Image recognition device, image recognition method and program

Abstract

To provide an image recognition device which improves image recognition accuracy while suppressing time and cost required for image recognition.SOLUTION: An image recognition device 15 comprises: an acquisition part 3A which acquires a camera image indicating outside of an own vehicle 1; a determination part 3B which determines a type of a target included in the camera image to calculate a reliability score of the type of the target; a recognition part 3C which recognizes, when the reliability score is a score threshold value or more, the type of the target determined by the determination part 3B is included in the camera image; a threshold value change part 3D which changes the score threshold value; and a base station number calculation part 3E which calculates number of radio communication base stations transmitting a signal that can be received by a radio communication apparatus 13 mounted on an own vehicle 1 where signal strength thereof is a signal intensity threshold value or more. For example, the threshold value change part 3D decreases, when the number of the radio communication base stations is the base station number threshold value or less, the score threshold value that is compared with the reliability score of an animal except a human being relative to a case that the number of the radio communication base stations is larger than the base station number threshold value.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an image recognition device, an image recognition method, and a program. [Background technology]

[0002] Patent Document 1 describes a technology for performing image recognition in response to changes in environmental conditions around a detection target object. The technology described in Patent Document 1 performs image recognition on an input image by assigning weights according to environmental conditions determined from the external environment. Furthermore, in the technology described in Patent Document 1, the external environment includes urban areas, suburban areas, mountainous areas, etc. Furthermore, in order to acquire information about the external environment, the technology described in Patent Document 1 uses a communication device that receives regional information provided by an external organization, an input device that inputs map information stored in a map database, etc. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-178736 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in Patent Document 1 performs image recognition in response to changes in environmental conditions because the visibility of the detected object changes depending on the environmental conditions. In other words, the technology described in Patent Document 1 does not incorporate a concept of avoiding the time and cost required to determine whether a camera image taken in a mountainous area contains a pedestrian, since pedestrians are rarely present in the mountainous area, or avoiding the time and cost required to determine whether a camera image taken in a human-populated area contains an animal other than a human, since the probability of an animal other than a human being appearing in a human-populated area is low. Therefore, the technology described in Patent Document 1 cannot improve image recognition accuracy while reducing the time and cost required for image recognition.

[0005] In view of the above, an object of the present disclosure is to provide an image recognition device, an image recognition method, and a program that can improve image recognition accuracy while reducing the time and cost required for image recognition. is. [Means for solving the problem]

[0006] (1) One aspect of the present disclosure includes an acquisition unit that acquires a camera image showing the outside of a vehicle captured by a camera; a determination unit that determines a type of a target included in the camera image and calculates a reliability score for the type of target; a recognition unit that recognizes that a target of the type determined by the determination unit is included in the camera image when the reliability score calculated by the determination unit is equal to or greater than a score threshold; a threshold change unit that changes the score threshold; and a base station number calculation unit that calculates the number of wireless communication base stations that transmit signals that can be received by a wireless communication device mounted on the vehicle with a signal strength equal to or greater than a signal strength threshold, wherein the types of targets determined by the determination unit include at least pedestrians, and animals excluding humans, and the threshold change unit performs at least one of: when the number of wireless communication base stations calculated by the base station number calculation unit is equal to or less than a base station number threshold, reducing the score threshold to be compared with the reliability score of the animals excluding humans, more than when the number of wireless communication base stations calculated by the base station number calculation unit is greater than the base station number threshold; and when the number of wireless communication base stations calculated by the base station number calculation unit is greater than the base station number threshold, reducing the score threshold to be compared with the reliability score of the pedestrian, more than when the number of wireless communication base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold.

[0007] (2) In the image recognition device of (1), the types of targets determined by the determination unit may include cliffs, and the threshold change unit may reduce the score threshold to be compared with the reliability score of the cliff when the number of wireless communication base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold, more than when the number of wireless communication base stations calculated by the base station number calculation unit is greater than the base station number threshold.

[0008] (3) In the image recognition device of (1), the wireless communication base stations include WiFi base stations and Bluetooth (registered trademark) base stations, and the threshold change unit may perform at least one of: when the number of Bluetooth base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold, reducing the score threshold to be compared with the reliability score of the animal excluding humans more than when the number of Bluetooth base stations calculated by the base station number calculation unit is greater than the base station number threshold; and when the number of Bluetooth base stations calculated by the base station number calculation unit is greater than the base station number threshold, reducing the score threshold to be compared with the reliability score of the pedestrian more than when the number of Bluetooth base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold.

[0009] (4) One aspect of the present disclosure includes an acquisition step in which an image recognition device acquires a camera image showing the outside of the vehicle captured by a camera; a determination step in which the image recognition device determines a type of target included in the camera image and calculates a reliability score for the type of target; a recognition step in which the image recognition device recognizes that the camera image includes a target of the type determined in the determination step if the reliability score calculated in the determination step is equal to or greater than a score threshold; a threshold change step in which the image recognition device changes the score threshold; and a base station number calculation step in which the image recognition device calculates the number of wireless communication base stations that transmit signals that can be received by a wireless communication device mounted on the vehicle with a signal strength equal to or greater than a signal strength threshold, and in the threshold changing step, when the number of wireless communication base stations calculated in the number of base stations calculation step is equal to or less than a base station number threshold, at least one of reducing the score threshold to be compared with the reliability score of the animals excluding humans more than a case where the number of wireless communication base stations calculated in the number of base stations calculation step is greater than the base station number threshold and reducing the score threshold to be compared with the reliability score of the pedestrian more than a case where the number of wireless communication base stations calculated in the number of base stations calculation step is equal to or less than the base station number threshold.

[0010] (5) One aspect of the present disclosure is a program for causing a processor to execute an acquisition step of acquiring a camera image showing the outside of a vehicle captured by a camera, a determination step of determining a type of target included in the camera image and calculating a reliability score for the type of target, a recognition step of recognizing that a target of the type determined in the determination step is included in the camera image if the reliability score calculated in the determination step is equal to or greater than a score threshold, a threshold change step of changing the score threshold, and a base station number calculation step of calculating the number of wireless communication base stations that transmit signals that can be received by a wireless communication device mounted on the vehicle with a signal strength equal to or greater than a signal strength threshold, The types include at least pedestrians and animals excluding humans, and the threshold change step executes at least one of: when the number of wireless communication base stations calculated in the base station number calculation step is equal to or less than a base station number threshold, reducing the score threshold to be compared with the reliability score of the animals excluding humans more than when the number of wireless communication base stations calculated in the base station number calculation step is greater than the base station number threshold; and when the number of wireless communication base stations calculated in the base station number calculation step is greater than the base station number threshold, reducing the score threshold to be compared with the reliability score of the pedestrian more than when the number of wireless communication base stations calculated in the base station number calculation step is equal to or less than the base station number threshold. [Effects of the Invention]

[0011] According to the present disclosure, it is possible to improve the accuracy of image recognition while reducing the time and cost required for image recognition. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram showing an example of a host vehicle to which an image recognition device according to a first embodiment is applied; [Figure 2] FIG. 2 is a diagram illustrating an example of a relationship between a vehicle and a wireless communication base station. [Figure 3]FIG. 10 is a diagram showing an example of a learning camera image used for learning a model used by a determination unit. [Figure 4] 4 is a flowchart illustrating an example of processing executed by a processor of the image recognition device according to the first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of an image recognition device, an image recognition method, and a program according to the present disclosure will be described with reference to the drawings.

[0014] First Embodiment 1 is a diagram showing an example of a host vehicle 1 to which an image recognition device 15 of the first embodiment is applied. In the example shown in FIG. 1, the host vehicle 1 includes a camera 11, an HMI (Human Machine Interface) 12, a wireless communication device 13, a vehicle control device 14, a steering actuator 14A, a braking actuator 14B, a drive actuator 14C, and the image recognition device 15. The camera 11 captures an image of the outside of the vehicle 1 and transmits the camera image showing the outside of the vehicle 1 to the image recognition device 15. The HMI 12 has functions such as accepting various operations by the driver of the vehicle 1, and transmits a signal indicating the operation of the driver of the vehicle 1 to the vehicle control device 14. The wireless communication device 13 communicates with wireless communication base stations WB1, ..., WBN (see Figure 2) outside the vehicle 1.

[0015] Fig. 2 is a diagram showing an example of the relationship between the vehicle 1 and the wireless communication base stations WB1, ..., WBN. In detail, Fig. 2 shows an example of the relationship between the vehicle 1 traveling in an urban area and the wireless communication base stations WB1, ..., WBN present around the vehicle 1. In the example shown in Fig. 2, the wireless communication device 13 of the vehicle 1 receives signals (radio waves) transmitted from WiFi base stations (WiFi access points) serving as wireless communication base stations WB1 and WB3 at signal strengths equal to or greater than a signal strength threshold. The wireless communication device 13 also receives signals (radio waves) transmitted from Bluetooth (registered trademark) base stations (BLE (Bluetooth Low Energy) icons) serving as wireless communication base stations WB2, WB4, and WBN at signal strengths equal to or greater than a signal strength threshold. Furthermore, the wireless communication device 13 also receives signals (radio waves) transmitted from a mobile phone base station serving as wireless communication base station WB5 at signal strengths equal to or greater than a signal strength threshold. In the example shown in FIG. 2, the wireless communication device 13 does not receive a signal (WiFi signal, Bluetooth signal, etc.) transmitted from a mobile terminal such as a smartphone, a laptop computer, a game console, etc., but in other examples, the wireless communication device 13 may receive a signal transmitted from a mobile terminal as a signal transmitted from a wireless communication base station. 2, where the vehicle 1 is traveling in an urban area, the wireless communication device 13 of the vehicle 1 receives signals transmitted from a large number of wireless communication base stations WB1, ..., WBN at signal strengths equal to or greater than the signal strength threshold. On the other hand, when the vehicle 1 is traveling in a mountainous area, the number of signals transmitted from the wireless communication base stations that the wireless communication device 13 of the vehicle 1 receives at signal strengths equal to or greater than the signal strength threshold becomes smaller than in the example shown in FIG.

[0016] In the example shown in FIG. 1, the vehicle control device 14 controls the steering actuator 14A, the braking actuator 14B, and the drive actuator 14C based on a signal transmitted from the HMI 12 (a signal indicating the operation of the driver of the vehicle 1), the results of image recognition performed by the image recognition device 15 (described later), and the like. 1, the vehicle control device 14 has a driving assistance function. Specifically, when a camera image including a pedestrian crossing a road on which the host vehicle 1 is traveling is transmitted from the camera 11 to the image recognition device 15, the vehicle control device 14 has a function to operate the brake actuator 14B based on the result of image recognition by the image recognition device 15 in order to avoid a collision between the host vehicle 1 and the pedestrian, without the driver of the host vehicle 1 having to perform an operation to operate the brake actuator 14B. In another example, the vehicle control device 14 may have an automatic driving function that controls the steering actuator 14A, braking actuator 14B, and drive actuator 14C based on a driving plan, the results of image recognition by the image recognition device 15, etc., without the need for operation by the driver of the vehicle 1, and causes the vehicle 1 to drive autonomously.

[0017] 1, the image recognition device 15 is configured by a microcomputer including a communication interface (I / F) 151, a memory 152, and a processor 153. The communication interface 151 has an interface circuit for connecting the image recognition device 15 to the camera 11, the HMI 12, the wireless communication device 13, and the vehicle control device 14. The memory 152 stores programs and various data used in the processing executed by the processor 153. The processor 153 has a function as an acquisition unit 3A, a function as a determination unit 3B, a function as a recognition unit 3C, a function as a threshold value change unit 3D, and a function as a base station number calculation unit 3E. The acquisition unit 3A acquires camera images showing the outside of the host vehicle 1 captured by the camera 11. The determination unit 3B determines the type of target included in the camera images acquired by the acquisition unit 3A and calculates a reliability score for the target type. The determination unit 3B determines the type of target included in the camera images acquired by the acquisition unit 3A by using a model obtained by learning using training data, which is a data set of training camera images captured by a camera mounted on a training vehicle and labels indicating the type of target included in the training camera images. The recognition unit 3C recognizes that a target of the type determined by the determination unit 3B is included in the camera images acquired by the acquisition unit 3A when the reliability score calculated by the determination unit 3B is equal to or greater than a score threshold.

[0018] 3A and 3B are diagrams showing examples of training camera images used in training a model used by the determination unit 3B. Specifically, FIG. 3A shows an example of a training camera image used in training a model that enables the determination unit 3B to determine that the type of a target included in a camera image acquired by the acquisition unit 3A is a pedestrian (specifically, a pedestrian crossing the road on which the host vehicle 1 is traveling). FIG. 3B shows an example of a training camera image used in training a model that enables the determination unit 3B to determine that the type of a target included in a camera image acquired by the acquisition unit 3A is an animal other than a human (specifically, an animal other than a human crossing the road on which the host vehicle 1 is traveling). FIG. 3C shows an example of a training camera image used in training a model that enables the determination unit 3B to determine that the type of a target included in a camera image acquired by the acquisition unit 3A is a cliff (specifically, a cliff beside the road on which the host vehicle 1 is traveling). A training camera image such as that shown in Figure 3(A) (a camera image including a pedestrian crossing a road on which a training vehicle is traveling) can be obtained relatively easily, for example, by driving a training vehicle through an urban area.

[0019] On the other hand, training camera images such as those shown in FIG. 3(B) (training camera images including animals other than humans crossing a road on which the training vehicle is traveling) cannot be easily obtained, for example, when the training vehicle travels through a mountainous area. Furthermore, in order for the determination unit 3B to determine that the type of target included in the camera image acquired by the acquisition unit 3A is an animal other than humans as shown in FIG. 3(B) and to calculate a high reliability score as the reliability score for the type of target (animal other than humans), it is necessary to prepare a large number of training camera images such as those shown in FIG. 3(B) for model training. Preparing a large number of training camera images such as those shown in FIG. 3(B) requires the training vehicle to travel a long distance, which increases costs. On the other hand, if the determination unit 3B calculates a relatively low reliability score as the reliability score for the target type (animals excluding humans) and the reliability score is less than the score threshold, the recognition unit 3C will not recognize that the target type (animals excluding humans) determined by the determination unit 3B is included in the camera image, and it will take some time for the image recognition results of the image recognition device 15 to be obtained. Furthermore, a training camera image such as that shown in FIG. 3(C) (a training camera image including a cliff beside the road on which the training vehicle is traveling) cannot be obtained by traveling in an urban area, but can only be obtained by traveling in a mountainous area. Furthermore, in order for the determination unit 3B to determine that the type of target included in the camera image acquired by the acquisition unit 3A is a cliff such as that shown in FIG. 3(C) and to calculate a high reliability score as the reliability score for the type of target (cliff), it is necessary to prepare a large number of training camera images such as those shown in FIG. 3(C) for model training. Preparing a large number of training camera images such as those shown in FIG. 3(C) requires the training vehicle to travel a long distance, which increases costs. In view of the above points, the image recognition device 15 of the first embodiment has the following measures implemented, utilizing the above-mentioned property that when the vehicle 1 is traveling in a mountainous area, the number of signals transmitted from a wireless communication base station that the wireless communication device 13 of the vehicle 1 receives with a signal strength equal to or greater than the signal strength threshold is less than in the example shown in Figure 2.

[0020] 1, the threshold change unit 3D has a function of changing the score threshold. The base station number calculation unit 3E calculates the number of wireless communication base stations WB1, ..., WBN (see FIG. 2) that transmit signals that can be received by the wireless communication device 13 with a signal strength equal to or greater than the signal strength threshold.

[0021] In detail, in the example shown in FIG. 1, when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (for example, when the vehicle 1 is traveling in a mountainous area), the threshold change unit 3D reduces the score threshold to be compared with the reliability score of animals excluding humans (i.e., makes it easier for the recognition unit 3C to recognize that animals excluding humans are included in the camera image) compared to when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is greater than the base station number threshold (for example, when the vehicle 1 is traveling in an urban area). In addition, when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (for example, when the vehicle 1 is traveling in a mountainous area), the threshold change unit 3D reduces the score threshold to be compared with the cliff reliability score (i.e., makes it easier for the recognition unit 3C to recognize that a cliff is included in the camera image) compared to when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is greater than the base station number threshold (for example, when the vehicle 1 is traveling in an urban area). Furthermore, when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (for example, when the vehicle 1 is traveling in a mountainous area), the threshold change unit 3D increases the score threshold to be compared with the pedestrian's confidence score (i.e., makes it less likely that the recognition unit 3C will recognize a pedestrian if it is included in the camera image) compared to when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is greater than the base station number threshold (for example, when the vehicle 1 is traveling in an urban area).

[0022] Furthermore, in the example shown in Figure 1, when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is greater than the base station number threshold (for example, when the vehicle 1 is traveling in an urban area), the threshold change unit 3D increases the score threshold to be compared with the reliability score of animals excluding humans (in other words, making it less likely that the recognition unit 3C will recognize that animals excluding humans are included in the camera image) compared to when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (for example, when the vehicle 1 is traveling in a mountainous area). In addition, when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is greater than the base station number threshold (for example, when the vehicle 1 is traveling in an urban area), the threshold change unit 3D increases the score threshold to be compared with the cliff reliability score (i.e., makes it less likely that the recognition unit 3C will recognize that a cliff is included in the camera image) compared to when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (for example, when the vehicle 1 is traveling in a mountainous area). Furthermore, when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is greater than the base station number threshold (for example, when the vehicle 1 is traveling in an urban area), the threshold change unit 3D reduces the score threshold to be compared with the pedestrian's confidence score (i.e., makes it easier for the recognition unit 3C to recognize that a pedestrian is included in the camera image) compared to when the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (for example, when the vehicle 1 is traveling in a mountainous area).

[0023] 1, the threshold change unit 3D changes the score threshold to be compared with the reliability scores of animals excluding humans, changes the score threshold to be compared with the reliability scores of cliffs, and changes the score threshold to be compared with the reliability scores of pedestrians, based on the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E. Therefore, in the example shown in Fig. 1, it is possible to improve the accuracy of image recognition while suppressing the time and cost required for image recognition.

[0024] FIG. 4 is a flowchart illustrating an example of processing executed by the processor 153 of the image recognition device 15 according to the first embodiment. In the example shown in FIG. 4, the acquisition unit 3A acquires a camera image showing the outside of the vehicle 1 captured by the camera 11 in step S10. In step S11, the base station number calculation unit 3E calculates the number of wireless communication base stations WB1, . . . , WBN that transmit signals that can be received by the wireless communication device 13 with a signal strength equal to or greater than the signal strength threshold. In step S12, for example, the threshold value changing unit 3D determines whether the number of wireless communication base stations WB1, ..., WBN calculated in step S12 is greater than the base station number threshold value. If YES, proceed to step S13, and if NO, proceed to step S16.

[0025] In step S13, the threshold change unit 3D reduces the score threshold to be compared with the pedestrian's reliability score compared to when the number of wireless communication base stations WB1, . . . , WBN is equal to or less than the base station number threshold. In step S14, the threshold change unit 3D increases the score threshold to be compared with the reliability scores of animals excluding humans, compared to when the number of wireless communication base stations WB1, . . . , WBN is equal to or less than the base station number threshold. In step S15, the threshold change unit 3D increases the score threshold to be compared with the cliff reliability score, compared to when the number of wireless communication base stations WB1, ..., WBN is equal to or less than the base station number threshold. Then, the process proceeds to step S19.

[0026] In step S16, the threshold change unit 3D increases the score threshold to be compared with the pedestrian's reliability score more than when the number of wireless communication base stations WB1, . . . , WBN is greater than the base station number threshold. In step S17, the threshold change unit 3D reduces the score threshold to be compared with the reliability scores of animals other than humans, compared to when the number of wireless communication base stations WB1, . . . , WBN is greater than the base station number threshold. In step S18, the threshold change unit 3D reduces the score threshold to be compared with the cliff reliability score when the number of wireless communication base stations WB1, ..., WBN is greater than the base station number threshold. Then, the process proceeds to step S19.

[0027] In step S19, the determination unit 3B determines the type of the target included in the camera image acquired in step S10. In step S20, the determination unit 3B calculates a reliability score for the type of target. In step S21, for example, the recognition unit 3C determines whether the reliability score calculated in step S20 is equal to or greater than a score threshold value. If YES, the process proceeds to step S22, and if NO, the process shown in FIG. 4 ends. In step S22, the recognition unit 3C recognizes that the target of the type determined in step S19 is included in the camera image acquired in step S10.

[0028] 1, the base station number calculation unit 3E calculates the number of wireless communication base stations WB1, ..., WBN (more specifically, the total number of WiFi base stations, Bluetooth base stations, and mobile phone base stations) that transmit signals that can be received with a signal strength equal to or greater than the signal strength threshold in the wireless communication device 13. Furthermore, the threshold change unit 3D changes the score threshold based on the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E. In another example, emphasis is placed on the number of Bluetooth base stations (WB2, WB4, WBN) among the wireless communication base stations WB1, ..., WBN. In detail, in this example, the base station number calculation unit 3E calculates the number of Bluetooth base stations that transmit signals that can be received by the wireless communication device 13 with a signal strength equal to or greater than the signal strength threshold. Furthermore, the threshold change unit 3D changes the score threshold based on the number of Bluetooth base stations calculated by the base station number calculation unit 3E. Specifically, when the number of Bluetooth base stations calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold, the threshold change unit 3D reduces the score threshold to be compared with the reliability score of animals excluding humans compared to when the number of Bluetooth base stations calculated by the base station number calculation unit 3E is greater than the base station number threshold (i.e., makes it easier to recognize that animals excluding humans are included in the camera image). Also, when the number of Bluetooth base stations calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold, the threshold change unit 3D reduces the score threshold to be compared with the reliability score of cliffs compared to when the number of Bluetooth base stations calculated by the base station number calculation unit 3E is greater than the base station number threshold (i.e., makes it easier to recognize that cliffs are included in the camera image). Furthermore, when the number of Bluetooth base stations calculated by the base station number calculation unit 3E is greater than the base station number threshold, the threshold change unit 3D reduces the score threshold to be compared with the reliability score of pedestrians compared to when the number of Bluetooth base stations calculated by the base station number calculation unit 3E is equal to or less than the base station number threshold (i.e., makes it easier to recognize that pedestrians are included in the camera image).

[0029] In yet another example, the number of wireless communication base stations WB1, ..., WBN calculated by the base station number calculation unit 3E may not include the number of wireless communication base stations (e.g., smartphones, laptops, game consoles, etc. that transmit Bluetooth signals, WiFi signals, etc.) present inside the vehicle 1 (e.g., carried by passengers of the vehicle 1). A signal whose received signal strength by the wireless communication device 13 does not change while the vehicle 1 is traveling at a certain speed or above for a certain period of time or more can be considered to be a signal transmitted from a wireless communication base station present inside the vehicle 1.

[0030] Second Embodiment The vehicle 1 to which the image recognition device 15 of the second embodiment is applied is configured in the same manner as the vehicle 1 to which the image recognition device 15 of the first embodiment described above is applied, except for the points described below.

[0031] As described above, in one example of processing executed by the processor 153 of the image recognition device 15 of the first embodiment (the example shown in FIG. 4), in step S16, the threshold change unit 3D increases the score threshold to be compared with the reliability score of pedestrians more than when the number of wireless communication base stations WB1, ..., WBN is greater than the base station number threshold, and in step S17, the threshold change unit 3D decreases the score threshold to be compared with the reliability score of animals excluding humans more than when the number of wireless communication base stations WB1, ..., WBN is greater than the base station number threshold. In the host vehicle 1 to which the image recognition device 15 of the second embodiment is applied, step S17 may be executed without executing step S16.

[0032] Furthermore, in one example of processing executed by the processor 153 of the image recognition device 15 of the first embodiment as described above (the example shown in FIG. 4), in step S13, the threshold change unit 3D reduces the score threshold to be compared with the reliability score of pedestrians compared to when the number of wireless communication base stations WB1, ..., WBN is equal to or less than the base station number threshold, and in step S14, the threshold change unit 3D increases the score threshold to be compared with the reliability score of animals excluding humans compared to when the number of wireless communication base stations WB1, ..., WBN is equal to or less than the base station number threshold. In the host vehicle 1 to which the image recognition device 15 of the second embodiment is applied, step S13 may be executed without executing step S14.

[0033] <Third embodiment> The vehicle 1 to which the image recognition device 15 of the third embodiment is applied is configured in the same manner as the vehicle 1 to which the image recognition device 15 of the first embodiment described above is applied, except for the points described below.

[0034] As described above, in an example of processing executed by the processor 153 of the image recognition device 15 of the first embodiment (the example shown in FIG. 4), if the determination in step S12 is YES, steps S13, S14, and S15 are executed, and if the determination is NO, steps S16, S17, and S18 are executed. In a host vehicle 1 to which the image recognition device 15 of the third embodiment is applied, if the judgment in step S12 is YES, only step S14 is executed (steps S13 and S15 are not executed), and if the judgment is NO, only step S17 is executed (steps S16 and S18 may not be executed).

[0035] <Fourth embodiment> The vehicle 1 to which the image recognition device 15 of the fourth embodiment is applied is configured in the same manner as the vehicle 1 to which the image recognition device 15 of the first embodiment described above is applied, except for the points described below.

[0036] In a host vehicle 1 to which the image recognition device 15 of the fourth embodiment is applied, if the judgment in step S12 is YES, only step S13 is executed (steps S14 and S15 are not executed), and if the judgment is NO, only step S16 is executed (steps S17 and S18 may not be executed).

[0037] As described above, embodiments of the image recognition device, image recognition method, and program of the present disclosure have been described with reference to the drawings. However, the image recognition device, image recognition method, and program of the present disclosure are not limited to the above-described embodiments and may be modified as appropriate without departing from the spirit and scope of the present disclosure. The configurations of the above-described embodiments may be combined as appropriate. In the above-described embodiments, the processing performed by the image recognition device 15 has been described as software processing performed by executing a program. However, the processing performed by the image recognition device 15 may also be processing performed by hardware. Alternatively, the processing performed by the image recognition device 15 may be processing that combines both software and hardware. Furthermore, the program stored in the memory 152 of the image recognition device 15 (the program that realizes the functions of the processor 153 of the image recognition device 15) may be recorded on a computer-readable storage medium such as a semiconductor memory, a magnetic recording medium, an optical recording medium, etc., and provided, distributed, etc. [Explanation of symbols]

[0038] 1...Own vehicle, 11...Camera, 12...HMI, 13...Wireless communication device, 14...Vehicle control device, 14A...Steering actuator, 14B...Braking actuator, 14C...Driving actuator, 15...Image recognition device, 151...Communication interface, 152...Memory, 153...Processor, 3A...Acquisition unit, 3B...Determination unit, 3C...Recognition unit, 3D...Threshold value change unit, 3E...Base station number calculation unit

Claims

1. an acquisition unit that acquires a camera image showing the outside of the host vehicle captured by a camera; a determination unit that determines the type of a target included in the camera image and calculates a reliability score of the type of the target; a recognition unit that recognizes that a target of the type determined by the determination unit is included in the camera image when the reliability score calculated by the determination unit is equal to or greater than a score threshold; a threshold change unit that changes the score threshold; a base station number calculation unit that calculates the number of wireless communication base stations that transmit signals that can be received with a signal strength equal to or greater than a signal strength threshold in the wireless communication device mounted on the vehicle; The types of targets determined by the determination unit include at least pedestrians and animals other than humans, The threshold value changing unit When the number of wireless communication base stations calculated by the base station number calculation unit is equal to or less than a base station number threshold, reducing the score threshold to be compared with the reliability score of the non-human animal compared to when the number of wireless communication base stations calculated by the base station number calculation unit is greater than the base station number threshold; and An image recognition device that performs at least one of reducing the score threshold to be compared with the pedestrian's reliability score when the number of wireless communication base stations calculated by the base station number calculation unit is greater than the base station number threshold, more than when the number of wireless communication base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold.

2. The types of targets determined by the determination unit include a cliff, The threshold value changing unit 2. The image recognition device according to claim 1, wherein when the number of wireless communication base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold, the score threshold to be compared with the cliff reliability score is reduced more than when the number of wireless communication base stations calculated by the base station number calculation unit is greater than the base station number threshold.

3. The wireless communication base station includes a Wi-Fi base station and a Bluetooth (registered trademark) base station; The threshold value changing unit When the number of Bluetooth base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold, the score threshold that is compared with the reliability score of the animal other than humans is reduced more than when the number of Bluetooth base stations calculated by the base station number calculation unit is greater than the base station number threshold; and 2. The image recognition device of claim 1, wherein when the number of Bluetooth base stations calculated by the base station number calculation unit is greater than the base station number threshold, the image recognition device performs at least one of reducing the score threshold to be compared with the pedestrian's reliability score more than when the number of Bluetooth base stations calculated by the base station number calculation unit is equal to or less than the base station number threshold.

4. an acquisition step in which the image recognition device acquires a camera image showing the outside of the host vehicle captured by a camera; a determination step in which the image recognition device determines a type of a target included in the camera image and calculates a reliability score of the type of the target; a recognition step in which the image recognition device recognizes that a target of the type determined in the determination step is included in the camera image when the reliability score calculated in the determination step is equal to or greater than a score threshold; a threshold changing step of changing the score threshold by the image recognition device; a base station number calculation step in which the image recognition device calculates the number of wireless communication base stations that transmit signals that can be received by the wireless communication device mounted on the vehicle with a signal strength equal to or greater than a signal strength threshold, The types of the target determined in the determining step include at least a pedestrian and an animal other than a human, In the threshold value changing step, When the number of wireless communication base stations calculated in the base station number calculation step is equal to or less than a base station number threshold, reducing the score threshold to be compared with the reliability score of the non-human animal compared to when the number of wireless communication base stations calculated in the base station number calculation step is greater than the base station number threshold; and An image recognition method in which, when the number of wireless communication base stations calculated in the base station number calculation step is greater than the base station number threshold, at least one of reducing the score threshold to be compared with the pedestrian's reliability score is performed more than when the number of wireless communication base stations calculated in the base station number calculation step is equal to or less than the base station number threshold.

5. The processor an acquisition step of acquiring a camera image showing the outside of the host vehicle captured by a camera; a determination step of determining a type of a target included in the camera image and calculating a reliability score of the type of the target; a recognition step of recognizing that a target of the type determined in the determination step is included in the camera image when the reliability score calculated in the determination step is equal to or greater than a score threshold; a threshold changing step of changing the score threshold; a base station number calculation step of calculating the number of wireless communication base stations that transmit signals that can be received with a signal strength equal to or greater than a signal strength threshold in the wireless communication device mounted on the vehicle, The types of the target determined in the determining step include at least a pedestrian and an animal other than a human, In the threshold value changing step, When the number of wireless communication base stations calculated in the base station number calculation step is equal to or less than a base station number threshold, reducing the score threshold to be compared with the reliability score of the non-human animal compared to when the number of wireless communication base stations calculated in the base station number calculation step is greater than the base station number threshold; and A program that executes at least one of reducing the score threshold to be compared with the pedestrian's reliability score when the number of wireless communication base stations calculated in the base station number calculation step is greater than the base station number threshold, more than when the number of wireless communication base stations calculated in the base station number calculation step is equal to or less than the base station number threshold.

Images