the driver seat, the driver seat estimation unit 24 executes a second image recognition process that is different from the first image recognition process and detects the front of the succeeding vehicle 4. A seat with high probability that is a driver seat among the seats is estimated, and the position of the driver seat in the front image is estimated. Furthermore, when the position of the driver's seat cannot be estimated by the second image recognition process, the driver's seat estimating unit 24 refers to the traffic classification database 33 and the probability that the driver's seat is the driver's seat among the front seats of the succeeding vehicle 4. A high seat is estimated, and the position of the driver's seat in the front image is estimated. As a result, even if the driver's seat of the succeeding vehicle 4 cannot be specified by the first image recognition process, a highly probable seat that is the driver's seat is estimated by the second image recognition process, and an alarm is issued at an appropriate timing. Can be emitted.

  The alarm device 100 according to the second embodiment can employ various modifications similar to those of the alarm device 100 according to the first embodiment.

  Further, both the front camera 10 and the oncoming vehicle detection unit 22 shown in FIG. 1 and the rear camera 10a and the following vehicle detection unit 22a shown in FIG. 10 are provided, and an alarm to the driver of the oncoming vehicle 2 and the following vehicle 4 are provided. It is good also as a structure which can output both the warning to the driver | operator of this.

  As described above, the alarm device 100 according to the second embodiment performs the first image recognition process on the rear image in which the rear camera 10a provided in the host vehicle 1 captures the subsequent vehicle 4, and the subsequent vehicle 4 in the rear image. And a second image recognition process different from the first image recognition process for the rear image when the driver's seat position cannot be specified by the driver's seat specifying unit 23. The driver seat estimation unit 24 that estimates the position of the driver seat in the rear image and the third image recognition process for the region including the driver seat in the rear image are performed, and the driver of the succeeding vehicle 4 When the attention state determination unit 25 determines whether or not the vehicle is in a forward careless state, and the attention state determination unit 25 determines that the driver is in a forward careless state, the output device 50 provided in the host vehicle 1 is Use to guard the driver And an alarm output unit 26 that emits. Even when the driver's seat of the succeeding vehicle 4 cannot be specified by the first image recognition process, a highly probable seat that is the driver's seat can be estimated by the second image recognition process, and an alarm can be issued at an appropriate timing. .

  In the present invention, within the scope of the invention, any combination of the embodiments, or any modification of any component in each embodiment, or omission of any component in each embodiment is possible. .

  DESCRIPTION OF SYMBOLS 1 Own vehicle, 2 Oncoming vehicle, 3 Server, 4 Subsequent vehicle, 10 Front camera, 10a Rear camera, 20 Control part, 21 Own vehicle position calculation part, 22 Oncoming vehicle detection part, 22a Subsequent vehicle detection part, 23 Driver's seat specification 24, driver's seat estimation unit, 25 attention state determination unit, 26 alarm output unit, 30 storage device, 31 emblem database, 32 vehicle type database, 33 traffic classification database, 40 GPS receiver, 50 output device, 51 horn, 52 heads Lamp, 53 hazard lamp, 54a tail lamp, 55a brake lamp, 60 wireless communication device, 61 transmitter, 62 receiver, 71 processing circuit, 72 memory, 73 processor, 100 alarm device.

Claims (11)

A driver seat identifying unit for performing a first image recognition process on an image obtained by photographing a vehicle with a camera provided in the host vehicle, and identifying a position of a driver seat of the other vehicle in the image;
When the position of the driver's seat cannot be specified by the driver's seat specifying unit, a second image recognition process different from the first image recognition process for the image is executed, and the position of the driver's seat in the image A driver seat estimation unit for estimating
A caution state determination unit that determines whether or not the driver of the other vehicle is in a forward careless state by executing a third image recognition process on a region including the driver seat in the image;
An alarm output unit that issues an alarm to the driver using an output device provided in the host vehicle, when the attention state determination unit determines that the driver is in a careless state ahead;
An alarm device comprising:   The driver seat identifying unit detects a passenger sitting in a front seat of the other vehicle by the first image recognition process, and when the number of passengers is one, the driver seat specifying unit The alarm device according to claim 1, wherein the seat on which the passenger sits is specified as the driver's seat.   The driver seat identifying unit detects the handle of the other vehicle by the first image recognition process, and identifies that the seat provided with the handle among the front seats of the other vehicle is the driver seat. The alarm device according to claim 1, wherein: The driver seat identification unit is
When a passenger sitting on the front seat of the other vehicle is detected by the first image recognition processing and the number of passengers is one, the seat on which the passenger sits is one of the front seats. Processing for identifying the driver's seat;
A process of detecting a handle of the other vehicle by the first image recognition process and specifying that the seat of the front seat provided with the handle is the driver seat;
The alarm device according to claim 1, wherein the position of the driver's seat in the image is specified by executing at least one of the processes.   The driver seat estimation unit detects an adult included in a passenger seated in a front seat of the other vehicle by the second image recognition process, and when the number of adults is one, the driver seat estimation unit The alarm device according to claim 1, wherein a seat on which the adult sits is estimated to be the driver's seat.   The driver seat estimation unit detects a fender lamp of the other vehicle by the second image recognition process, and estimates that a seat farther from the fender lamp among the front seats of the other vehicle is the driver seat. The alarm device according to claim 1.   The driver seat estimation unit detects a manufacturer emblem of the other vehicle by the second image recognition process, and whether or not the right-hand drive vehicle is manufactured by the vehicle manufacturer indicated by the maker emblem and the manufacture of the left-handle vehicle. The alarm device according to claim 1, wherein the position of the driver's seat in the image is estimated by acquiring information indicating presence / absence from a database.   The driver seat estimation unit detects the appearance of the other vehicle by the second image recognition process to identify the vehicle type of the other vehicle, and determines whether the right handle type and the left handle type are present in the vehicle type. The alarm device according to claim 1, wherein the position of the driver's seat in the image is estimated by acquiring information to be shown from a database. The driver seat estimation unit
When the second image recognition process detects an adult included in a passenger seated in the front seat of the other vehicle and the number of adults is one, the adult in the front seat sat down A process of estimating that the seat is the driver's seat;
A process of detecting a fender lamp of the other vehicle by the second image recognition process and estimating that a seat farther from the fender lamp of the front seats is the driver seat;
A maker emblem of the other vehicle is detected by the second image recognition process, and information indicating whether the right-hand drive vehicle is manufactured and whether the left-hand drive vehicle is manufactured by the vehicle manufacturer indicated by the maker emblem is stored from the database. A process of estimating the position of the driver's seat in the image by obtaining;
The appearance of the other vehicle is detected by the second image recognition process to identify the vehicle type of the other vehicle, and information indicating the presence / absence of the type of the right steering wheel and the type of the left steering wheel in the vehicle type is acquired from the database. Thus, processing for estimating the position of the driver's seat in the image,
The warning device according to claim 1, wherein the position of the driver's seat in the image is estimated by executing at least one of the processes.   When the driver's face is facing a different direction from the front of the other vehicle as a result of the third image recognition process, the caution state determination unit is different from the traveling direction of the other vehicle. When facing the direction, when the driver closes his eyes, when the driver is bowed, or when the driver is talking using a mobile terminal, the driver The alarm device according to claim 1, wherein the alarm device is determined to be in a forward careless state.   11. The alarm according to claim 1, wherein the alarm output unit uses a horn, a head lamp, a tail lamp, a brake lamp, or a hazard lamp of the host vehicle for the output device. apparatus.

Images