KR102839672B1 - Smart emergency bell system equipped with a fusion sensor that improve recognition rate of de-identification images - Google Patents

Abstract

The present invention relates to an emergency bell system that clearly derives a low-resolution de-identified image of a surveillance subject using a fusion sensor that links an infrared array sensor and a ToF (Time to Flight) array sensor, and detects illegal filming or an emergency situation in a specific security target space based on the derived low-resolution de-identified image, and performs automatic reporting and warning broadcast transmission based on the detection result, and more specifically, the system comprises: a fusion sensor unit that includes two or more heterogeneous detection sensors arranged in a specific security target space, and obtains a de-identified image of a surveillance subject by fusing data output from each detection sensor and outputs the obtained de-identified image; an analysis device that receives the de-identified image transmitted by the fusion sensor unit, analyzes the received de-identified image to determine the motion and mobility of the surveillance subject, derives a behavioral pattern based on the determined motion and mobility of the surveillance subject, and detects and outputs whether the surveillance subject is behaving abnormally from the derived behavioral pattern; And it is characterized by including an emergency bell device that selectively generates a rescue request signal based on a user operation or a detection result output from the analysis device and transmits the same to one or more management agencies so that it is transmitted visually or audibly in the field.

Description

Smart emergency bell system equipped with a fusion sensor that improves recognition rate of de-identification images

The present invention relates to an emergency bell system that uses a fusion sensor that links an infrared array sensor and a ToF (Time to Flight) array sensor to clearly derive a low-resolution, non-identifiable image of a surveillance subject, and detects illegal filming or an emergency situation in a specific surveillance target space based on the derived low-resolution, non-identifiable image, and performs automatic reporting and warning broadcast transmission in conjunction with the detection result.

Our country has been continuously carrying out public restroom modernization projects since around the year 2000, and is dedicated to improving aging sanitary facilities and creating a world-class public restroom environment.

However, according to statistics from the Korea Local Administration Research Institute, public restroom users continue to raise safety issues such as illegal filming and random assault due to inconveniences in using the restroom.

Accordingly, with the revision of the Public Restroom Act in June 2023, the installation of "emergency bells" in public restrooms became mandatory, providing a mechanism to respond to safety/crime. However, there is a problem of waste of public power due to the passive reporting system and false reports, and it is urgent to prepare countermeasures to resolve this.

Recently, various services utilizing ICT technology have been proposed to respond to illegal filming in public restrooms, random assaults, and various safety accidents. However, due to the environmental characteristics that prioritize personal information protection, the technology and scope of application are extremely limited. In particular, in the case of public restrooms, it is impossible to establish a monitoring environment using video equipment, so there are many difficulties in implementing the services.

To replace this, technologies are being introduced that utilize 4D imaging radar sensors, ToF sensors, infrared sensors, etc. to visualize the subject of surveillance as a dot image, and based on this, recognize movements and detect dangerous situations.

However, the data reception method and technical characteristics of each sensor limit the scope of surveillance target recognition. For example, 4D imaging radar sensors can interpret structural features across an entire space, but their ability to specifically identify (or act upon) surveillance targets based on changes in those features is severely limited.

And in the case of ToF sensors (or lidar, radar, etc.), specific identification is possible through the distance difference between the surveillance target and the background, but it is impossible to identify whether the surveillance target is a person or an object, which can lower the accuracy of the motion recognition rate.

In addition, infrared sensors can use thermal information from surveillance footage to determine whether a person is present, but they are sensitive to external temperature changes and have limitations in recognizing small, low-temperature extremities (hands, arms, etc.) of the target.

Republic of Korea Patent No. 10-2337790 Republic of Korea Patent No. 10-2110945 Republic of Korea Publication Patent No. 10-2022-0134821 Republic of Korea Patent No. 10-2486299 Daehinmanguk Publication Patent No. 10-2024-0098199

Accordingly, the present invention has been devised to solve the above-described problems, and the purpose of the present invention is to provide a smart emergency bell system that can improve the reliability of detection results of abnormal behaviors such as illegal filming or emergency situations by applying multimodal sensing technology that fuses heterogeneous sensors to more accurately derive various motion recognition and corresponding behavioral patterns of a surveillance subject.

As a means for solving the above-described problem, a smart emergency bell system equipped with a fusion sensor that improves the motion recognition rate of an unidentifiable image of the present invention (hereinafter referred to as the "system of the present invention") comprises: an infrared array sensor installed in a specific security target space, in which a plurality of infrared sensors are arranged to obtain pixel data digitizing the temperature of a surveillance target; a ToF (Time to Flight) array sensor in which a plurality of ToF sensors are arranged to obtain pixel data digitizing the distance of the surveillance target; a fusion sensor unit that fuses data output from each sensor to obtain an unidentifiable image of the surveillance target and outputs the obtained unidentifiable image; an analysis device that receives an unidentifiable image transmitted from the fusion sensor unit, analyzes the received unidentifiable image to determine motion and mobility of the surveillance target, derives a behavioral pattern based on the determined motion and mobility of the surveillance target, and detects and outputs whether the surveillance target exhibits abnormal behavior from the derived behavioral pattern; And an emergency bell device that selectively generates a rescue request signal according to a user operation or a detection result output from the analysis device and transmits the same to one or more management agencies while operating to be visually or audibly transmitted in the field; wherein the analysis device comprises: a motion recognition unit that specifies the shape of an unidentifiable image received from the fusion sensor unit and classifies the specified image shape to recognize it as a selected motion; a mobility analysis unit that analyzes mobility including a movement direction and movement speed by tracking pixel changes according to the movement of a surveillance subject in the unidentifiable image received from the fusion sensor unit; a pattern derivation unit that combines the motion of the surveillance subject recognized by the motion recognition unit and the mobility of the surveillance subject analyzed by the mobility analysis unit to derive a behavior pattern in a time series format; an abnormal behavior detection unit that detects any one of illegal filming, fainting, falling, and assault based on the behavior pattern derived from the pattern derivation unit; and a signal transmission unit that selectively outputs one or more response signals according to the detection result of the abnormal behavior detection unit and transmits the output one or more response signals to the emergency bell device. Including, the fusion sensor unit is provided with two or more to build a sensor network, and the two or more fusion sensor units are installed at mutually spaced locations within the specific security target space, and the analysis device is characterized in that it collects each individual non-identifiable image output from the two or more fusion sensor units installed at mutually spaced locations to create a multi-view non-identifiable image set, and analyzes the created multi-view non-identifiable image set to determine a three-dimensional movement of the surveillance target.

delete

delete

As an example, the fusion sensor unit is characterized by including a coordinate system setting module that corrects a coordinate system error according to a focal difference between the infrared array sensor and the ToF array sensor, a resolution correction module that corrects the displayed shape of a surveillance subject according to a difference in field of view between the infrared array sensor and the ToF array sensor, and a data fusion module that matches and fuses pixel data of the infrared array sensor and pixel data of the ToF array sensor corrected by the coordinate system setting module and the resolution correction module to generate an unidentifiable image.

delete

delete

As an example, the emergency bell device is characterized by including one or more situation transmission modules that receive a response signal transmitted from the analysis device and operate to visually or audibly transmit a dangerous situation at the site in response to input of the response signal.

As an example, the emergency bell device is characterized by including an automatic reporting module that receives a response signal transmitted from the analysis device and generates the rescue request signal in response to input of the response signal.

The system of the present invention has the effect of being able to derive a clearer low-resolution non-identifiable image of a surveillance target in a specific security target space by compensating for the shortcomings of each sensor in obtaining a non-identifiable image by applying a fusion sensor that links an infrared array sensor and a ToF (Time to Flight) array sensor.

In addition, it has the effect of further improving the reliability of detection results for abnormal behaviors such as illegal filming or emergency situations by enabling recognition of various movements of the subject of surveillance and deriving behavioral patterns based on the non-identifiable images derived from the fusion sensor that fuses heterogeneous sensors.

Additionally, when abnormal behavior of the subject of surveillance is detected, it automatically reports the situation to the management agency by linking with the emergency bell device and transmits a warning broadcast on site, which has the effect of enabling the manager to quickly check the site and take response measures.

Figure 1 is a drawing schematically showing the configuration of a system according to one embodiment of the present invention.
Figure 2 is a block diagram showing the configuration of a fusion sensor unit according to one embodiment of the present invention.
Figures 3a to 3c are diagrams showing a data fusion process of a fusion sensor unit according to one embodiment of the present invention.
Figure 4 is a block diagram showing the configuration of an analysis device according to one embodiment of the present invention.
FIG. 5 is a drawing showing a method for recognizing the motion of a surveillance subject according to one embodiment of the present invention.
FIG. 6 is a diagram illustrating a method for analyzing the mobility of a surveillance subject according to an embodiment of the present invention.
FIG. 7 is a diagram illustrating a method for recognizing the motion of a surveillance subject using re-point de-identified image data according to an embodiment of the present invention.

In describing the present invention, terms or words used in this specification and claims should be interpreted as meanings and concepts that conform to the technical idea of the present invention, based on the principle that the inventor can appropriately define the concept of the term to describe his or her own invention in the best way.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.

Referring to FIG. 1, the system of the present invention can be configured to include a fusion sensor unit (10), an analysis device (30), and an emergency bell device (20).

The above emergency bell device (20) is installed in a specific security target space (1) and can generate a rescue request signal according to user operation, and transmit it to one or more selected management agencies to notify the relevant agencies of an emergency situation such as a disaster, calamity, or crime.

In addition, the above-mentioned emergency bell device (20) can be operated to transmit an alarm sound or warning broadcast in conjunction with the generation of a rescue request signal in the above-mentioned emergency situation.

The above emergency bell device (20) can optionally attempt to connect a call with a person in charge of the management agency along with generating a rescue request signal, and can not only help the person in charge of the management agency determine the severity of the emergency situation through the information obtained through the call about the on-site situation, but can also instruct and guide the reporter on the initial response through the person in charge's voice.

Here, the above management agency may be one or more of the local government police station or fire station, the local government civil complaint situation room, the district office in charge of the crime prevention target area, or the autonomous crime prevention unit.

In addition, the above management agency can monitor the rescue request signal transmitted from the above emergency bell device (20), and can support a quick response to an emergency situation such as a crime by dispatching and deploying management personnel such as the police to the scene by considering the severity of the emergency situation through a call with the reporter, etc.

The above emergency bell device (20) is not shown in the drawing, but is equipped with a rescue request button exposed externally to generate and output a rescue request signal according to the user's command, and may include a communication module that provides data communication with the management agency and a handset for call connection with the management agency.

The above fusion sensor unit (10) is composed of heterogeneous sensors and fuses data detected from each sensor to obtain an unidentifiable image, thereby further improving the motion recognition rate for a surveillance target within the specific security target space (1).

Here, the above-mentioned emergency bell device (20) can be installed in a place that is easily exposed to crime, for example, the specific crime prevention target space (1) can be a public restroom.

In the case of public restrooms, since it is impossible to establish a monitoring environment using video equipment due to the environmental characteristics that place importance on personal information protection, the fusion sensor unit (10) utilizes an unidentifiable image for the purpose of detecting the subject of surveillance in the space.

Accordingly, the above-mentioned fusion sensor unit (10) may include two or more heterogeneous detection sensors (100, 110) placed in a specific security target space (1), and may fuse data output from each detection sensor (100, 110) to obtain an unidentifiable image of the surveillance target and output the obtained unidentifiable image.

For example, the fusion sensor unit (10) may include an infrared array sensor (100) in which a plurality of infrared sensors are arranged to obtain pixel data that digitizes the temperature of the subject of surveillance, and a ToF (Time to Flight) array sensor (110) in which a plurality of ToF sensors are arranged to obtain pixel data that digitizes the distance of the subject of surveillance.

The above infrared sensor cannot detect features of the extremities of the human body (hands, arms, legs, etc.) with relatively low blood flow due to its thermal pattern detection characteristics, whereas the ToF sensor can detect such features.

In addition, in the case of the above ToF sensor, since it cannot distinguish between a human body and an object, misrecognition may occur due to other objects (bags, clothing, etc.) in the surveillance area, whereas the above infrared sensor can identify the central shape of the human body through heat information.

In the present invention, by combining and utilizing the characteristics of these two sensors, it is possible to more clearly derive basic data necessary for detecting the motion of a subject of surveillance, that is, a low-resolution, non-identifiable image that cannot identify personal information such as the face of the subject of surveillance.

First, the plurality of pixel data output from the infrared array sensor (100) and the ToF array sensor (110) respectively need to be corrected due to structural and functional differences such as focus and field of view (FoV).

Accordingly, the fusion sensor unit (10) according to one embodiment of the present invention may include a coordinate system setting module (120), a resolution correction module (130), and a data fusion module (140), as illustrated in FIG. 2, to fuse data output from the infrared array sensor (100) and the ToF array sensor (11) to create an unidentifiable image that can be used for motion analysis.

The above coordinate system setting module (120) can correct the coordinate system error according to the focal difference of the infrared array sensor and the ToF array sensor.

And the above resolution correction module (130) can correct the displayed shape of the surveillance target according to the difference in the angle of view of the infrared array sensor (100) and the ToF array sensor (110).

In addition, the data fusion module (140) can match and fuse the pixel data of the infrared array sensor (100) corrected by the coordinate system setting module (120) and the resolution correction module (130) and the pixel data of the ToF array sensor (110) to create an unidentifiable image.

Hereinafter, the data fusion process of the fusion sensor unit (10) according to one embodiment of the present invention will be described in more detail with reference to FIGS. 3a to 3c.

Figure 3a shows measurements of a surveillance subject received from an infrared sensor and a ToF sensor, which are then visualized. Since the infrared sensor provides temperature information of the surveillance area in numerical form, low-quality data that makes target identification difficult may be produced when the temperature difference between the space and the surveillance subject is small. This characteristic can be confirmed in ① of Figure 3a.

Preprocessing filters can be applied to obtain a motion profile of a surveillance subject that is robust to environmental factors. For example, as shown in ② of Figure 3a, a sigmoid filter can be applied to isolate the temperature difference between the surveillance subject and the space. However, patterns not included in the surveillance subject may still be identified.

Meanwhile, the ToF array sensor (110) quantifies the distance to the surveillance target and shows a clear contrast between the space and the surveillance target, as shown in ③ of Fig. 3a. However, it can be confirmed that the information displayed on the surveillance target differs due to differences in focus and field of view between the two sensors.

As in ② and ③ of Fig. 3a, the difference in the center axis of the target may be caused by the difference in focus between the two sensors, and the displayed pixel size of the surveillance target may be caused by the difference in the field of view between the two sensors.

Accordingly, in the fusion sensor unit (10) of the present invention, in order to fuse the data output from the infrared sensor and the ToF sensor, the ToF sensor's zero-resolution alignment preprocessing can be performed.

For example, the coordinate system setting module (120) can adjust the ToF input data of Fig. 3b ① to a shape like Fig. 3b ②. At this time, the image coordinate system can be adjusted using prior information such as the focal length and angle of view of the infrared sensor and the ToF sensor.

In addition, the data of FIG. 3b ② can be corrected to the shape of FIG. 3b ③ through the above resolution correction module (130).

At this point, padding pixels can be added to the original image and reduced through a pooling process. The periphery of the subject (the extremities) will display low numerical data, while the center will display high numerical data. These preprocessed image characteristics reflect the characteristics of infrared imaging of the center and extremities of the human body.

Afterwards, the above fusion sensor unit (10) can obtain an unidentifiable image that can clearly recognize the movement (shape) of the surveillance subject through the data fusion module (140), which can be confirmed through FIG. 3c.

For example, the infrared image of Fig. 3c ① and the ToF preprocessing image values of Fig. 3c ② can be used to process a weighted masking image. This allows for the acquisition of a high-quality, non-identifiable image, such as Fig. 3c ③. The non-identifiable image thus output minimizes the spatial thermal information visible in the infrared image, allowing for a clearer representation of the subject's movements.

Meanwhile, the above analysis device (30) is installed adjacent to the fusion sensor unit (10) in a specific security target space (1) such as the fusion sensor unit (10) and can receive a low-resolution unidentifiable image output from the fusion sensor unit (10).

This analysis device (30) performs the function of a gateway that interconnects the fusion sensor unit (10) and the emergency bell device (20). At this time, communication with the fusion sensor unit (10) can utilize any one of Lora, Bluetooth, Zigbee, RFID, IR, and RF, which are short-range wireless communication networks, and a communication module that can apply the corresponding communication network can be mounted.

The above analysis device (30) can analyze the non-identifiable image received from the fusion sensor unit (10) to determine the movement and mobility of the subject of surveillance.

And the above analysis device (30) can derive a behavioral pattern based on the determined movement and mobility of the surveillance subject and detect whether the surveillance subject is behaving abnormally from the derived behavioral pattern.

Specifically, the above analysis device (30) may be configured to include a motion recognition unit (300), a mobility analysis unit (310), a pattern extraction unit (320), an abnormal behavior detection unit (330), and a signal transmission unit (340), as illustrated in FIG. 4.

The above motion recognition unit (300) can specify the shape of an unidentifiable image received from the fusion sensor unit (10) and classify the specified image shape to recognize it as a selected motion.

For example, the above motion recognition unit (300) performs motion shape concretization work to identify the motion of a surveillance subject in a low-resolution image as illustrated in FIG. 3, and this can be implemented by applying a known interpolation method.

And the motion shape specification data can perform motion classification tasks using artificial intelligence techniques including deep learning algorithms.

The above mobility analysis unit (310) can analyze mobility, including movement direction and movement speed, by tracking pixel changes according to the movement of the surveillance subject in the non-identifiable image received from the fusion sensor unit (10).

Referring to FIG. 6, the mobility analysis unit (310) can obtain multiple image frames sequentially in time from the non-identifiable images transmitted from the fusion sensor unit (10), and can confirm the movement direction by the difference between the current image frame and the previous image frame in the multiple image frames obtained.

And by calculating the movement speed using the time between each image frame and the distance the pixels moved, the motion vector can be extracted.

The above pattern derivation unit (320) combines the movement of the surveillance subject recognized by the motion recognition unit (300) and the mobility of the surveillance subject analyzed by the mobility analysis unit (310) to derive a behavior pattern in a time series format, thereby ultimately estimating the type of behavior of the surveillance subject.

The above abnormal behavior detection unit (330) can detect any one of the abnormal behaviors among illegal filming, fainting, falling, and assault based on the behavioral pattern derived from the above pattern derivation unit (320).

For example, the above-mentioned abnormal behavior detection unit (330) can detect whether the behavior pattern derived from the above-mentioned pattern derivation unit (320) corresponds to any one of the above-mentioned abnormal behaviors through comparative analysis with various abnormal behaviors defined through a known artificial intelligence learning model.

That is, the above-mentioned abnormal behavior detection unit (330) analyzes the similarity between the behavior type estimated by the above-mentioned pattern derivation unit (320) and the learned abnormal behavior to detect whether the behavior type of the subject of surveillance is an illegal act such as illegal filming or random assault, or an emergency situation such as fainting or falling.

At this time, the means for detecting abnormal behavior can be selected and applied by selecting an appropriate one among the known artificial intelligence algorithms such as RNN (Recurrent Neural Networks) and LSTM (Long Short-Term Memory model), so a detailed description thereof is omitted.

The signal transmission unit (340) can selectively output one or more response signals based on the detection result of the abnormal behavior detection unit (330), that is, detection of abnormal behavior of the subject of surveillance, and transmit the output one or more response signals to the emergency bell device (20).

Although this type of analysis device (30) is shown as a separate configuration from the emergency bell device (20) in FIG. 1, the present invention is not limited thereto and it is obvious that it may have a form built into the emergency bell device (20).

Meanwhile, the above fusion sensor unit (10) can be provided in two or more units within the security target space (1) to build a sensor network, and can output two or more individual low-resolution unidentifiable images to the emergency bell device (20).

At this time, two or more fusion sensor units (10) may be installed at mutually spaced locations within the specific security target space (1) or may be installed collectively in one casing.

And the above analysis device (20) can compile each low-resolution non-identifiable image output from two or more fusion sensor units (10) to create a multi-view non-identifiable image set, and analyze the created multi-view non-identifiable image set to determine the three-dimensional movement of the surveillance subject.

At this time, the above-mentioned re-identified image can be used as input data for a classification artificial intelligence model for motion analysis.

That is, as illustrated in FIG. 7, the system of the present invention comprises two or more fusion sensor units (10) to construct a sensor network within a security target space (1) and obtain two or more individual low-resolution non-identifiable images from them to recognize depth information about a surveillance target, thereby enabling recognition of more diverse motions and thus ensuring recognition accuracy.

This not only complements the disadvantage of conventional infrared thermal imaging cameras that cannot identify the depth information of a target, but also has the advantage of being able to economically recognize the depth information of a surveillance target without using expensive products such as laser scanners.

Meanwhile, the above emergency bell device (20) may include one or more situation transmission modules (200) that receive a response signal output in response to detection of abnormal behavior of a subject of surveillance from the analysis device (30) and operate to visually or audibly transmit a dangerous situation at the scene in response to the input of the response signal.

For example, the above situation transmission module (200) may include a speaker for outputting an alarm sound or a warning guidance sound, and a warning light or display, etc., that allows the user to visually confirm an emergency situation, and may output an alarm sound or operate a warning light, etc. in conjunction with the input of the response signal, so as to notify the surroundings of an emergency situation or a crime situation at the scene.

In addition, the emergency bell device (20) may further include an automatic signal module (210) that receives a response signal transmitted from the analysis device (30) and generates the rescue request signal in response to the input of the response signal.

That is, the above emergency bell device (20) not only generates a rescue request signal at the user's initiative, such as by operating the rescue request button as mentioned above, but also automatically generates and outputs a rescue request signal when the analysis device (30) detects abnormal behavior of the subject of surveillance, thereby enabling a quick report to be made to the management agency (40).

Although the preferred embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims also fall within the scope of the present invention.

10: Fusion sensor unit 20: Emergency bell device
30: Analysis device 40: Management agency
100: Infrared array sensor 110: ToF array sensor
120: Coordinate system setting module 130: Resolution correction module
140: Data Fusion Module 200: Situation Propagation Module
210: Automatic reporting module 300: Motion recognition unit
310: Mobility Analysis Department 320: Pattern Derivation Department
330: Abnormal behavior detection unit 340: Signal transmission unit

Claims (8)

An infrared array sensor installed in a specific surveillance target space, comprising a plurality of infrared sensors arranged to obtain pixel data digitizing the temperature of a surveillance target, and a ToF (Time to Flight) array sensor arranged in a plurality of ToF sensors arranged to obtain pixel data digitizing the distance of a surveillance target, and a fusion sensor unit that fuses data output from each sensor to obtain an unidentifiable image of the surveillance target and outputs the obtained unidentifiable image;
An analysis device that receives an unidentifiable image transmitted from the above fusion sensor unit, analyzes the received unidentifiable image to determine the movement and mobility of the surveillance subject, derives a behavioral pattern based on the determined movement and mobility of the surveillance subject, and detects and outputs whether the surveillance subject is performing abnormal behavior from the derived behavioral pattern; and
An emergency bell device that selectively generates a rescue request signal based on a user operation or a detection result output from the analysis device and transmits the signal to one or more management agencies so that it is transmitted visually or audibly in the field;
The above analysis device,
It includes a motion recognition unit that specifies the shape of an unidentifiable image received from the fusion sensor unit and classifies the specified image shape to recognize it as a selected motion, a mobility analysis unit that analyzes mobility including the direction of movement and the speed of movement by tracking pixel changes according to the movement of a surveillance subject in the unidentifiable image received from the fusion sensor unit, a pattern derivation unit that derives a behavior pattern in a time series format by combining the motion of the surveillance subject recognized by the motion recognition unit and the mobility of the surveillance subject analyzed by the mobility analysis unit, an abnormal behavior detection unit that detects any one of illegal filming, fainting, falling, and assault based on the behavior pattern derived from the pattern derivation unit, and a signal transmission unit that selectively outputs one or more response signals according to the detection result of the abnormal behavior detection unit and transmits the output one or more response signals to the emergency bell device.
The above fusion sensor unit is provided in two or more parts to build a sensor network, and two or more fusion sensor units are installed at mutually separated locations within the specific security target space.
The above analysis device,
A smart emergency bell system equipped with a fusion sensor that improves the recognition rate of de-identified images, characterized in that it collects individual de-identified images output from two or more fusion sensor units installed at mutually separated locations to create a multi-view de-identified image set, and analyzes the created multi-view de-identified image set to determine the three-dimensional movement of the subject of surveillance.
delete delete In paragraph 1,
The above fusion sensor part,
A smart emergency bell system equipped with a fusion sensor that improves the recognition rate of an unidentifiable image, characterized by including a coordinate system setting module that corrects a coordinate system error according to a focal difference between the infrared array sensor and the ToF array sensor, a resolution correction module that corrects the displayed shape of a surveillance subject according to a difference in field of view between the infrared array sensor and the ToF array sensor, and a data fusion module that matches and fuses pixel data of the infrared array sensor and pixel data of the ToF array sensor corrected by the coordinate system setting module and the resolution correction module to generate an unidentifiable image.
delete delete In paragraph 1,
The above emergency bell device,
A smart emergency bell system equipped with a fusion sensor that improves the recognition rate of anonymized images, characterized by including one or more situation transmission modules that receive a response signal transmitted from the above analysis device and operate to visually or audibly transmit a dangerous situation at the scene in response to the input of the response signal.
In paragraph 1,
The above emergency bell device,
A smart emergency bell system equipped with a fusion sensor that improves the recognition rate of anonymized images, characterized by including an automatic reporting module that receives a response signal transmitted from the above analysis device and generates the rescue request signal in response to the input of the response signal.