KR101124371B1 - Speech recognition system for crime prevention in dangerous area - Google Patents

Abstract

The present invention relates to a voice recognition crime prevention alarm system, comprising: a voice image input module (100) for inputting a voice and a video signal inside a dangerous area; A storage block module 200 for storing audio and video signals input from the audio video input module 100; A voice analysis module 300 which analyzes the voice signal input from the voice image input module 100 to classify the voice signal according to the risk level or generates an alarm based on the classified voice signal; It characterized in that it comprises a power block 600 for applying power to the voice image input module 100, the storage block module 200, the voice analysis module 300.
According to the present invention, it is possible to provide a crime prevention warning system in a dangerous area that detects a crime occurrence situation in a dangerous area and informs it to outside through voice recognition, and in particular, a criminal activity such as an elevator or a security weak area rarely occurs. Detects and reads voice signals, words promised in advance, and abnormal noises that may have occurred inside a dangerous situation, and informs external recipients of the situation where the dangerous situation occurred, along with information on the situation according to the level of danger. By providing a danger signal transmission and reception system that can record and transmit the front and rear state as video and audio data, there is an effect that can effectively implement a prior crime prevention.

Description

Speech recognition system for crime prevention in dangerous area}

The present invention relates to a crime prevention warning system in a dangerous area that detects a crime occurrence situation in a dangerous area in real time through voice recognition and informs it to the outside.

In general, as residential buildings and office buildings become larger, frequent accidents occur in rare places such as elevators and parking lots. In particular, since elevators are closed from the outside in a closed place, criminal activity has occurred frequently in recent years. In closed places such as elevators or parking lots, or in places where there are no people, the voices of victims screaming or asking for help are often not transmitted to the outside, leading to great damage.

Of course, in order to solve this problem, the unmanned security facilities such as CCTVs are equipped to monitor and prevent the situation of crime, but in reality, there is a problem that the personnel who can continuously check these unmanned expenses cannot reside. Since it is used as a data for post-mortem visits that identify the criminal's personal situation and the circumstances of the incident after occurrence, the effectiveness of crime prevention is practically low.

Therefore, when a criminal act occurs in a dangerous area such as a confined space, it analyzes the pattern of the voice signal of the victim and informs it to the outside in real time, and at the same time alerts the offender to alert the offender to prevent crime in advance. The need for an existing system is growing.

The present invention has been made to solve the above-described problems, an object of the present invention can provide a crime prevention alarm system in the danger zone to detect and notify the crime occurrence situation in the danger zone through the voice recognition, in particular, Places where dangerous situations occur to external destinations by detecting and reading voice signals, word promised in advance, abnormal noise, etc. It is to provide a danger signal transmission and reception system that can inform the situation information according to the risk stage along with the information, and record and transmit the state before and after the risk situation as video and audio data.

The configuration of the present invention for solving the above problems, the audio image input module 100 and the audio and video scene holol input from the audio image input module 100 for inputting audio and video signals in the danger zone The voice analysis module 300 analyzes the voice signal input from the storage block module 200 and the voice image input module 100 to classify the voice signal according to the risk level or to generate an alarm based on the classified voice signal. It can provide a voice recognition crime prevention alarm system comprising a power block 600 for applying power to the voice image input module 100, the storage block module 200, the voice analysis module 300. Make sure

In particular, the speech analysis module 300, A / D converter 310 for converting an analog signal into a digital signal; A first voice pattern analysis module 320 for comparing and analyzing whether or not the danger setting reference voice signal is through the digital signal; An alarm generator 340 which generates an alarm sound when the first sound pattern analysis module 320 is classified as a danger setting reference voice signal; A voice synthesizer 330 for synthesizing a signal determined as a risk setting reference voice signal with a square wave in the first voice pattern analysis module 320; A pulse controller 350 for synthesizing the speech signal synthesized with the square wave with a triangular wave to generate a square wave having a different duty; It may be implemented as a voice recognition crime prevention alarm system comprising a; first control module 360 for controlling the components of the voice analysis module 300.

In addition, the voice recognition crime prevention alarm system according to the present invention may further comprise a communication module 400 for transmitting the input voice signal to the external server 500.

In this case, the external server 500 may include: a receiver 510 for receiving a voice signal classified as an input voice signal or a dangerous signal; A filter unit 520 for filtering a frequency of an input voice signal; A second voice pattern analysis module 530 for comparing and analyzing whether the input voice signal is a voice signal before being classified as a dangerous signal, whether the voice signal is a danger setting reference voice signal by itself; When the input voice signal is a signal classified as a dangerous signal or is classified as a danger setting reference voice signal by the second voice pattern analysis module 530, the first control module 360 or the second control module 550 itself. It may be configured to include; alarm signal indicating unit 540 for instructing the generation of the alarm sound through.

In the system according to the present invention, the storage block module includes: a booting means for executing a booting and operating environment of the storage block module; and a temporary storage unit for temporarily storing audio and video; And a portable storage unit capable of storing audio and video in a portable storage unit. In addition, the filter unit may be implemented to pass only a signal between 30KHz ~ 100MHz.

According to the present invention, it is possible to provide a crime prevention warning system in a dangerous area that detects a crime occurrence situation in a dangerous area and informs the outside through voice recognition, and in particular, a criminal activity such as a rare elevator or a security weak area frequently occurs. Detects and reads voice signals, words promised in advance, and abnormal noises that may have occurred inside a dangerous situation, and informs external recipients of the situation where the dangerous situation occurred, along with information on the situation according to the level of danger. By providing a danger signal transmission and reception system that can record and transmit the front and rear state as video and audio data, there is an effect that can effectively implement a prior crime prevention.

1 is a block diagram showing the overall configuration according to the present invention.
Figure 2 is a block diagram showing a strip block module according to the present invention.
3 is a conceptual view illustrating the operation of the system and the storage block according to the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail the configuration and operation according to the present invention. In the description with reference to the accompanying drawings, the same components are given the same reference numerals regardless of the reference numerals, and duplicate description thereof will be omitted. The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

According to the present invention, a particular word or a special sound or word group mainly generated in a specific crime scene is prepared through a voice recognition module, and at the same time, a crime event is concerned by integrating a video information device such as CCTV or DVR. The main point is to provide an alarm system that can be installed at frequent places to grasp the level of danger and alarm the site and central monitoring system while recording and storing the audio and video situation before and after the occurrence of the event.

Figure 1 shows an overall block diagram according to the present invention.

Voice recognition crime prevention alarm system according to the present invention (hereinafter referred to as "this system") is in the audio image input module 100 and the audio image input module 100 for inputting the voice and video signal inside the danger zone Analyzing the audio signal input from the storage block module 200 and the audio image input module 100 to store the input voice and the image sinhol, and classify the audio signal by risk level, or alarm based on the classified voice signal Generating a voice analysis module 300 and the voice image input module 100, the storage block module 200, and a power block 600 for applying power to the voice analysis module 300 is configured.

In particular, the audio image input module 100 includes equipment for inputting audio and video generated in a crime situation in the system installed in a specific region. Means a closed and unhuman place, such as a parking lot, a toilet, a rooftop, a parking lot of a private house, a trailing place, a park, a crime-prone area, etc.), and the audio image input module 100 includes a voice input module. It can be configured as a voice sensor for picking up a voice signal, and generally can be configured as a condenser microphone. In addition, it is preferable to include an image input module (CAM) 114 that is composed of a DVR or CCTV that can shoot the image in the specific region described above. In this case, the voice input module receives the external voice through the microphone 111, amplifies the voice signal through the voice amplifier 112, converts the digital signal into a digital signal, and transmits the digital signal to a voice pattern analysis module to be described later.

The image input module receives an image of an external camera and transfers the image to the CPU. That is, the voice amplifier 112 amplifies the signal input from the voice input microphone 111 by 100 times or more by using a preamp to more surely detect the voice signal. Of course, in this case, it is preferable to have a filter to remove the ambient noise, and the sound pressure that is recognized through the voice detection is passed through a switch that operates when a predetermined level or more (85 to 100db). The voice signal from the voice amplifier 112 is introduced into the voice analysis module 300, in which case the analog signal is converted into a digital signal through the A / D converter 310, more preferably the voice recognition chip In order to facilitate this detection, the processor is processed once more. The processing of the processor preferably uses a digital sound processor to compensate for the digital error in the process of converting the analog signal to the digital signal or to optimize the signal so that the analog voice is close to the algorithm to be implemented in the present invention.

In addition, the voice analysis module 300, the A / D converter 310 for converting an analog signal, which is a voice signal input from the voice and video input module 100, into a digital signal and the risk setting reference voice signal through the digital signal. A first voice pattern analysis module 320 for analyzing whether or not the recognition is performed, an alarm generator 340 for generating an alarm sound when the first voice pattern analysis module 320 is classified as a risk setting reference voice signal, In the first voice pattern analysis module 320, a voice synthesizer 330 for synthesizing a signal determined as a danger setting reference voice signal with a square wave, and generating a square wave having a different duty by synthesizing the voice signal synthesized with the square wave with a triangular wave. It may be configured to include a pulse control unit 350 and the first control module 360 for controlling the components of the voice analysis module 300. The voice analysis module 300 is particularly the first voice pattern analysis. In module 320 It is classified as a risk setting reference voice signal by itself, and when it is determined as a dangerous signal, it operates to drive an alarm generator 340 that generates an alarm sound by itself, or transmits a voice signal to an external server and analyzes it in an external server. It can also be implemented to generate an alarm.

(1) When driven by sending to an external server

The first voice pattern analysis module 320 may specifically determine whether or not a risk situation by using the input digital voice signal, and in case of a danger signal, may transmit the warning signal to an external server.

That is, the risk setting reference voice signal is set in advance, and it compares and analyzes whether or not it matches the reference voice signal, and the control of the voice pattern analysis is performed by the first control module 340 implemented by a microprocessor. do. If it is determined that the analysis result corresponds to the risk setting reference voice signal, the detected voice signal is synthesized with the square wave by the voice synthesizer 330. Thereafter, the speech signal synthesized with the square wave is combined with the triangular wave to pass a pulse control unit (PWM) 350 for generating square waves having different duty. The pulse controller 350 is preferably formed of PWM (pluse width modulation) for pulse width modulation. The voice signal determined as the danger signal is transmitted to the external server 500 through the communication module 400. The communication module 400 may use various communication methods, and may be implemented by using power line communication or by mounting an interface IC such as a LAN.

The external server 500 is a concept of managing a device that can receive a signal transmitted through the communication module. In this embodiment, the receiver 510 for receiving an input signal and a filter 520 for filtering out unnecessary noise. ) May be included. In the filter unit 510, the 60Hz signal, which is a general household frequency signal, may be cut off and set to pass only a 30KHz to 100MHz signal. In addition, when the input signal is a signal classified as a dangerous signal in the first pattern analysis module 320 of the voice analysis module 300 is configured to further include an alarm signal transmitting unit 540 that can immediately send an alarm signal. This is preferred.

In particular, in the present system, when the signal transmitted from the voice analysis module 300 is not a voice signal classified as a dangerous classification, but a general voice signal (unclassified word), the external server 500 independently signals a dangerous signal. It may be configured to further include a second voice pattern analysis module 530 to determine whether or not. In this case, even if the input voice signal is not classified as a dangerous classification, a more accurate classification can be performed based on a larger volume of voice classification information. In 540, the alarm signal may be simultaneously generated through the first control module 360 and the second control module 550. The alarm signal may be formed of an alarm lamp such as an LED or a structure in which an alarm of a voice type for generating a siren sound is combined.

(2) In case of not transmitting to external server

In the present system, when a voice signal cannot be transmitted to an external server, for example, even when a communication module does not operate, the voice pattern is analyzed in the voice analysis module 300 itself, and then classified as a dangerous signal. The control module 360 may be configured to drive the alarm generator 340 to generate an alarm sound. This can be reduced in terms of the capacity of the classification of risk words compared to the method of combining with an external server, but it is very useful in that the crime prevention effect can be realized at all times by simply attaching a convenient terminal (main system) in case of emergency. Can be.

In particular, the above-described system is advantageous in that it is inexpensive and secures mobility. In particular, it can be implemented as a stand-alone device in the form of an embedded device without a PC, and can be operated independently, and has a voice black box function before and after an event. Through the recording of audio / video data according to the situation, there is an advantage of providing a powerful evidence capability as well as crime prevention function.

The storage block module 200 according to the present invention for performing such a voice black box function will be described with reference to FIGS. 1 to 3.

The storage block module 200 according to the present invention is a bootable means for executing a booting and operating environment of the storage block module (Booting) 210 and a temporary storage unit 220 for temporarily storing audio and video and a removable audio and video. It may be made including a portable storage unit 230 that can be stored in the storage means.

The booting means 210 performs a function of controlling a booting and a dynamic environment using a NOR flash, and the temporary storage unit 220 is a space for storing a program, voice, and projection. It can be implemented with SDRAM. In addition, the portable storage unit may include a NAND flash program as a modified FAT system as a space for storing audio and video, and an SD card or T-Flash used as a removable medium for recording audio and video.

In the storage block module 200, the voice and video are pre-recorded from a predetermined time (about 1 minute before the occurrence of the event) and recorded in an internal temporary storage unit, and the buffer can be buffered so that the pre-condition can be stored afterwards. In addition, audio and video data over a certain time (about 2 minutes or more) is recorded in the file system. In this case, recording starts automatically by the volume of the voice or the word measure of the recognized voice. In addition, the internal power supply (battery) is implemented so that it can be operated safely even when the supply of external power is cut off, and when an event occurs, a warning is generated externally using the communication module through the voice analysis module simultaneously. Make it work. Video recording can be recorded with MPEG4 SP, and audio can be recorded with AAC 22.05KHz. The storage block module 200 may be miniaturized or independently implemented to ensure external impact resistance and to facilitate installation.

Figure 4 shows an example of the implementation of the software for operating the system including the speech analysis module of the present system.

1) Central processing unit (Renesas SH mobile)

The central processing unit (Renesas SH mobile) is a CPU for controlling the speech analysis module in the structure according to the present invention described above, the Renesas Techonlogy's CPU is configured as a base model. Of course, if the CPU has a voice processing function, it can be used with other CPUs.

2) Driving system (U-ITRON (Real time O / S))

It consists of software as a real-time operating system (Operating System) for coordinating the entire resource on the H / W structure of the CPU (CPU). It handles functions (Event Flag, Semaphore) imposed internally on memory allocation and return, and various synchronization functions, and enables multi-task execution.

3) Driver System

It is a software module that manages various communication and HW interface, and builds environment for application program to work.

4) FAT SYSTEM

It uses the FAT method as a file system to store and read internal data on NAND and external storage media.

5) Compressor (MP4 / AAC Firmware)

MP4: MPEG4 compression technology is firmware for compressing video signals. The video inputted by the camera, which is an image input module, needs to be compressed and stored because there is a lot of data. The compressor used at this time.

AAC: MPEG audio compression technology that compresses audio data while maintaining the original sound. Firmware that uses this technique is built-in.

6) Sound Recognition Engine

As a speech recognition engine, it is an engine that processes externally inputted voices, converts them into specific words, and extracts base words for processing by the system. It is composed of software and is a software module with advanced voice processing technology.

7) OKYES HTML UI Engine

It is a software module developed by OKYES Electronics Co., Ltd. to operate HTML processing device in embedded system. It displays the operation of the system to USER by printing letters and pictures on the screen.

In the foregoing detailed description of the present invention, specific examples have been described. However, various modifications are possible within the scope of the present invention. The technical idea of the present invention should not be limited to the embodiments of the present invention but should be determined by the equivalents of the claims and the claims.

100: audio image input module
111: voice input module
112: negative amplifier
113: Frequency Analysis Welch Switch
114: video input module
200: storage block module
210: boot section
220: temporary storage
230: removable storage
300: voice analysis module
310: analog digital converter
320: first voice pattern analysis module
330: voice synthesizer
340: alarm generating unit
350: pulse width control unit
360: first control module
400: communication module
500: external server
510: receiver
520: filter
530: second speech pattern analysis module
540: alarm signal transmitting unit
550: second control module

Claims (6)

An audio image input module 100 for inputting an audio and video signal inside the dangerous area; A storage block module 200 for storing audio and video signals input from the audio video input module 100; A voice analysis module 300 which analyzes the voice signal input from the voice image input module 100 to classify the voice signal according to the risk level or generates an alarm based on the classified voice signal; And a power block 600 for applying power to the voice image input module 100, the storage block module 200, and the voice analysis module 300.
The voice analysis module 300, the A / D converter 310 for converting an analog signal into a digital signal; A first voice pattern analysis module 320 for comparing and analyzing whether or not the danger setting reference voice signal is through the digital signal; An alarm generator 340 which generates an alarm sound when the first sound pattern analysis module 320 is classified as a danger setting reference voice signal; A voice synthesizer 330 for synthesizing a signal determined as a risk setting reference voice signal with a square wave in the first voice pattern analysis module 320; A pulse controller 350 for synthesizing the speech signal synthesized with the square wave with a triangular wave to generate a square wave having a different duty; And a first control module 360 for controlling the components of the voice analysis module 300.
The storage block module 200 includes: a booting means (Booting 210) for executing a booting and operating environment of the storage block module; and a temporary storage unit 220 for temporarily storing audio and video; And a portable storage unit 230 capable of storing audio and video in a portable storage unit.
Voice recognition crime prevention alarm system, characterized in that comprises a.
delete The method according to claim 1,
The voice recognition crime prevention alarm system,
Voice recognition crime prevention alarm system further comprises a communication module for transmitting the input voice signal to the external server (500).
The method according to claim 3,
The external server 500,
A receiver 510 for receiving an input voice signal or a voice signal classified as a dangerous signal;
A filter unit 520 for filtering a frequency of an input voice signal;
A second voice pattern analysis module 530 for comparing and analyzing whether the input voice signal is a voice signal before being classified as a dangerous signal, whether the voice signal is a danger setting reference voice signal by itself;
When the input voice signal is a signal classified as a dangerous signal or is classified as a danger setting reference voice signal by the second voice pattern analysis module 530, the first control module 360 or the second control module 550 itself. An alarm signal transmitter 540 instructing generation of an alarm sound through;
Voice recognition crime prevention alarm system, characterized in that consisting of.
delete The method of claim 4,
The filter unit is a voice recognition crime prevention alarm system, characterized in that passing only the signal between 30KHz ~ 100MHz.

Images