Crime prevention system

Abstract

To provide a crime prevention system capable of preventing a case to a vehicle, a case generated around the vehicle, and the like.SOLUTION: A crime prevention system comprises: a detection part for detecting a suspicious person that is present around a vehicle by image analysis processing for a video image photographed with a camera installed in the vehicle; a sentence generation part for generating a hail sentence to hail to the suspicious person by verbalization processing for the video image including the suspicious person; a voice generation part for generating voice data for converting the hail sentence to voice; and an output part for outputting hail voice obtained by converting the hail sentence to voice based on the voice data.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] FIELD OF THE DISCLOSURE The present invention relates to a security system. [Background technology]

[0002] In a system for preventing incidents such as vehicle theft and vandalism, images captured by an on-board camera such as a drive recorder may be used. For example, when an incident occurs with a vehicle, there is a technology in which information including the date, time, location, etc. of the incident is transmitted from an agency such as the police to the vehicle, and images corresponding to the date, time, location, etc. among the images captured by the on-board camera mounted on the vehicle are stored for a certain period of time (Patent Document 1, etc.). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2011-180693 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, while the conventional technology can improve responses after an incident has occurred, it is difficult to prevent an incident from occurring.

[0005] One of the problems that the embodiments of the present invention aim to solve is to provide a crime prevention system that can prevent crimes against vehicles and crimes that occur around vehicles. [Means for solving the problem]

[0006] One embodiment of the security system according to the present invention comprises a detection unit that detects suspicious persons in the vicinity of the vehicle by image analysis processing of video captured by a camera mounted on the vehicle, a sentence generation unit that generates a message to call out to the suspicious person by verbalizing the video including the suspicious person, a voice generation unit that generates voice data to vocalize the message, and an output unit that outputs a voice that vocalizes the message based on the voice data.

[0007] According to the above configuration, when a suspicious person is detected in the image of the camera mounted on the vehicle, a message to be spoken to the suspicious person is generated based on the image, and the message is converted into a voice and output. This makes it possible to dissuade the suspicious person from committing a crime or a nuisance, and to prevent incidents against the vehicle or incidents occurring around the vehicle. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of a hardware configuration of a security system according to a first embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of a functional configuration of the security system according to the first embodiment. [Diagram 3] FIG. 3 is a flowchart showing an example of processing in the security system of the first embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a hardware configuration of the security system according to the second embodiment. [Diagram 5] FIG. 5 is a diagram illustrating an example of a functional configuration of the security system according to the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Exemplary embodiments of the present invention are disclosed below. The configurations of the embodiments shown below, and the actions, results, and effects brought about by the configurations are merely examples. The present invention can be realized with configurations other than those disclosed in the following embodiments, and it is possible to obtain at least one of various effects and derivative effects based on the basic configurations.

[0010] (First embodiment) 1 is a diagram showing an example of a hardware configuration of a security system 1 according to a first embodiment. The security system 1 according to the present embodiment includes a vehicle 10 and a server 20.

[0011] The vehicle 10 is an automobile or the like capable of traveling on a roadway, and is equipped with a camera 11, an edge AI (Artificial Intelligence) device 12, an ECU (Electronic Control Unit) 13, a communication device 14, a speaker 15, a horn 16, lights 17, etc.

[0012] The camera 11 is a device capable of capturing images (including still images and moving images) of the surroundings of the vehicle 10, and may be, for example, a digital camera constituting part of a drive recorder.

[0013] The edge AI device 12 is an information processing device that enables processing to be performed locally (inside the vehicle 10) using a trained model generated by machine learning (deep learning). The hardware configuration of the edge AI device 12 is not particularly limited and may be configured by appropriately using publicly known or new technologies, and may be configured by using, for example, a central processing unit (CPU), a graphics processing unit (GPU), a neural network processing unit (NPU), a memory, etc. The edge AI device 12 of this embodiment executes image analysis processing and the like using a trained model, etc., using data based on an image captured by the camera 11 (for example, feature quantities indicating the characteristics of the image, etc.) as input.

[0014] The ECU 13 is an information processing device that executes arithmetic processing, control processing, etc. to realize the functions of the security system 1. The hardware configuration of the ECU 13 is not particularly limited and may be configured by appropriately utilizing publicly known or new technologies, for example, it may be configured by using a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), a memory, etc. The ECU 13 of this embodiment controls the communication device 14, the speaker 15, the horn 16, the light 17, etc. based on the processing result by the edge AI device 12, data acquired from the server 20, etc.

[0015] The communication device 14 is a device that enables communication between an in-vehicle device such as the ECU 13 mounted on the vehicle 10 and an external device such as the server 20 via a communication network N including the Internet. The communication device 14 of this embodiment transmits data required to realize the functions of the security system 1 from the vehicle 10 to the server 20, and receives data transmitted from the server 20.

[0016] The speaker 15 is a device capable of outputting sound to the outside of the vehicle 10. The speaker 15 may be an internal speaker of an audio or navigation system that outputs sound to passengers inside the vehicle, or may be a dedicated speaker suitable for outputting sound to the outside of the vehicle 10.

[0017] The horn 16 is a device that outputs a warning sound to the outside of the vehicle 10.

[0018] The light 17 is a device that emits light toward the outside of the vehicle 10, and may be, for example, a headlight, a parking light, a hazard lamp, a backlight, or the like.

[0019] The server 20 is an information processing device that executes predetermined processing to realize the functions of the security system 1. The hardware configuration of the server 20 is not particularly limited and may be configured by appropriately utilizing publicly known or new technologies, for example, it may be configured by utilizing a CPU, a GPU, an NPU, an ASIC, an FPGA, a memory, a communication interface, a user interface, etc. The server 20 of this embodiment generates data required to realize the functions of the security system 1 based on data acquired from the vehicle 10 via the communication network N, and transmits the data to the vehicle 10 via the communication network N.

[0020] The crime prevention system 1 of this embodiment has a function for preventing incidents against the vehicle 10 and incidents occurring around the vehicle 10.

[0021] Fig. 2 is a diagram showing an example of a functional configuration of the security system 1 of the first embodiment. In Fig. 2, functions provided in the vehicle 10 and the server 20 are illustrated.

[0022] The vehicle 10 of this embodiment includes an acquisition unit 101, a detection unit 102, a transmission unit 103, a reception unit 104, an output unit 105, and a recording unit 106. These functional units are configured by the cooperation of hardware and software (programs, etc.) mounted on the vehicle 10. Furthermore, at least one of these functional units may be configured by dedicated hardware (circuits, etc.).

[0023] The acquisition unit 101 acquires an image captured around the vehicle 10. The image may be, for example, an image captured by a camera 11 constituting a part of a drive recorder mounted on the vehicle 10.

[0024] The detection unit 102 detects a suspicious person present in the vicinity of the vehicle 10 by image analysis processing of the video acquired by the acquisition unit 101, and outputs a suspicious person video that is a video including the suspicious person. The image analysis processing of this embodiment includes an algorithm configured using a trained model generated by machine learning, and is executed by the edge AI device 12 mounted on the vehicle 10.

[0025] The image analysis process of this embodiment includes a person detection process, a tracking process, and a posture estimation process. The person detection process is a process for determining whether or not a person is included in an image, for example, a process for detecting a person from each of a plurality of frames constituting an image as a moving image. The tracking process is a process for tracking a person detected by the person detection process in an image that changes in time series. The posture estimation process is a process for estimating the posture of a person detected by the person detection process. The posture referred to here is a posture related to a criminal act, a nuisance act, or the like, and may be, for example, a posture of touching or about to touch the vehicle 10, or a posture of a person about to harm another person, or the like.

[0026] The detection unit 102 of this embodiment determines whether or not there is a suspicious person around the vehicle 10, in other words, whether or not a person around the vehicle 10 is a suspicious person, based on the results of the person detection process, the tracking process, and the posture estimation process. Note that the detection unit 102 may detect a suspicious person by using information other than the results of the person detection process, the tracking process, and the posture estimation process.

[0027] The transmitting unit 103 transmits the suspicious person video output from the detecting unit 102 to the server 20.

[0028] The receiving unit 104 receives the voice data transmitted from the server 20. The voice data will be described later.

[0029] The output unit 105 outputs a voice message to call out to the suspicious person based on the voice data received by the receiving unit 104. The voice message is output using, for example, a speaker 15 capable of outputting voice to the outside of the vehicle 10.

[0030] The recording unit 106 records the suspicious person video when a predetermined condition is satisfied. The predetermined condition may be, for example, that the suspicious person who was the target of the voice call continues to exist even after the output unit 105 outputs the voice call a predetermined number of times. The recording may be, for example, a process of storing the data of the suspicious person video in a storage device mounted on the vehicle 10 for a fixed period or for an indefinite period. The storage device that records the suspicious person video is not limited to one mounted on the vehicle 10, but may be one mounted on an external device such as the server 20.

[0031] The server 20 of this embodiment includes a receiving unit 201, a text generating unit 202, a voice generating unit 203, and a transmitting unit 204. These functional units are configured by cooperation between hardware and software constituting the server 20. Furthermore, at least one of these functional units may be configured by dedicated hardware (circuitry, etc.).

[0032] The receiving unit 201 receives the suspicious person video transmitted by the transmitting unit 103 of the vehicle 10 .

[0033] The sentence generation unit 202 generates a calling sentence for calling out to a suspicious person included in the suspicious person video by performing a verbalization process on the suspicious person video received by the receiving unit 201. The verbalization process of this embodiment includes captioning and natural language processing.

[0034] Captioning is generally referred to as image caption generation or the like, and is a process of generating a caption including a phrase, sentence, or the like that describes the suspicious person video received by the receiving unit 201. The specific method of the captioning is not particularly limited and may be realized by appropriately utilizing publicly known or new technology, but captioning may be realized by, for example, utilizing an encoder-decoder model that utilizes a neural network.

[0035] Natural language processing generates a message to the suspicious person by using the caption of the suspicious person video generated by captioning. The specific method of the natural language processing is not particularly limited and may be configured by appropriately using publicly known or new technologies, but the natural language processing may be realized by using, for example, an encoder-decoder model using a neural network, or a sentence generation AI available via the Internet, etc.

[0036] The voice generation unit 203 generates voice data for vocalizing the call sentence generated by the sentence generation unit 202. The specific method of vocalization is not particularly limited and may be configured by appropriately utilizing publicly known or new technologies, and may be realized by, for example, waveform concatenative voice synthesis, parametric voice synthesis, a text-to-speech conversion model generated by machine learning, or the like.

[0037] The transmitting unit 204 transmits the voice data generated by the voice generating unit 203 to the vehicle 10 that transmitted the suspicious person video corresponding to the voice data.

[0038] According to the above-mentioned configuration, when a suspicious person is detected in the image of the camera 11 mounted on the vehicle 10, a message to be called out to the suspicious person is generated based on the suspicious person image including the suspicious person, and voice data for vocalizing the message is generated. Then, a voice to be called out to the suspicious person is output from the speaker 15 mounted on the vehicle 10 based on the voice data.

[0039] 3 is a flowchart showing an example of processing in the security system 1 of the first embodiment. In step S101, the acquisition unit 101 acquires an image of the periphery of the vehicle 10 (image captured by the camera 11). In step S102, the detection unit 102 determines whether or not there is a suspicious person in the periphery of the vehicle 10 by image analysis processing of the image. In step S102, if there is no suspicious person (S102: No), this routine ends, and if there is a suspicious person (S102: Yes), in step S103, the transmission unit 103 transmits the suspicious person image to the server 20.

[0040] In step S104, the sentence generation unit 202 of the server 20 performs captioning on the suspicious person video received from the vehicle 10, and generates a caption that describes the suspicious person video. After that, in step S105, the sentence generation unit 202 performs natural language processing using a sentence generation AI or the like using the generated caption, and generates a sentence to call out to the suspicious person.

[0041] In step S106, the voice generating unit 203 of the server 20 generates voice data for vocalizing the generated call out message. In step S107, the transmitting unit 204 of the server 20 transmits the generated voice data to the vehicle 10.

[0042] In step S108, the output unit 105 of the vehicle 10 outputs a voice message obtained by converting the message into a voice message based on the voice data received from the server 20 from the speaker 15 mounted on the vehicle 10. At this time, a warning sound may be output from the horn 16, and the light 17 may be turned on or flashed.

[0043] After that, in step S109, the detection unit 102 determines whether or not the suspicious person (the suspicious person detected in step S102 and who was the target of the call) is no longer present. In step S109, if the suspicious person is no longer present (S109: Yes), this routine ends.

[0044] In step S109, if the suspicious person is not absent (S109: Yes), that is, if the suspicious person continues to be present even after the voice calling out is output, in step S110, the output unit 105 determines whether or not the number of times the voice calling out has been output has reached a predetermined threshold value. In step S110, if the number of times the voice calling out has been output has not reached the threshold value (S110: No), the processes from step S108 onwards are executed again, and the voice calling out continues to be output.

[0045] In step S110, if the number of times the voice calling out reaches the threshold value (S110: Yes), that is, if the suspicious person continues to exist even after the voice calling out is output multiple times, in step S111, the recording unit 106 records the suspicious person video image showing the suspicious person in a storage device (for example, the memory of a drive recorder, etc.). At this time, a process of notifying a specific organization such as the police or a security company of the presence of the suspicious person may be performed.

[0046] As described above, according to this embodiment, when a suspicious person is detected in the image captured by the camera 11 mounted on the vehicle 10, a message to be spoken to the suspicious person is generated based on the image (image of the suspicious person), and the message is converted into a voice and output from the vehicle 10. This makes it possible to dissuade the suspicious person from committing a crime or a nuisance, and to prevent incidents against the vehicle 10 or incidents occurring around the vehicle 10.

[0047] In the above embodiment, the verbalization process includes captioning that generates a caption to describe a video including a suspicious person (suspicious person video), and natural language processing that generates a call-out sentence using the caption. This makes it possible to generate appropriate call-out sentences for various suspicious people.

[0048] In the above embodiment, the image analysis process includes a person detection process for determining whether or not a person is included in the video, a tracking process for tracking the person detected by the person detection process, and a posture estimation process for estimating the posture of the person detected by the person detection process, and the detection unit 102 determines whether or not a suspicious person is present based on the results of the person detection process, the tracking process, and the posture estimation process. This allows the presence or absence of a suspicious person to be determined with high accuracy.

[0049] In the above embodiment, if the suspicious person continues to be present after the voice calling out to the user is output a predetermined number of times, a video including the suspicious person (suspicious person video) is recorded. This makes it possible to record only the video of suspicious people who are likely to be committing crimes, etc., and to avoid unnecessary recording of video of people who immediately retreat when called out to.

[0050] Other embodiments will be described below with reference to the drawings, but descriptions of parts that are the same as or similar to the first embodiment will be omitted as appropriate.

[0051] Second embodiment In the first embodiment, the verbalization process to generate a call-out sentence from the suspicious person video and the voice conversion process to convert the call-out sentence into voice are performed by the server 20, whereas in the second embodiment, the verbalization process and the voice conversion process are performed by the vehicle 10.

[0052] 4 is a diagram showing an example of a hardware configuration of the security system 2 of the second embodiment. The vehicle 10 in the security system 2 of the present embodiment includes a camera 11, an edge AI device 12, an ECU 13, a speaker 15, a horn 16, a light 17, and a neuromorphic system 31.

[0053] The neuromorphic system 31 is an information processing system configured using hardware or software designed to mimic neural circuits. The neuromorphic system 31 is connected to the ECU 13 and the like via a network such as a Controller Area Network (CAN), and is capable of performing advanced information processing, learning processing, and the like using a trained model, etc., based on data output from the ECU 13 and the like.

[0054] Fig. 5 is a diagram showing an example of the functional configuration of the security system 2 of the second embodiment. As shown in Fig. 5, the sentence generation unit 202 and the voice generation unit 203 of this embodiment are realized by using the neuromorphic system 31 mounted on the vehicle 10. That is, according to this embodiment, detection of a suspicious person by image analysis processing of the video (captured by the camera 11) acquired by the acquisition unit 101, generation of a message to call out to the suspicious person by verbalization processing, and vocalization of the message to call out are performed inside the vehicle 10.

[0055] As described above, according to this embodiment, it is possible to call out to a suspicious person even when communication with an external device such as a server cannot be established.

[0056] The program for causing an information processing device to execute the process for realizing the functions of the security systems 1 and 2 of the above-mentioned embodiments may be provided as a computer program product by being stored in a computer-readable storage medium such as a CD-ROM, CD-R, memory card, DVD (Digital Versatile Disk), or flexible disk (FD) in an installable or executable format file. The program may also be provided by being stored on a computer connected to a network such as the Internet and downloaded via the network. The program may also be provided or distributed via a network such as the Internet.

[0057] Although the embodiment of the present invention has been described above, the above embodiment is presented as an example and is not intended to limit the scope of the invention. This new embodiment can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the gist of the invention. This embodiment and its modifications are included in the scope and gist of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0058] 1, 2... security system, 10... vehicle, 11... camera, 20... server, 101... acquisition unit, 102... detection unit, 105... output unit, 106... recording unit, 202... text generation unit, 203... voice generation unit

Claims

1. A detection unit that detects a suspicious person present around the vehicle by image analysis processing of an image captured by a camera mounted on the vehicle; a sentence generation unit that generates a message to call out to the suspicious person by performing a verbalization process on the video including the suspicious person; a voice generating unit that generates voice data for vocalizing the greeting; an output unit that outputs a voice that vocalizes the voice message based on the voice data; A security system equipped with

2. The verbalization process includes a captioning process that generates a caption that describes the video including the suspicious person, and a natural language process that uses the caption to generate the call-out sentence. The security system of claim 1.

3. the image analysis process includes a person detection process for determining whether or not a person is included in the video, a tracking process for tracking the person detected by the person detection process, and a posture estimation process for estimating a posture of the person detected by the person detection process; the detection unit determines whether or not the suspicious person is present based on a result of the person detection process, a result of the tracking process, and a result of the posture estimation process. The security system of claim 1.

4. a recording unit that records an image including the suspicious person when the suspicious person continues to exist after the voice calling out is output a predetermined number of times; The security system according to any one of claims 1 to 3, further comprising:

Images