Crew detection device and crew detection method

Abstract

To provide an occupant detection device and an occupant detection method capable of precisely estimating an occupant's posture.SOLUTION: An occupant detection device includes: an acquisition part for acquiring point group information for showing a position of an object moving within a room as a point group comprising at least one detection point on a three-dimensional map corresponding to a space within the room on the basis of intensity of reflection waves generated by reflection of transmission waves transmitted toward the inside of the room of a moving body by an object existing in the room; a calculation part for calculating point group density for showing density of detection points for each of multiple regions of interest set so as to be separated mutually relative to one seating region on the basis of the point group information; and a posture estimation part for estimating a seating posture of an occupant who is seated in one seating region on the basis of point group density of each of multiple regions of interest corresponding to the one seating region.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] Embodiments of the present invention relate to an occupant detection device and an occupant detection method.

Background Art

[0002] In a vehicle control system or the like, a technique for estimating the posture and build of an occupant based on data acquired by a radio wave sensor or an image sensor installed in the vehicle interior is used.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] An embodiment of the present invention aims to provide an occupant detection device and an occupant detection method capable of accurately estimating the posture of an occupant.

Means for Solving the Problems

[0005] An occupant detection device according to an embodiment of the present invention includes an acquisition unit that acquires point group information indicating the position of an object moving in a room as a point group composed of one or more detection points on a three-dimensional map corresponding to the space in the room based on the intensity of a reflected wave generated by reflecting a transmitted wave transmitted toward the interior of a moving body by an object existing in the room, an arithmetic unit that calculates a point group density indicating the density of detection points for each of a plurality of regions of interest set so as to be separated from each other with respect to one seating region based on the point group information, and a posture estimation unit that estimates the seating posture of an occupant sitting in the seating region based on the point group density of each of the plurality of regions of interest corresponding to one seating region.

[0006] According to the above configuration, the occupant's posture is estimated based on the point cloud density of each of the multiple regions of interest set for a single seating area. This allows for highly accurate estimation of the occupant's posture.

[0007] Furthermore, in the above configuration, the multiple regions of interest may include a rear region set at the rear of the seating area and a front region set at the front of the seating area, and the posture estimation unit may detect the occupant's leaning posture based on the point cloud density of the rear region and the point cloud density of the front region.

[0008] With the above configuration, the forward-leaning posture of the occupant can be detected with high accuracy.

[0009] Furthermore, in the above configuration, the multiple regions of interest include a central region set in the center of the seating area, and at least one of the leftmost region set at the left end of the seating area or the rightmost region set at the right end of the seating area, and the posture estimation unit may detect a leaning posture in which the occupant is tilted to either the left or right based on the point cloud density of the central region and the point cloud density of the leftmost region, or the point cloud density of the central region and the point cloud density of the rightmost region.

[0010] With the above configuration, the occupant's leaning posture can be detected with high precision.

[0011] Furthermore, in the above configuration, the multiple regions of interest may include a left-end region set at the left end of the seating area and a right-end region set at the right end of the seating area, and the posture estimation unit may detect a leaning posture in which the occupant is tilted to either the left or right based on the point cloud density of the left-end region and the point cloud density of the right-end region.

[0012] With the above configuration, changes in the occupant's leaning posture (for example, a change from leaning left to leaning right, or from leaning right to leaning left) can be detected with high accuracy.

[0013] Furthermore, in the above configuration, the occupant detection device may further include a body size estimation unit that, based on the estimation results from the posture estimation unit, extracts normal point cloud information acquired when the occupant is in a normal posture from the point cloud information acquired by the acquisition unit, and estimates the occupant's body size based on the normal point cloud information.

[0014] With the above configuration, the physique of the crew can be estimated with high accuracy.

[0015] Furthermore, another embodiment of the present invention, a method for detecting an occupant, includes the steps of: acquiring point cloud information that indicates the position of an object moving in a room as a point cloud consisting of one or more detection points on a three-dimensional map corresponding to the space inside the room, based on the intensity of reflected waves generated when a transmitted wave sent toward the room of a moving object is reflected by an object present in the room; calculating a point cloud density that indicates the density of detection points for each of a plurality of regions of interest set to be spaced apart from each other with respect to a single seating area, based on the point cloud density of each of the plurality of regions of interest corresponding to a single seating area, and estimating the seating posture of an occupant seated in a seating area. [Brief explanation of the drawing]

[0016] [Figure 1] Figure 1 is a side view showing an example of the interior configuration of a vehicle equipped with the occupant detection device of the first embodiment. [Figure 2] Figure 2 is a block diagram showing an example of the hardware configuration of the occupant detection device according to the first embodiment. [Figure 3] Figure 3 is a block diagram showing an example of the functional configuration of the occupant detection device according to the first embodiment. [Figure 4] Figure 4 shows an example of point cloud information corresponding to a bird's-eye view of the vehicle interior in the first embodiment. [Figure 5] Figure 5 shows an example of point cloud information corresponding to a side view of the vehicle interior in the first embodiment. [Figure 6] Figure 6 shows an example of point cloud information corresponding to a top view of the vehicle interior in the first embodiment. [Figure 7]FIG. 7 is a diagram showing an example of point cloud information corresponding to a rear view of the passenger compartment of the first embodiment. [Figure 8] FIG. 8 is a diagram showing an example of a region of interest in the first embodiment. [Figure 9] FIG. 9 is a diagram showing an example of the relationship between the point cloud density and the posture of the occupant in the first embodiment. [Figure 10] FIG. 10 is a flowchart showing an example of processing in the occupant detection device of the first embodiment. [Figure 11] FIG. 11 is a diagram showing an example of a region of interest in the second embodiment. [Figure 12] FIG. 12 is a diagram showing an example of the relationship between the point cloud density and the posture of the occupant in the second embodiment. [Figure 13] FIG. 13 is a flowchart showing an example of processing in the occupant detection device of the second embodiment. [Figure 14] FIG. 14 is a diagram showing an example of a region of interest in the third embodiment. [Figure 15] FIG. 15 is a diagram showing an example of the relationship between the point cloud density and the posture of the occupant in the third embodiment. [Figure 16] FIG. 16 is a flowchart showing an example of processing in the occupant detection device of the third embodiment.

MODE FOR CARRYING OUT THE INVENTION

[0017] Hereinafter, exemplary embodiments of the present invention will be disclosed. The configurations of the embodiments shown below, as well as the actions, results, and effects brought about by the configurations, are examples. The present invention can be realized by configurations other than those disclosed in the following embodiments, and it is possible to obtain at least one of various effects and derivative effects based on the basic configuration.

[0018] (First Embodiment) Figure 1 is a side view showing an example of the configuration inside the passenger compartment R of a vehicle C in which the occupant detection device 1 of the first embodiment is installed. In the figure, the X direction corresponds to the direction from the rear to the front of the vehicle C, the Y direction corresponds to the direction from the left side to the right side of the vehicle C, and the Z direction corresponds to the direction from the bottom to the top of the vehicle C.

[0019] Vehicle C is an example of a mobile vehicle, and may be, for example, an automobile powered by an internal combustion engine, a motor, or both (hybrid mechanism). Vehicle C in this embodiment is a so-called three-row seat vehicle, and is equipped with a first seat row S1, a second seat row S2, and a third seat row S3. The first seat row S1 includes the driver's seat and the passenger seat.

[0020] The occupant detection device 1 is a device that acquires occupant information about occupants present in the vehicle compartment R. The occupant information includes information indicating the occupant's posture, physique, etc. In this embodiment, the case of acquiring occupant information about occupants located behind the first seat row S1, i.e., in the second seat row S2 and the third seat row S3, will be described.

[0021] The occupant detection device 1 comprises a sensor 2 and an information processing device 3. The sensor 2 is installed on the ceiling 5 of the vehicle C, transmits a wave toward the passenger compartment R, and receives reflected waves generated when the transmitted wave is reflected by an object present in the passenger compartment R. In this embodiment, the sensor 2 is installed behind the backrest of the first row of seats S1, and in the example shown in Figure 1, it is located approximately above the backrest of the second row of seats S2. The information processing device 3 is installed in the dashboard and is connected to the sensor 2 via a network such as CAN (Controller Area Network). Note that the installation locations of the sensor 2 and the information processing device 3 are not limited to those described above. Also, multiple sensors 2 may be installed.

[0022] Figure 2 is a block diagram showing an example of the hardware configuration of the occupant detection device 1 of the first embodiment. Sensor 2 comprises a transmitter 21, a receiver 22, an ECU (Electronic Control Unit) 23, and an input / output unit 24. Transmitter 21 is a device that transmits (irradiates) electromagnetic waves of a predetermined frequency (e.g., 60 GHz to 65 GHz, etc.) as a transmitted wave into the vehicle interior R. Receiver 22 is a device that receives reflected waves generated when the transmitted wave is reflected by an object present in the vehicle interior R, and generates an electrical signal indicating the intensity of the reflected wave. Transmitter 21 and receiver 22 can be configured using, for example, an oscillator circuit, a piezoelectric element, an AD converter, an amplifier, a filter circuit, etc. Transmitter 21 and receiver 22 may be configured as separate units or as an integrated unit. ECU 23 is a microcontroller configured using a CPU, memory, etc., and performs processing related to the control of transmitter 21 and receiver 22, and the generation of data based on the reflected wave received by receiver 22. The input / output unit 24 is an interface device that establishes communication with the information processing device 3 and other devices in accordance with a predetermined standard such as CAN.

[0023] The information processing device 3 comprises a CPU (Central Processing unit) 31, a memory 32, and an input / output unit 33. The CPU 31 executes various arithmetic processes according to the program stored in the memory 32. The memory 32 may be configured using appropriate volatile and non-volatile memory. The memory 32 stores programs that cause the CPU 31 to execute various processes to realize the functions of the occupant detection device 1, setting data, data acquired from the sensor 2, data generated by the CPU 31, etc. The input / output unit 33 is an interface device that establishes communication with the sensor 2 and other devices in accordance with predetermined standards such as CAN.

[0024] Note that the hardware configuration shown in Figure 2 is an example, and the hardware configuration of the occupant detection device 1 is not limited to that shown.

[0025] Figure 3 is a block diagram showing an example of the functional configuration of the occupant detection device 1 of the first embodiment. The occupant detection device 1 of this embodiment includes an acquisition unit 101, a calculation unit 102, a posture estimation unit 103, and a body size estimation unit 104. These functional units 101 to 104 can be configured by the cooperation of hardware and software (programs), for example, as illustrated in Figure 2. Furthermore, some or all of these functional units 101 to 104 may be configured by dedicated hardware (circuits, etc.) such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field-Programmable Gate Array).

[0026] The acquisition unit 101 acquires point cloud information that shows the position of an object moving within the vehicle interior R as a point cloud consisting of one or more detection points on a three-dimensional map corresponding to the space within the vehicle interior R, based on the intensity of the reflected wave received by the sensor 2.

[0027] Figure 4 shows an example of point cloud information 201 corresponding to a bird's-eye view of the vehicle compartment R in the first embodiment. Figure 5 shows an example of point cloud information 201 corresponding to a side view (XZ plan view) of the vehicle compartment R in the first embodiment. Figure 6 shows an example of point cloud information 201 corresponding to a top view (XY plan view) of the vehicle compartment R in the first embodiment. Figure 7 shows an example of point cloud information 201 corresponding to a rear view (YZ plan view) of the vehicle compartment R in the first embodiment.

[0028] The point cloud information 201 illustrated here is information plotted on a three-dimensional map (e.g., a voxel map) corresponding to the space inside the vehicle compartment R, with one or more points (e.g., voxels, etc.) corresponding to the position of a moving object (e.g., the area occupied by the moving part of the occupant's body 10) plotted as detection points P. The type of point to be used as a detection point P is determined appropriately according to the function of the sensor 2 used, but for example, a detection point P may be a point where the change in the intensity of the reflected wave per unit time is greater than a predetermined threshold.

[0029] Figures 4 to 7 illustrate a scenario where three seating areas R21 to R23 are set on the second seat row S2, and two seating areas R31 and R32 are set on the third seat row S3, with an occupant seated in the rightmost seating area R23 of the second seat row S2. In such a case, as shown in Figures 4 to 7, a point cloud G consisting of multiple detection points P corresponding to the occupant's head, chest, back, etc., appears in the space corresponding to the seating area R23 of the three-dimensional map. The point cloud information 201 includes information about such a point cloud G, that is, information indicating the position, number, etc., of the multiple detection points P that appear on the three-dimensional map.

[0030] Based on the point cloud information 201 described above, the calculation unit 102 (see Figure 3) calculates the point cloud density, which indicates the density of detected points P, for each of multiple regions of interest that are set to be spaced apart from one seating region (at least one of the seating regions R21~R23, R31, and R32).

[0031] Figure 8 shows an example of a region of interest in the first embodiment. Figure 8 illustrates a rear region Rb and a front region Rf as regions of interest set in the seating area R23 at the right end of the second seat row S2. The rear region Rb is a region set at the rear of the seating area R23, and in this example, it is set to include the space from the seat surface near the rear end of the seating area R23 up to a predetermined height (for example, a position higher than the upper end of the backrest). The front region Rf is a region set at the front of the seating area R23, and in this example, it is set to include the space from the seat surface near the front end of the seating area R23 up to a predetermined height. Such a rear region Rb and front region Rf are set for each seating area. Note that the method of setting the rear region Rb and front region Rf is not limited to the above and should be determined appropriately according to the seat shape, the characteristics of the sensor 2, etc.

[0032] Figure 9 shows an example of the relationship between the point cloud density and the posture of the occupant 10 in the first embodiment. As shown in Figure 9, when the occupant 10 is seated in a normal posture, the number of detected points P in the rear region Rb increases, and the number of detected points P in the front region Rf decreases. That is, when the occupant 10 is in a normal posture, the point cloud density in the rear region Rb increases, and the point cloud density in the front region Rf decreases. On the other hand, when the occupant 10 is in a hunched-over posture with the upper body bent forward, the number of detected points P in the front region Rf increases, and the number of detected points P in the rear region Rb decreases. That is, when the occupant 10 is in a hunched-over posture, the point cloud density in the front region Rf increases, and the point cloud density in the rear region Rb decreases. The calculation unit 102 of this embodiment calculates the point cloud density for the rear region Rb and the front region Rf as described above.

[0033] The posture estimation unit 103 (see Figure 3) estimates the seating posture of an occupant 10 seated in a seating area (e.g., seating area R23) based on the point cloud density of multiple regions of interest (in this embodiment, the rear region Rb and the front region Rf) corresponding to a single seating area (e.g., seating area R23), and generates posture information indicating the estimated seating posture. The posture estimation unit 103 in this embodiment detects the occupant 10's hunched-over posture based on the point cloud density of the rear region Rb and the point cloud density of the front region Rf. The method for determining whether or not the occupant is hunched over is not particularly limited, but for example, if the point cloud density of the rear region Rb decreases and the point cloud density of the front region Rf increases, it can be determined that the occupant 10 is hunched over. For example, if the difference value obtained by subtracting the point cloud density of the rear region Rb from the point cloud density of the front region Rf is greater than or equal to a threshold, it can be determined that the occupant 10 is hunched over.

[0034] The body size estimation unit 104 estimates the body size of the occupant 10 based on the point cloud information acquired by the acquisition unit 101 and the posture information generated by the posture estimation unit 103, and generates body size information indicating the estimated body size. The body size information is output to various ECUs for controlling the vehicle C and used for various purposes (e.g., seat belt reminder). In this embodiment, the body size estimation unit 104 extracts normal point cloud information, which is point cloud information acquired when the occupant 10 is in a normal posture, from the point cloud information acquired by the acquisition unit 101 based on the posture information, and estimates the body size of the occupant 10 based on the normal point cloud information. This improves the accuracy of body size estimation.

[0035] Figure 10 is a flowchart illustrating an example of processing in the occupant detection device 1 of the first embodiment. When the acquisition unit 101 acquires point cloud information for the rear region Rb and the front region Rf of a seating area (hereinafter referred to as seating area R23) (S101), the calculation unit 102 calculates the point cloud density ρb of the rear region Rb and the point cloud density ρf of the front region Rf (S102). Subsequently, the posture estimation unit 103 determines whether the difference value obtained by subtracting the point cloud density ρb of the rear region Rb from the point cloud density ρf of the front region Rf is greater than the threshold K1 (ρf-ρb>K1) (S103). If ρf-ρb>K1 (S103:Yes), the posture estimation unit 103 determines that the occupant 10 seated in seating area R23 is in a hunched-over posture (S104). Subsequently, the body size estimation unit 104 extracts normal point cloud information from the point cloud information of the seating area R23 acquired by the acquisition unit 101 when the occupant 10 was not in a hunched-over position (i.e., in a normal position) (S106), and estimates the occupant 10's body size based on the normal point cloud information (S107).

[0036] If ρf-ρb > K1 (S103: No), the posture estimation unit 103 determines that the occupant 10 seated in the seating area R23 is in a normal posture (S105), and the body size estimation unit 104 estimates the body size of the occupant 10 based on the current point cloud information (S107).

[0037] As described above, according to this embodiment, the posture of the occupant 10 is estimated based on the point cloud densities of the rear region Rb and the front region Rf, which are multiple regions of interest set for one seating region (at least one of the seating regions R21-R23, R31, and R32). This makes it possible to detect the occupant 10's hunched posture with high accuracy. Furthermore, since the occupant 10's physique is estimated based on the point cloud information acquired when the occupant 10 is in a normal posture, the occupant 10's physique can be estimated with high accuracy.

[0038] Other embodiments will be described below with reference to the drawings, but descriptions of parts that are the same as or similar to those in the first embodiment may be omitted.

[0039] (Second Embodiment) Figure 11 shows an example of a region of interest in the second embodiment. The region of interest in this embodiment includes the leftmost region Rl, the rightmost region Rr, and the central region Rc. The leftmost region Rl is set at the leftmost end of the seating areas R21-R23, R31, and R32. The rightmost region Rr is set at the rightmost end of the seating areas R21-R23, R31, and R32. The central region Rc is set in the center of the seating areas R21-R23, R31, and R32. In the example shown in Figure 11, the leftmost region Rl is the area outside the leftmost end of the seating area R23 and is set to include a part of the console 51. The rightmost region Rr is the area outside the rightmost end of the seating area R23 and is set to include the space from the rightmost end of the seating area R23 to the inner surface of the side door 52. The central region Rc is set to include the central part of the seating region R23, but not to include the left and right edges of the seating region R23. Note that the setting methods for the left edge region Rl, right edge region Rr, and central region Rc are not limited to those described above.

[0040] Figure 12 shows an example of the relationship between point cloud density and the posture of the occupant 10 in the second embodiment. As shown in Figure 12, when the occupant 10 is seated in a normal posture, the number of detected points P in the central region Rc increases, and the number of detected points P in the leftmost region Rl and the rightmost region Rr decreases. In other words, when the occupant 10 is in a normal posture, the point cloud density in the central region Rc increases, and the point cloud density in the leftmost region Rl and the rightmost region Rr decreases. When the occupant 10 is leaning to the left with their upper body tilted to the left and leaning against an object in the passenger compartment R (e.g., console 51), the number of detected points P in the leftmost region Rl increases, and the number of detected points P in the central region Rc decreases. When the occupant 10 is leaning to the right with their upper body tilted to the right and leaning against an object in the passenger compartment R (e.g., side door 52), the number of detected points P in the rightmost region Rr increases, and the number of detected points P in the central region Rc decreases.

[0041] The calculation unit 102 of this embodiment calculates the point cloud density for the leftmost region Rl, the rightmost region Rr, and the central region Rc, respectively, as described above. The attitude estimation unit 103 of this embodiment detects the left-leaning and right-leaning postures of the occupant 10 based on the point cloud density of the leftmost region Rl, the point cloud density of the rightmost region Rr, and the point cloud density of the central region Rc. The method for determining whether or not the occupant is in a leaning posture is not particularly limited, but for example, if the point cloud density of the central region Rc decreases and the point cloud density of the leftmost region Rl increases, it can be determined that the occupant 10 is in a left-leaning posture. For example, if the difference value obtained by subtracting the point cloud density of the central region Rc from the point cloud density of the leftmost region Rl is greater than or equal to a threshold, it can be determined that the occupant 10 is in a left-leaning posture. Alternatively, if the point cloud density of the central region Rc decreases and the point cloud density of the rightmost region Rr increases, it can be determined that the occupant 10 is in a right-leaning posture. For example, if the difference obtained by subtracting the point cloud density of the central region Rc from the point cloud density of the rightmost region Rr is greater than or equal to a threshold, it may be determined that the occupant 10 is in a right-leaning posture.

[0042] Figure 13 is a flowchart showing an example of processing in the occupant detection device 1 of the second embodiment. When the acquisition unit 101 acquires point cloud information for the leftmost region Rl, the rightmost region Rr, and the central region Rc of a seating area (hereinafter referred to as seating area R23) (S201), the calculation unit 102 calculates the point cloud density ρl of the leftmost region Rl, the point cloud density ρr of the rightmost region Rr, and the point cloud density ρc of the central region Rc (S202). Subsequently, the attitude estimation unit 103 determines whether the difference value obtained by subtracting the point cloud density ρc of the central region Rc from the point cloud density ρl of the leftmost region Rl is greater than the threshold K2 (ρl-ρc>K2) (S203). If ρl-ρc>K2 (S203:Yes), the attitude estimation unit 103 determines that the occupant 10 seated in seating area R23 is in a left-leaning posture (S204). Subsequently, the body size estimation unit 104 extracts normal point cloud information obtained when the occupant 10 was in a normal posture from the point cloud information of the seating area R23 acquired by the acquisition unit 101 (S205), and estimates the body size of the occupant 10 based on the normal point cloud information (S206).

[0043] If ρl-ρc is not > K2 (S203: No), the posture estimation unit 103 determines whether the difference obtained by subtracting the point cloud density ρc of the central region Rc from the point cloud density ρr of the rightmost region Rr is greater than the threshold K3 (ρr-ρc>K3) (S207). If ρr-ρc>K3 (S207: Yes), the posture estimation unit 103 determines that the occupant 10 seated in the seating region R23 is in a right-leaning posture (S208). Subsequently, the body size estimation unit 104 extracts normal point cloud information obtained when the occupant 10 was in a normal posture from the point cloud information of the seating region R23 obtained by the acquisition unit 101 (S205), and estimates the body size of the occupant 10 based on the normal point cloud information (S206).

[0044] If ρr-ρc is not > K3 (S207: No), the posture estimation unit 103 determines that the occupant 10 seated in the seating area R23 is in a normal posture (S209), and the body size estimation unit 104 estimates the body size of the occupant 10 based on the current point cloud information (S206).

[0045] As described above, according to this embodiment, the posture of the occupant 10 is estimated based on the point cloud densities of the leftmost region Rl, the rightmost region Rr, and the central region Rc, which are multiple regions of interest set for one seating area (at least one of the seating areas R21-R23, R31, and R32). This makes it possible to detect the occupant 10's left-leaning and right-leaning postures with high accuracy.

[0046] In the above example, the leftmost region Rl and the rightmost region Rr are set, but either the leftmost region Rl or the rightmost region Rr may be set alone. For example, if only the central region Rc and the leftmost region Rl are set, the leftward leaning posture of the occupant 10 can be detected based on the relationship between the point cloud density of the central region Rc and the point cloud density of the leftmost region Rl. Similarly, if only the central region Rc and the rightmost region Rr are set, the rightward leaning posture of the occupant 10 can be detected based on the relationship between the point cloud density of the central region Rc and the point cloud density of the rightmost region Rr.

[0047] (Third embodiment) Figure 14 shows an example of the region of interest in the third embodiment. The region of interest in this embodiment includes the leftmost region Rl and the rightmost region Rr. That is, it differs from the second embodiment in that the central region Rc is not set. The leftmost region Rl is set at the left end of the seating areas R21-R23, R31, and R32. The rightmost region Rr is set at the right end of the seating areas R21-R23, R31, and R32. In the example shown in Figure 14, the leftmost region Rl is the area outside the left end of the seating area R23 and is set to include a part of the console 51. The rightmost region Rr is the area outside the right end of the seating area R23 and is set to include the space from the right end of the seating area R23 to the inner surface of the side door 52. Note that the method of setting the leftmost region Rl and the rightmost region Rr is not limited to the above.

[0048] Figure 15 shows an example of the relationship between point cloud density and the posture of the occupant 10 in the third embodiment. As shown in Figure 15, when the occupant 10 is leaning to the left, the number of detected points P in the leftmost region Rl increases, and the number of detected points P in the rightmost region Rr decreases. When the occupant 10 is leaning to the right, the number of detected points P in the rightmost region Rr increases, and the number of detected points P in the leftmost region Rl decreases.

[0049] The calculation unit 102 of this embodiment calculates the point cloud density for the leftmost region Rl and the rightmost region Rr, respectively, as described above. The attitude estimation unit 103 of this embodiment detects the left-leaning and right-leaning attitudes of the occupant 10 based on the point cloud density of the leftmost region Rl and the point cloud density of the rightmost region Rr. For example, if the point cloud density of the rightmost region Rr decreases and the point cloud density of the leftmost region Rl increases, it can be determined that the attitude of the occupant 10 has changed from a right-leaning attitude to a left-leaning attitude. For example, if the difference value obtained by subtracting the point cloud density of the rightmost region Rr from the point cloud density of the leftmost region Rl is greater than or equal to a threshold, it can be determined that the occupant 10 is in a left-leaning attitude. Also, if the point cloud density of the leftmost region Rl decreases and the point cloud density of the rightmost region Rr increases, it can be determined that the attitude of the occupant 10 has changed from a left-leaning attitude to a right-leaning attitude. For example, if the difference obtained by subtracting the point cloud density of the leftmost region Rl from the point cloud density of the rightmost region Rr is greater than or equal to a threshold, it may be determined that the occupant 10 is in a right-leaning posture. Alternatively, if the occupant 10 is neither in a left-leaning nor right-leaning posture, it may be determined that the occupant 10 is in a normal posture.

[0050] Figure 16 is a flowchart showing an example of processing in the occupant detection device 1 of the third embodiment. When the acquisition unit 101 acquires point cloud information of the leftmost region Rl and the rightmost region Rr of a seating area (hereinafter referred to as seating area R23) (S301), the calculation unit 102 calculates the point cloud density ρl of the leftmost region Rl and the point cloud density ρr of the rightmost region Rr (S302). Subsequently, the attitude estimation unit 103 determines whether the difference value obtained by subtracting the point cloud density ρl of the leftmost region Rl from the point cloud density ρr of the rightmost region Rr is greater than the threshold K4 (ρr-ρl>K4) (S303). If ρr-ρl>K4 (S303:Yes), the attitude estimation unit 103 determines that the occupant 10 seated in seating area R23 has changed from a left-leaning posture to a right-leaning posture (S304). Subsequently, the body size estimation unit 104 extracts normal point cloud information obtained when the occupant 10 was in a normal posture from the point cloud information of the seating area R23 acquired by the acquisition unit 101 (S305), and estimates the body size of the occupant 10 based on the normal point cloud information (S306).

[0051] If ρr-ρl is not > K4 (S303: No), the posture estimation unit 103 determines whether the difference obtained by subtracting the point cloud density ρr of the rightmost region Rr from the point cloud density ρl of the leftmost region Rl is greater than the threshold K5 (ρl-ρr>K5) (S307). If ρl-ρr>K5 (S307: Yes), the posture estimation unit 103 determines that the occupant 10 seated in the seating region R23 has changed from a right-leaning posture to a left-leaning posture (S308). Subsequently, the body size estimation unit 104 extracts normal point cloud information obtained when the occupant 10 was in a normal posture from the point cloud information of the seating region R23 obtained by the acquisition unit 101 (S305), and estimates the body size of the occupant 10 based on the normal point cloud information (S306).

[0052] If ρl-ρr > K5 (S307: No), the posture estimation unit 103 determines that the occupant 10 seated in the seating area R23 is in a normal posture (S309), and the body size estimation unit 104 estimates the body size of the occupant 10 based on the current point cloud information (S306).

[0053] As described above, according to this embodiment, the posture of the occupant 10 is estimated based on the point cloud density of the leftmost region Rl and the rightmost region Rr, which are set as regions of interest for each seating region R21-R23, R31, and R32. This makes it possible to detect the occupant 10's left-leaning and right-leaning postures with high accuracy.

[0054] The program that causes a computer (information processing device 3) to execute the processing necessary to realize the functions of the occupant detection device 1 as described above may be provided as a computer program product by being stored in an installable or executable file format on a computer-readable storage medium such as a CD-ROM, CD-R, memory card, DVD (Digital Versatile Disk), or flexible disk (FD). Alternatively, the program may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Furthermore, the program may be provided or distributed via a network such as the Internet.

[0055] Although embodiments of the present invention have been described above, these embodiments are presented as examples only and are not intended to limit the scope of the invention. This novel embodiment can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. This embodiment and its variations are included in the scope and spirit of the invention, as well as in the claims of the invention and its equivalents. [Explanation of Symbols]

[0056] 1... Occupant detection device, 2... Sensor, 3... Information processing device, 5... Ceiling, 10... Occupant, 21... Transmitter, 22... Receiver, 23... ECU, 24... Input / Output unit, 31... CPU, 32... Memory, 33... Input / Output unit, 101... Acquisition unit, 102... Calculation unit, 103... Posture estimation unit, 104... Body size estimation unit, 201... Point cloud information, C... Vehicle, G... Point cloud, P... Detection point, R... Passenger compartment, R21~R23, R31, R32... Seating area, Rb... Rear area, Rc... Central area, Rf... Front area, Rl... Left end area, Rr... Right end area, S1... First seat row, S2... Second seat row, S3... Third seat row

Claims

1. An acquisition unit acquires point cloud information that indicates the position of an object moving in a room as a point cloud consisting of one or more detection points on a three-dimensional map corresponding to the space of the room, based on the intensity of the reflected wave generated when a transmitted wave sent toward the room of a moving object is reflected by an object present in the room. A calculation unit calculates the point cloud density, which indicates the density of detected points, for each of a plurality of regions of interest set to be spaced apart from one seating region, based on the point cloud information. A posture estimation unit that estimates the seating posture of an occupant seated in a seating area based on the point cloud density of each of the multiple areas of interest corresponding to one seating area, Equipped with, The multiple regions of interest include a left-end region set at the left end of the seating area and a right-end region set at the right end of the seating area. The posture estimation unit detects, based on the point cloud density of the leftmost region and the point cloud density of the rightmost region, that the occupant has changed from a left-leaning posture to a right-leaning posture, or from a right-leaning posture to a left-leaning posture, and is an occupant detection device.

2. Based on the estimation results by the posture estimation unit, a body size estimation unit extracts normal point cloud information acquired when the occupant is in a normal posture from the point cloud information acquired by the acquisition unit, and estimates the occupant's body size based on the normal point cloud information. The occupant detection device according to claim 1, further comprising:

3. A step of acquiring point cloud information that indicates the position of an object moving in a room as a point cloud consisting of one or more detection points on a three-dimensional map corresponding to the space of the room, based on the intensity of the reflected wave generated when a transmitted wave sent toward the room of a moving object is reflected by an object present in the room, A step of calculating the point cloud density, which indicates the density of detected points, for each of a plurality of regions of interest set to be spaced apart from each other, including a left-end region set at the left end of a seating region and a right-end region set at the right end of the seating region, based on the point cloud information; A step of detecting, based on the point cloud density of each of the multiple regions of interest corresponding to one of the seating regions, that an occupant seated in the seating region has changed from a left-leaning posture to a right-leaning posture, or from a right-leaning posture to a left-leaning posture, Crew detection method including

Images