Unmanned flight body and defense system

Abstract

To provide an unmanned flight body and a defense system for a purpose of defending against an intruder who entered a monitored area.SOLUTION: Image data, including an intruder X who has entered a monitored area 1, is acquired by image capture means. Based on the image data, characteristic information regarding the intruder X's appearance is acquired. Based on the characteristic information, text data indicating an alarm, including the item regarding the appearance, is created. The text data is converted into audio data. An unmanned flight body 3 reproduces the audio data through a speaker. In other words, the unmanned flight body 3 outputs an alarm indicating the audio data by the audio. The characteristic information is, for example, information regarding the clothes the intruder X wears.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present application relates to unmanned aerial vehicles and defense systems using unmanned aerial vehicles. [Background technology]

[0002] Conventionally, unmanned aerial vehicles known as drones have been used in work areas such as factories or warehouses. For example, Patent Document 1 describes a guidance system that includes a manned guided vehicle operated by an operator, one unmanned aerial vehicle capable of hovering in the air, and a management device that controls the unmanned aerial vehicle.

[0003] In the guidance system described in Patent Document 1, the unmanned aerial vehicle is equipped with a projector that projects a guidance image onto a road surface. The guidance image, for example, displays an arrow pointing in a specific direction and is projected onto the road surface in front of the manned guided vehicle. As a result, an operator operating the manned guided vehicle can be guided to the loading position by checking the guidance image.

[0004] In recent years, automated guided vehicles (AGVs) have been used instead of manned vehicles to reduce work loads. However, when an automated system is used in a work area such as a factory or warehouse, there is a risk that unauthorized intruders may steal or destroy luggage or equipment stored on shelves in the work area. In this regard, if there were a means to effectively warn intruders, it would be possible to expel intruders or deter them from entering the work area. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2020-52629 Summary of the Invention [Problem to be solved by the invention]

[0006] The present application provides an unmanned aerial vehicle and defense system for the purpose of defending against intruders into a monitored area. [Means for solving the problem]

[0007] The unmanned aerial vehicle of the present application is An unmanned aerial vehicle configured to fly in a predetermined surveillance area, A photographing device, a determination unit that detects an intruder who has entered the monitored area based on image data acquired by the image capturing device; a feature identification unit that acquires feature information relating to the appearance of the intruder based on image data including the intruder acquired by the photographing device; a warning generation unit that generates text data indicating a warning including matters related to the appearance based on the characteristic information; a conversion unit that converts the text data into audio data; and a speaker for outputting the warning indicated by the audio data.

[0008] The defense system of this application is at least one unmanned air vehicle configured to fly in a monitored area; a management device that wirelessly communicates with the unmanned aerial vehicle and manages the unmanned aerial vehicle; an imaging means for imaging the monitoring area; a voice conversion unit, the management device includes a determination unit, a feature identification unit, a warning generation unit, and a control instruction unit; the at least one unmanned air vehicle includes an unmanned air vehicle equipped with a speaker; the determination unit detects the intruder based on image data acquired by the imaging means, the characteristic identification unit acquires characteristic information relating to the appearance of the intruder based on image data including the intruder acquired by the imaging means; the warning creation unit creates text data indicating a warning including matters related to the appearance based on the characteristic information; the speech conversion unit converts the text data into speech data; The control command unit sends an audio output command to the unmanned aerial vehicle equipped with the speaker, causing the speaker of the unmanned aerial vehicle to output the warning indicated by the audio data.

[0009] In the defense system, the voice conversion unit may be provided in the unmanned aerial vehicle equipped with the speaker, and the control command unit may transmit the text data to the unmanned aerial vehicle together with the voice output command.

[0010] Alternatively, the voice conversion unit may be provided in the management device, and the control command unit may transmit the voice data together with the voice output command to the unmanned aerial vehicle equipped with the speaker.

[0011] In the defense system, the at least one unmanned aerial vehicle may include an unmanned aerial vehicle equipped with the speaker and an unmanned aerial vehicle equipped with a photographing device as the photographing means.

[0012] In the defense system, the at least one unmanned aerial vehicle may include an unmanned aerial vehicle equipped with a camera device as the imaging means and the speaker.

[0013] The determination unit may detect the intruder by determining whether the intruder is included in the image data acquired by the photographing device and wirelessly transmitted from the unmanned aerial vehicle to the management device.

[0014] Alternatively, the imaging means may be a plurality of monitoring cameras installed in the monitoring area.

[0015] A model storage medium is provided in which a trained model generated by performing machine learning using training data in which image data including a person is used as input data and feature information related to the appearance of the person is used as output data; and The feature identification unit The system may further include an acquisition unit that inputs acquired image data including the intruder into the trained model to acquire the characteristic information of the intruder.

[0016] The feature specifying unit further a text conversion unit that converts the feature information acquired by the acquisition unit into text, The warning generation unit The text data indicating the warning may be created using the feature information that has been converted into text.

[0017] The characteristic information may be clothing information relating to the clothing of the intruder.

[0018] The clothing information includes at least one of the color of the clothing, the type of the clothing, and the pattern of the clothing, The warning may include information regarding at least one of the color of the clothing, the type of the clothing, and the pattern of the clothing. [Brief explanation of the drawings]

[0019] [Figure 1] FIG. 1 is a schematic diagram illustrating an exemplary defense system. [Figure 2] FIG. 2 is a diagram illustrating an exemplary unmanned air vehicle. [Figure 3] FIG. 3 is a block diagram of an example controller for an unmanned air vehicle. [Figure 4] FIG. 4 is a conceptual diagram of exemplary training data. [Figure 5] 5A and 5B show exemplary feature information (clothing information). [Figure 6] FIG. 6 is a flow diagram of the defense process performed by the unmanned aerial vehicle in one embodiment. [Figure 7] FIG. 7 is a schematic diagram illustrating an example defense process. [Figure 8] FIG. 8 is a block diagram of another exemplary control device. [Figure 9] FIG. 9 is a block diagram of an example management device. [Figure 10]FIG. 10 is a schematic diagram illustrating yet another exemplary defense system. [Figure 11] FIG. 11 is a schematic diagram illustrating yet another exemplary defense system. [Figure 12] FIG. 12 is a block diagram of yet another exemplary management device. DETAILED DESCRIPTION OF THE INVENTION

[0020] An unmanned aerial vehicle and defense system according to an embodiment of the present invention will now be described with reference to the accompanying drawings, which are merely exemplary of the present invention. The drawings are merely schematic and may not be drawn to scale. It should also be understood that the same reference numerals are used throughout the figures to represent the same or similar components.

[0021] [First embodiment] 1 is a schematic diagram of an example defense system. The defense system photographs an intruder X who has invaded a predetermined area to be monitored (hereinafter referred to as monitored area 1), issues a warning, and defends monitored area 1 (e.g., objects and facilities within monitored area 1) from the intruder X.

[0022] The defense system in Figure 1 comprises a monitored area 1, at least one work vehicle 2, at least one unmanned aerial vehicle 3, and a management device 4 that remotely manages and controls the work vehicle 2 and the unmanned aerial vehicle 3. The management device 4 is capable of wireless communication with the unmanned aerial vehicle 3 and the work vehicle 2, and can send and receive information and control commands. The management device 4 is composed of a computer equipped with an arithmetic unit and a storage device.

[0023] The monitored area 1 is an area monitored by the unmanned aerial vehicle 3, and the unmanned aerial vehicle 3 is configured to fly within the monitored area 1. For example, the monitored area 1 is a work area within any facility, including a factory or warehouse. In the example shown in FIG. 1, the work area is provided with a plurality of shelves 10, and cargo is stored on the shelves 10. The shelves 10 may be fixed shelves, or may be movable shelves configured to be movable under the control of the management device 4, or may include both. The monitored area 1 may be outdoors.

[0024] The work vehicle 2 is a vehicle that performs work within a work area. The work vehicle 2 is, for example, a forklift that performs loading and unloading work. The work vehicle 2 may be a manned work vehicle or an unmanned work vehicle. In the example of FIG. 1, the work vehicle 2 is an unmanned forklift, and the work area (for example, the wall of a facility) is provided with multiple reflectors (not shown) for laser guidance of the work vehicle 2.

[0025] 1 projects a laser from a laser scanner into the surrounding area, detects the light reflected from multiple reflectors, and estimates its own position in real time based on the principle of triangulation. This allows the work vehicle 2 to travel along the movement route determined by the management device 4 based on its estimated self-position and perform loading and unloading on the shelves 10.

[0026] 1, one unmanned aerial vehicle 3 is used. Multiple unmanned aerial vehicles 3 may also be used. As shown in FIG. 2, the unmanned aerial vehicle 3 includes a main body 30, rotors 31, legs 32, a camera 33, a light 34, a speaker 35, a communication device 36, and a control device 37.

[0027] The main body 30 comprises a body incorporating a light 34, a speaker 35, a communication device 36, and a control device 37, and a plurality of arms (four in this example) extending radially from the body. A rotor 31 is provided at the top end of each arm, and a pair of legs 32 and a camera device 33 are attached to the bottom of the body.

[0028] The photographing device 33 is a camera that, under the control of the control device 37, photographs video and / or still images of the monitoring area 1 while the unmanned aerial vehicle 3 is flying, and wirelessly transmits the acquired image data to the management device 4 via the communication device 36. The photographing device 33 may be equipped with a light that emits light when photographing.

[0029] The light 34 outputs visible light and may be, for example, composed of at least one LED light. The light 34 may change the state of output of visible light. For example, the light may be a constantly lit light or a flashing light. The light 34 may also be equipped with an irradiation angle adjustment mechanism that can change the irradiation angle of the visible light.

[0030] The speaker 35 outputs sound under the control of the control device 37. The speaker 35 is configured to reproduce audio data. That is, the speaker 35 outputs the sound represented by the audio data. The speaker 35 is preferably a directional speaker.

[0031] The communication device 36 is for wireless communication with the management device 4, and thus allows transmission and reception of information to and from the management device 4.

[0032] The control device 37 controls the operation of each component of the aircraft based on the outputs of sensors (not shown) (such as an orientation sensor and an altitude sensor) and control commands from the management device 4. The control device 37 is composed of a calculation device and a storage device.

[0033] The configuration of the control device 37 of the first embodiment is shown in Fig. 3. The following functional units 370 to 374, 50 to 54 of the control device 37 are realized by one or more processors of the arithmetic device executing programs stored in a storage device (hard disk or memory).

[0034] The control device 37 includes a position information acquisition unit 370, a flight control unit 371, an imaging control unit 372, a light control unit 373, and a speaker control unit 374. The control device 37 includes a storage medium 55. The storage medium 55 is configured with a storage device. The storage medium 55 includes a map storage medium 550 that stores an environmental map of the monitored area 1, and a model storage medium 551 that stores a trained model (described later).

[0035] The position information acquisition unit 370 is configured to be able to acquire position information of the aircraft (main body 30), for example, using a satellite positioning system such as GPS. The position information includes, for example, latitude information, longitude information, and altitude information. The position information acquisition unit 370 outputs the acquired position information to the flight control unit 371 and transmits the position information to the management device 4. Note that the position information acquisition unit 370 is not limited to a satellite positioning system such as GPS, and any known configuration can be adopted, for example, the configuration described in Patent Document 1. However, when the configuration described in Patent Document 1 is adopted, a ceiling marker must be provided on the ceiling of the monitored area 1, and the position of the aircraft in the monitored area 1 may be acquired using the marker and an environmental map of the monitored area 1 stored in the map storage medium 550.

[0036] The flight control unit 371 performs flight control of the main body 30 based on the position information from the position information acquisition unit 370, control commands from the management device 4, and the like. Specifically, the flight control unit 371 controls the rotation of the rotors 31 as part of flight control of the main body 30. The flight control unit 371 can make the unmanned aerial vehicle 3 hover (stop in mid-air) by setting the rotation speeds of the electric motors of each rotor 31 to the same speed, raise the unmanned aerial vehicle 3 by simultaneously increasing the rotation speeds of each electric motor in the hovering state by the same amount, and change the direction of travel of the unmanned aerial vehicle 3 by changing the rotation speeds of the four electric motors at different rates. This allows the unmanned aerial vehicle 3 to fly within the monitored area 1 according to a flight path determined by the management device 4, and to track the intruder X after detecting the intruder X.

[0037] The photography control unit 372 causes the photography device 33 to take videos and / or still images during flight and transmits the image data obtained thereby to the management device 4 via the communication device 36. If the photography device 33 is equipped with lighting, the photography control unit 372 also controls the lighting. The photography control unit 372 temporarily stores the videos and / or still images in the storage medium 55 of the aircraft itself.

[0038] The control device 37 further includes a determination unit 50, a position identification unit 51, a feature identification unit 52, a warning generation unit 53, and a voice conversion unit .

[0039] When the unmanned aerial vehicle 3 is flying in the monitoring area 1 for monitoring purposes, the judgment unit 50 analyzes the image data acquired by the photographing device 33 by photographing the monitoring area 1 using a known method, and detects the intruder X by determining whether the image data includes a person (i.e., intruder X).

[0040] The position identification unit 51 is triggered by the detection of intruder X by the determination unit 50 to identify the position of intruder X. The position identification unit 51 acquires position information (position coordinates) of intruder X by analyzing image data using a known method. Alternatively, if the first unmanned aerial vehicle 3 is equipped with a distance sensor (e.g., LiDAR, millimeter-wave radar, or stereo camera) for measuring the distance to the target object (intruder X), the position identification unit 51 may identify the position of intruder X based on the position information of the unmanned aerial vehicle 3 (acquired from the position information acquisition unit 370), information related to the measurement results measured by the distance sensor, the direction in which the unmanned aerial vehicle 3 (image capture device 33) is facing, and an environmental map.

[0041] The feature identification unit 52 acquires feature information about the appearance of the intruder X based on image data including the intruder X acquired by the photographing device 33 (hereinafter referred to as image data of the intruder X), and further converts this feature information into text for creating a warning message, which will be described later. The feature information about the appearance here is clothing information about the clothing worn by the intruder X. The clothing includes, for example, a top, bottoms, a hat, a mask, and shoes. Hereinafter, the feature information will be described as clothing information.

[0042] More specifically, the clothing information of a person (intruder X) may include clothing color, clothing type, and / or clothing pattern. Clothing color may be, for example, red, blue, etc., and may be the color that makes up the majority of the clothing, since some clothing has designs. Clothing types include: tops (jerseys, T-shirts, jackets, etc.); bottoms (trousers (long pants), shorts, skirts, etc.); hats (knit hats, caps, not worn, etc.); masks (full-face masks, masks that cover only the eyes and nose), not worn, etc.; shoes (sneakers, sandals, not worn (barefoot), etc.), and the clothing type may also include not wearing a hat, mask, or shoes. Clothing patterns include solid colors, checks, stripes, stripes, etc. For example, it is not necessary to specify one or more of tops, bottoms, hats, masks, and shoes. For example, the types of hats, masks, and shoes may be classified simply by whether they are worn or not.

[0043] The feature identification unit 52 includes an acquisition unit 520 that acquires clothing information (an example of feature information) using a trained model, and a text conversion unit 521 that converts the clothing information acquired by the acquisition unit 520 into text.

[0044] The acquisition unit 520 is configured to input image data of the intruder X to a trained model and acquire clothing information using the output from the trained model. The trained model is stored in a model storage medium 551. The trained model is generated by machine learning a neural network using training data in which an image of a person is input data and clothing information about the clothes worn by the person is output data. In other words, the trained model is a clothing recognition model that, when it receives input image data including a person, outputs clothing information about the person.

[0045] An example of training data is conceptually shown in FIG. 4. The training data consists of image data (input data) including the entire image of a person wearing clothes, and clothing information (output data: correct answer data) of the person, including the color and type of hat, mask, top, bottom, and shoes. Note that if the mask is a "full-face" mask, since it covers not only the face but also the head, the type of hat may also be "full-face." For simplicity, FIG. 4 does not include information about the pattern of the clothing, but such a configuration is of course also possible. This exemplary trained model is generated by machine learning using such multiple training data. Thus, as described below, the acquisition unit 520 inputs image data of intruder X into the trained model to acquire information indicating the color and type of intruder X's hat, mask, top, bottom, and shoes.

[0046] The text conversion unit 521 converts the clothing information acquired by the acquisition unit 520 into text for use in creating a warning message, which will be described later. As shown in Figures 5A and 5B, the text conversion unit 521 in this example converts the clothing information into text for each piece of clothing (hat, mask, top, bottom, shoes). Note that if the person is not wearing a hat, mask, or shoes, that part does not need to be converted into text (see (null) in Figure 5A).

[0047] In this way, the feature identification unit 52 extracts clothing information of the person (intruder X) included in the image data and converts it into text. The obtained textual clothing information (clothing data) is stored in the storage medium 55.

[0048] The warning creation unit 53 creates text data indicating a warning including information about the clothing worn by the intruder X based on the clothing information (an example of characteristic information). The warning may be a warning calling for the intruder to leave the monitored area 1, surrendering, or threatening the intruder X. The warning creation unit 53 in this example creates the warning in the form of text data using part or all of the textual clothing information and a text template for the warning message.

[0049] An example of a text template is "An intruder wearing <clothing information> has been detected. Recording has begun. Police have been notified. Please do not move." Another example text template is "Intruder wearing <clothing information>, please leave here immediately."

[0050] The defense system may be configured to store multiple such text templates in storage medium 55, and allow a user of the defense system to communicate with unmanned aerial vehicle 3 via a user interface of management device 4 and select one template to be used by warning creation unit 53. The defense system may also be configured to create a text template to be used by warning creation unit 53 through input from the user interface and send it to unmanned aerial vehicle 3.

[0051] The warning creation unit 53 completes the warning by applying part or all of the textualized clothing information (text) of the intruder X to the <clothing information> in the template, and creates text data indicating the warning.

[0052] For example, when the textual clothing information (clothing data) of FIG. 5A is acquired, the warning creation unit 53 creates a warning message stating, "An intruder wearing a green jacket has been detected. Recording has begun. The police have been notified. Please do not move."

[0053] Also, for example, when the textual clothing information (clothing data) of FIG. 5B is acquired, the warning creation unit 53 creates a warning message stating, "An intruder wearing a white T-shirt has been detected. Recording has begun. The police have been notified. Please do not move."

[0054] In this example, the warning generator 53 uses the text for the top garment, but the warning generator 53 may also use the text for the hat, mask, bottom garment, and / or shoes in addition to or instead of the top garment. The defense system may be configured to allow input from a user interface to preselect the text for the hat, mask, top garment, bottom garment, and / or shoes to be used by the warning generator 53.

[0055] Note that feature identification unit 52 may apply a known clothing recognition algorithm to extract a person from image data and identify clothing information about the person (e.g., the color of their jacket) without using artificial intelligence technology such as machine learning, and warning creation unit 53 may then create text data indicating a warning (e.g., text data indicating "Intruder wearing red clothing...") based on the extracted clothing information. The configuration of warning creation unit 53 is not limited to the above.

[0056] The voice conversion unit 54 converts the text data indicating the warning into voice data. As a specific configuration of the voice conversion unit 54, a known text-to-speech synthesis technology such as text to speech (TTS) technology can be adopted.

[0057] Although details of the management device 4 are not described in the first embodiment, the management device 4 is configured to communicate wirelessly with the unmanned aerial vehicle 3. The management device 4 transmits a monitoring command to the unmanned aerial vehicle 3 along with a flight path for monitoring. Upon receiving the monitoring command, the unmanned aerial vehicle 3 (control device 37) executes the following defense processing.

[0058] 6, in step S1, in response to receiving a monitoring command, the unmanned aerial vehicle 3 flies along a monitoring flight path and monitors the monitoring area 1 by photographing the monitoring area 1 with the photographing device 33 (see FIG. 1). That is, the determination unit 50 determines whether or not the image data acquired by the photographing device 33 includes an intruder X.

[0059] If the determination unit 50 determines that the image data does not include an intruder X and does not detect the intruder X (NO in S2), the control device 37 continues the surveillance flight. If the determination unit 50 determines that the image data includes an intruder X and detects the intruder X (YES in S2), the control device 37 proceeds to the next step S3.

[0060] In step S3, the control device 37 refers to the position of the intruder X captured in the image capturing device 33, moves the unmanned aerial vehicle 3 closer to the intruder X, and then captures an image using the image capturing device 33 to obtain image data of the intruder X for obtaining clothing information (an example of characteristic information). At this stage, the control device 37 may also capture a video using the image capturing device 33 and record it in the storage medium 55.

[0061] Next, the control device 37 proceeds to step S4. In step S4, the feature identification unit 52 acquires and converts clothing information into text based on the image data of the intruder X. That is, the acquisition unit 520 inputs the image data into a trained model to acquire the clothing information of the intruder X, and the text conversion unit 521 converts the clothing information into text.

[0062] Next, the control device 37 proceeds to step S5. In step S5, the warning creation unit 53 uses the text-converted clothing information acquired in S4 to create text data indicating a warning including information about the clothing of the intruder X, as described above.

[0063] Next, the control device 37 proceeds to step S6. In step S6, an audio warning is issued to the intruder X. To this end, the audio conversion unit 54 converts the text data into audio data. The control device 37 has already brought the unmanned aerial vehicle 3 close to the intruder X, and while referring to the position of the intruder X using the imaging device 33, adjusts the position and orientation of the unmanned aerial vehicle 3 and points the directional speaker 35 toward the intruder X. The control device 37 then causes the speaker 35 to play audio data, thereby outputting the warning indicated by the audio data to the intruder X (see FIG. 7).

[0064] In this example, in the case of the clothing information of Fig. 5A, the unmanned aerial vehicle 3 outputs a voice message from the speaker 35 saying, "An intruder wearing a green jacket has been detected. Recording has begun. The police have been notified. Please do not move." Also, in the case of the clothing information of Fig. 5B, the unmanned aerial vehicle 3 outputs a voice message from the speaker 35 saying, "An intruder wearing a white T-shirt has been detected. Recording has begun. The police have been notified. Please do not move."

[0065] In this way, in the defense system, the unmanned aerial vehicle 3 reads out a warning that includes specific external characteristics of the intruder X, causing the intruder X to feel a strong sense of surprise, awe, and being watched, effectively driving the intruder X away from the monitored area 1. In other words, a high level of defensive effectiveness is expected. In Figure 1, it is expected that luggage, equipment, facilities, etc. within the work area for the unmanned work vehicle 2 can be more effectively defended.

[0066] The unmanned aerial vehicle 3 may return to a predetermined waiting position after issuing the warning, or may continue (or repeat) issuing the warning until the intruder X leaves the monitored area 1. When the intruder X is detected, the management device 4 may actually notify the police or a security company.

[0067] [Second embodiment] The second embodiment will be described below. The monitoring area 1 has the same configuration as that shown in Figure 1. The unmanned aerial vehicle 3 of this embodiment has the configuration shown in Figure 2 and, as shown in Figure 8, is equipped with control units 370-374 and a voice conversion unit 54 similar to those of the first embodiment. The imaging device 33 of the unmanned aerial vehicle 3 functions as the "imaging means" in the "defense system" of the present invention.

[0068] 9, the management device 4 of this embodiment includes a communication device 40, a user interface 41, a display 42, and a control command unit 43, and further includes a storage medium 55, a determination unit 50, a position identification unit 51, a feature identification unit 52, and a warning creation unit 53. Each of the units 43, 50 to 53 is realized by one or more processors of a computing device constituting the management device 4 executing a program stored in a storage device (hard disk or memory).

[0069] As such, the second embodiment differs from the first embodiment in that the functional units 50 to 53 are provided in the management device 4. The unmanned aerial vehicle 3 (control device 37) wirelessly transmits each piece of information (image data, location information, etc.) that it has acquired to the management device 4 in real time.

[0070] The communication device 40 is used for wireless communication with the unmanned aerial vehicle 3. The user interface 41 is used to select the start of the defense processing function, select the text template to be used as described above, and make various settings. The display 42 displays image data in real time showing the state of the monitored area 1 that is acquired by the imaging device 33 of the unmanned aerial vehicle 3 and transmitted wirelessly, and displays a message to that effect when an intruder X is detected.

[0071] The judgment unit 50 of this embodiment detects intruder X by analyzing image data wirelessly transmitted from the unmanned aerial vehicle 3 using a known method and determining whether the image data contains a person (intruder X).

[0072] The position determination unit 51 determines the position of the intruder X based on the environmental map of the monitored area 1 (stored in the map storage medium 550), the position information of the unmanned aerial vehicle 3 (wirelessly transmitted from the unmanned aerial vehicle 3), the distance between the intruder X and the unmanned aerial vehicle 3 determined by image processing, and the like.

[0073] The feature identification unit 52 and the warning generation unit 53 are the same as those in the first embodiment.

[0074] The control command unit 43 transmits control commands (monitoring commands, flight commands, photography commands, audio output commands, etc.) to the unmanned aerial vehicle 3 to control the operation of the unmanned aerial vehicle 3. The control command unit 43 also determines a flight path to the intruder X based on the position of the intruder X, the position of the unmanned aerial vehicle 3, an environmental map, etc.

[0075] In this example, when the control command unit 43 of the management device 4 sends a monitoring command to the unmanned aerial vehicle 3, the unmanned aerial vehicle 3 begins a monitoring flight, as in the first embodiment. The unmanned aerial vehicle 3 transmits image data captured by the imaging device 33 in real time, and the judgment unit 50 of the management device 4 detects the intruder X.

[0076] When an intruder X is detected, the control command unit 43 sends a photographing command, and when the unmanned aerial vehicle 3 receives this, it approaches the intruder X, takes a photograph to obtain characteristic information, and sends image data of the intruder X to the management device 4, as in the first embodiment.

[0077] Then, in the management device 4, the feature specifying unit 52 and the warning creating unit 53 create text data indicating a warning in the same manner as in the first embodiment.

[0078] Then, the control command unit 43 transmits the voice output command to the unmanned aerial vehicle 3 together with the text data.

[0079] When the unmanned aerial vehicle 3 receives the voice output command together with the text data, it outputs a warning to the intruder X by voice, indicated by the voice data converted by the voice conversion unit 54, as in the first embodiment (see FIG. 6).

[0080] [Third embodiment] The third embodiment is a modification of the second embodiment. For defense processing, as shown in Figure 10, two unmanned aerial vehicles 3a and 3b are prepared, with one unmanned aerial vehicle 3a being used for surveillance flight and photography, and the other unmanned aerial vehicle 3b being used for audio output. The unmanned aerial vehicles 3a and 3b have the configurations shown in Figures 2 and 8, but the unmanned aerial vehicle 3a does not necessarily have to have the speaker 35, speaker control unit 374, or audio conversion unit 54.

[0081] When unmanned aerial vehicle 3a detects intruder X during surveillance flight, control command unit 43 determines a flight route to intruder X for waiting unmanned aerial vehicle 3b and transmits a flight command along with the flight route. Upon receiving the flight command, unmanned aerial vehicle 3b flies along the flight route to the location of intruder X (a position where intruder X can be captured by its own camera device 33). Furthermore, when text data is created based on image data of intruder X acquired by the camera device 33 of unmanned aerial vehicle 3a, control command unit 43 transmits an audio output command together with the text data to unmanned aerial vehicle 3b. Once unmanned aerial vehicle 3b flies to the location of intruder X and receives the audio output command and text data, it outputs an audio warning in the same manner as above (see FIG. 10).

[0082] [Fourth embodiment] The fourth embodiment will be described below. As shown in Fig. 11, in the fourth embodiment, multiple surveillance cameras 11 are installed in a monitoring area 1 (for example, a wall, ceiling, or shelf 10 of a work area) and connected to a management device 4. These multiple surveillance cameras 11 serve as the "imaging means" of the "defense system" of the present invention. The unmanned aerial vehicle 3 has the configuration shown in Figs. 2 and 8.

[0083] In this embodiment, the determination unit 50 detects the intruder X by analyzing each image data acquired by each surveillance camera 11 using a known method and determining whether or not the image data includes a person (intruder X). The position identification unit 51 identifies the position of the intruder X based on an environmental map of the surveillance area, the position information of each surveillance camera 11, the distance between the intruder X and the surveillance camera 11 determined by image processing, and the like. The feature identification unit 52 uses the image data of the intruder X acquired by the surveillance camera 11.

[0084] 12, the management device 4 further includes a surveillance camera control unit 44 that controls each surveillance camera 11. Each surveillance camera 11 may be fixed, or may be moved under the control of the surveillance camera control unit 44 to scan a certain area. Each surveillance camera 11 may also have a function that enables it to take a magnified image of an object under the control of the surveillance camera control unit 44.

[0085] In the fourth embodiment, the defensive process is performed as follows.

[0086] For example, when a user of the defense system inputs a command to start defense processing via the user interface 41, the surveillance camera control unit 44 activates each surveillance camera 11 to monitor the surveillance area 1. The determination unit 50 detects an intruder X based on image data from each surveillance camera 11.

[0087] Then, triggered by the detection of intruder X by the determination unit 50, the surveillance camera control unit 44 controls the surveillance camera 11 that discovered the intruder X, enlarges the intruder X as necessary, photographs the intruder X to obtain feature information, and outputs the image data of the intruder X to the feature identification unit 52.

[0088] At the same time, the position identification unit 51 identifies the position of the intruder X, and the control command unit 43 determines a flight path to the intruder X and transmits a flight command to the waiting unmanned aerial vehicle 3 along with the determined flight path.

[0089] When the unmanned aerial vehicle 3 receives the flight command together with the flight path, it flies from the standby position to the intruder X (see FIG. 11).

[0090] The subsequent steps are the same as those in the second embodiment and will therefore be omitted.

[0091] As shown in Figure 11, it is possible that multiple intruders Xa and Xb may simultaneously invade the monitored area 1. In this case, the management device 4 may assign unmanned aerial vehicle 3c to intruder Xa and unmanned aerial vehicle 3d to intruder Xb, and perform defense processing individually.

[0092] As described above, the second to fourth embodiments can also be expected to have the same defensive effect as the first embodiment. In the second to fourth embodiments, the voice conversion unit 54 may be provided in the management device 4. In this case, the control command unit 43 transmits a voice output command from the management device 4 to the unmanned aerial vehicle 3 together with the converted voice data.

[0093] In the first to fourth embodiments, the characteristic information is clothing information, but instead of or in addition to this, it may also include the intruder's physique and / or height. [Explanation of symbols]

[0094] 1 Monitoring area 11. Surveillance cameras (examples of photography methods) 3(3a, 3b, 3c, 3d) Unmanned aerial vehicle 33 Photography device (an example of photography means) 35 speakers 4 Management device 43 Control Command Section 50 Judgment section 52 Feature Identification Unit 520 Acquisition Department 521 Text Conversion Department 53 Warning Creation Department 54 Voice conversion unit X(Xa, Xb) Intruder

Claims

1. An unmanned aerial vehicle configured to fly in a predetermined surveillance area, A photographing device, a determination unit that detects an intruder who has entered the monitored area based on image data acquired by the image capturing device; a feature specifying unit that obtains feature information relating to the appearance of the intruder based on image data including the intruder obtained by the photographing device; a warning generation unit that generates text data indicating a warning including matters related to the appearance based on the characteristic information; a conversion unit that converts the text data into audio data; a speaker for outputting the warning indicated by the audio data as audio. Unmanned aerial vehicle.

2. The unmanned aerial vehicle further comprises: a model storage medium storing a trained model generated by performing machine learning using training data in which image data including a person is used as input data and feature information related to the appearance of the person is used as output data; The feature identification unit an acquisition unit that inputs the acquired image data including the intruder into the trained model to acquire the feature information of the intruder; The unmanned aerial vehicle according to claim 1 .

3. The feature specifying unit further a text conversion unit that converts the feature information acquired by the acquisition unit into text, The warning generation unit creating the text data indicating the warning using the feature information converted into text; The unmanned aerial vehicle according to claim 1 .

4. The characteristic information is clothing information regarding the clothing of the intruder. The unmanned aerial vehicle according to claim 1 .

5. The clothing information includes at least one of the color of the clothing, the type of the clothing, and the pattern of the clothing, The warning includes information regarding at least one of the color of the clothing, the type of the clothing, and the pattern of the clothing. The unmanned aerial vehicle according to claim 4.

6. A defense system, at least one unmanned air vehicle configured to fly in a monitored area; a management device that wirelessly communicates with the unmanned aerial vehicle and manages the unmanned aerial vehicle; an imaging means for imaging the monitoring area; a voice conversion unit, the management device includes a determination unit, a feature identification unit, a warning generation unit, and a control instruction unit; the at least one unmanned air vehicle includes an unmanned air vehicle equipped with a speaker; the determination unit detects the intruder based on image data acquired by the imaging means, the characteristic identification unit acquires characteristic information relating to the appearance of the intruder based on image data including the intruder acquired by the imaging means; the warning creation unit creates text data indicating a warning including matters related to the appearance based on the characteristic information; the speech conversion unit converts the text data into speech data; The control command unit transmits an audio output command to the unmanned aerial vehicle equipped with the speaker, causing the speaker of the unmanned aerial vehicle to output the warning indicated by the audio data. Defense system.

7. The defense system further comprises: a model storage medium storing a trained model generated by performing machine learning using training data in which image data including a person is used as input data and feature information related to the appearance of the person is used as output data; The feature identification unit an acquisition unit that inputs the acquired image data including the intruder into the trained model to acquire the feature information of the intruder; The defense system of claim 6.

8. The feature specifying unit further a text conversion unit that converts the feature information acquired by the acquisition unit into text, The warning generation unit creating the text data indicating the warning using the feature information converted into text; The defense system of claim 7.

9. The characteristic information is clothing information regarding the clothing of the intruder. The defense system of claim 6.

10. The clothing information includes at least one of the color of the clothing, the type of the clothing, and the pattern of the clothing, The warning includes information regarding at least one of the color of the clothing, the type of the clothing, and the pattern of the clothing. The defense system of claim 9.

11. The voice conversion unit is provided in the unmanned aerial vehicle equipped with the speaker, and the control command unit transmits the text data together with the voice output command to the unmanned aerial vehicle. The defense system of claim 6.

12. The voice conversion unit is provided in the management device, and the control command unit transmits the voice data together with the voice output command to the unmanned aerial vehicle equipped with the speaker. The defense system of claim 6.

13. the at least one unmanned air vehicle, the unmanned aerial vehicle equipped with the speaker; and an unmanned aerial vehicle equipped with a photographing device as the photographing means. The defense system of claim 6.

14. The at least one unmanned aerial vehicle includes an unmanned aerial vehicle equipped with a camera as the imaging means and the speaker. The defense system of claim 6.

15. The determination unit detects the intruder by determining whether the intruder is included in the image data acquired by the photographing device and wirelessly transmitted from the unmanned aerial vehicle to the management device. Defense system according to claim 13 or claim 14.

16. The imaging means is a plurality of monitoring cameras installed in the monitoring area. The defense system of claim 6.

Images