KR102596183B1 - In-track train entry detection system and detection method using camera module and radar sensor module - Google Patents

Abstract

The present invention clearly detects the approach of a train (railroad vehicle, motor car, etc.) while performing maintenance work on the track by using a camera module and radar sensor module, and when it is determined that the train is certain to enter the track, the worker This relates to a safety warning system for line workers that prevents the possibility of safety accidents by generating an easily recognizable luminous warning or alarm sound.

Description

In-track train entry detection system and detection method using camera module and radar sensor module}

The present invention clearly detects when a train (railroad vehicle, motor car, etc.) approaches while the worker is performing maintenance work on the track, and when it is determined that the train is certain to enter the track, a light emitting alarm or alarm sound that can be easily recognized by the worker is provided. This relates to a safety warning system for line workers that prevents the possibility of safety accidents occurring.

Trains (railroad vehicles, motor cars, etc.) are a special means of transportation that allow passage on tracks. Normal tracks are exposed to the external environment and parts are subject to deterioration, deformation, and damage due to the periodic load of the railway. It happens.

Therefore, repair work is done frequently. Track repair is directly related to the safe operation of the railroad, and delays in work on the track cause delays in railroad operation, so workers work on the track at the risk of accidents. there is.

Various devices for safe operation of the railroad are installed around these railroad tracks, such as signals, track switching devices, track circuit devices, interlocking devices, and level crossing security devices, but these devices ensure the safety of workers working on the trackside. It was not installed for this purpose; it was just installed to ensure safe operation of the railway.

In such a working environment, accidents in which workers die by being struck by railroad tracks while performing maintenance work on the tracks frequently occur.

Meanwhile, even in the case of subways where the tracks are rarely exposed to the outside, track maintenance work is necessary depending on vehicle operation.

Various technologies have been proposed so far to prevent collision accidents between railroads and workers as described above, and a technology using a wireless alarm method is proposed.

Technologies such as Korea Registered Utility Model Publication No. 20-0219500 (title of the invention: Train approach warning system) have been developed.

The described train approach warning system consists of a transmitting device installed on the railroad to transmit radio waves in the direction of travel of the railroad, and a receiving device attached to the worker's hard hat or seat belt to alert the worker when the radio waves are received. there is.

However, this conventional warning system for line workers has the following problems.

First, there is an economic problem that each train (motor car, etc.) must be equipped with a transmitting device to transmit radio waves, and the cost for this is considerable.

Second, when a train (motor car, etc.) approaches the work space slowly, problems such as loud warnings or alarms that are unnecessary for railroad car work are emerging, reducing workers' concentration on work and causing noise. This is the situation.

Patent Document 1: Domestic Patent Publication No. 20-0219500 (Publication Date: 2001.04.02) Patent Document 2: Domestic Patent Publication No. 10-1817529 (Publication date: 2018.01.11) Patent Document 3: Domestic Patent Publication No. 10-2010-0044339 (Publication Date: 2010.04.30)

In order to solve the above problems, the present invention is a track constructed to prevent safety accidents, etc. The purpose is to provide an internal worker safety support system.

Additionally, the purpose of the present invention is to present a system configured to solve the problem of increased installation costs of the overall system due to a transmitting device mounted inside a train (motor car, etc.) moving on a track.

In addition, the purpose of the present invention is to present a system configured to accurately determine the entry state of a train (motor car, etc.) on the track and prevent work interference or noise generation due to unnecessary warning sounds, etc. in the work space.

In order to achieve the above object, the present invention is a train entry detection system in the track, including a camera module that generates image information when an object enters the track, and receives the image information generated by the camera module to read the image. A train reading device that reads whether a train is entering a track by comparing and analyzing it with previously stored train image data using technology, and generates and transmits a train entry signal with a specific frequency when read as a train entry; An administrator terminal that uploads train image data to the train reading device through an application; A repeater disposed on the track side and configured to receive a train entry signal with a specific frequency transmitted from the train reading device; and a communication alarm device for workers configured to generate vibration or an alarm when receiving a signal transmitted from the repeater, and to display information about the entry of a train on the LCD display. A radar sensor module and a camera module comprising a. We present an on-track train entry detection system using .

The train reading device includes a camera module that generates image information when an object enters the track; Photo information D/B in which train image data uploaded through a terminal application is stored in conjunction with the manager terminal; And an object analysis module that is linked to the photo information D/B and compares the image information received from the camera module with the train image information stored in the photo information D/B to generate and transmit train reading and train entry signals. It consists of:

In addition, in order to achieve the above object, in another embodiment of the present invention, in a train entry detection system using a camera module, when an object enters the track, a built-in sensor reads whether the train has entered the track, and the train enters the track. A train detection device configured to expose a customized image object to an external output monitor according to the entry reading result; An image control configured to photograph image objects exposed to the external output monitor of the train detection device, extract characters from the photographed image objects, and generate and transmit train entry signals with specific programmed frequencies according to the extracted characters. Device; A repeater disposed on the track side and configured to receive a train entry signal with a specific frequency transmitted from the artificial intelligence object analysis module; and a communication alarm device for workers configured to generate vibration or an alarm when receiving a signal transmitted from the repeater, and to display information about the entry of a train on the LCD display. A radar sensor module and a camera module comprising a. We present an on-track train entry detection system using .

The train detection device includes a smart streetlight configured to adjust the brightness of the LED according to the speed of the object detected within the detection distance of the radar sensor module; An illuminance sensor unit that measures and outputs the illuminance of the LED emitted from the smart streetlight; and a voltage detection module that switches according to a preset value based on the voltage output from the illuminance sensor unit and selects a signal to be output to the outside. It includes an external output monitor configured to selectively expose one of the customized image objects regarding train entry according to the signal output from the voltage detection module,

The image object exposed to the external output monitor consists of an image including a text area, and the text area is configured to include characters such as fast train entry speed, slow train entry speed, and no train entry situation. We present a train entry detection system using a sensor module.

The image control device includes a camera module that photographs an image object exposed to an external output monitor of the train detection device; And an algorithm trained to recognize a text area in an image object captured by the camera module, extract a text from the recognized text area, and generate and transmit a train entry signal with a specific frequency programmed according to the extracted text is stored. It consists of an object analysis module.

The size of the vibration or alarm generated by the communication alarm device varies depending on the train entry signal generated through the character recognition of the object analysis module, and the contents of the train entry displayed on the LCD display are changed according to the situation. It is structured as possible.

In addition, in order to achieve the above object, the present invention relates to a method for detecting train entry on a track using a camera module,

When an object enters the track through a train detection device, reading whether a train is entering based on the speed of the object, and exposing a customized image object to an external output monitor according to the train entry reading result; Photographing an image object exposed to an external output monitor of the train detection device through a camera module of the video control device; Through a character recognition-based artificial intelligence algorithm stored in the internal object analysis module of the video control device, character areas are recognized in image objects captured by the camera module, characters are extracted from the recognized character areas, and each program is programmed according to the extracted characters. It includes generating and transmitting a train entry signal with a specific frequency,

The character recognition method of the object analysis module using the character recognition-based artificial intelligence algorithm includes obtaining an original image object by photographing an image object exposed to an external output monitor of the train detection device through the camera module; removing noise from the obtained original image object, applying perspective transformation, and converting it to gray scale; Binarizing the converted image object using an adaptive binarization algorithm; Detecting all edges of the image object from the binarized image object using an edge detection algorithm and detecting a rectangular area with a thick border line predicted to be a character area; Extracting a Tesseract OCR result using the image within the detected rectangular area as input data for the Tesseract OCR engine; and recognizing the result of the Tesseract OCR as text. We present a method for detecting train entry on a track using a radar sensor module and a camera module.

The text in the text area of the image object is relatively strong against surrounding noise, uses black text on a white background to have a clear color contrast, and uses a thick rectangular border for the text area to make it easy to extract only the text area within a complex image object. It is composed.

The following effects can be achieved by the train entry detection system and detection method using the radar sensor module and camera module of the present invention.

First, the present invention receives a signal transmitted from a train reading device or video control device through a communication alarm device owned by the worker, thereby reducing the possibility of a collision accident with a railroad vehicle (motor car, etc.) of the worker on the track and evacuation time. It has the effect of preventing safety accidents, etc. from occurring in advance by securing and lowering the possibility of railroad car accidents caused by work tools placed on the tracks.

Second, the present invention has the effect of accurately determining the entry of a train (motor car, etc.) into the track through a train reading device or a train detection device to prevent work interference or noise generation due to unnecessary warning sounds, etc. in the work space. There is.

Figure 1 is a diagram depicting the working environment of a train line on which the system according to the present invention is implemented.
2 to 3 are diagrams showing the configuration of a worker safety support system when entering a train according to the first embodiment of the present invention.
Figure 4 is an operational configuration diagram of the train entry monitoring system according to the first embodiment of the present invention.
Figures 5 and 6 are diagrams showing the configuration of a worker safety support system when entering a train according to a second embodiment of the present invention.
Figure 7 is an external configuration diagram of a train detection device according to a second embodiment of the present invention.
Figure 8 is an internal configuration diagram of a train detection device according to a second embodiment of the present invention.
Figure 9 is a state diagram of an external output monitor included in the train detection device according to the second embodiment of the present invention.

Hereinafter, the train entry detection system and detection method within the track using the radar sensor module and camera module of the present invention will be described in detail with reference to the attached drawings.

Figure 1 is a diagram depicting the working environment of a train line on which the system according to the present invention is implemented.

Referring to the figure, workers who work at night are placed around the train tracks to work.

Trains, such as trams and motor cars, continuously run on the tracks, and many workers are placed around the tracks and around the tracks as obstructions that provide work tools and work tools are placed on the tracks for work such as train maintenance and track repair. Work is performed on the track.

Therefore, when a train is in operation, a system will be needed to support the safety of workers on the track and prevent neglect of work tools.

2 to 3 are diagrams showing the configuration of a worker safety support system when entering a train according to the first embodiment of the present invention. Figure 4 is an operational configuration diagram of the train entry monitoring system according to the first embodiment of the present invention.

Referring to Figure 2, it shows an on-track train entry detection system according to the first embodiment of the present invention.

In this embodiment, a train reading device 100 is configured to determine whether the train has entered while generating image information when the train enters.

More specifically, the train reading device 100 includes a camera module 110 that generates image information to determine whether the entering object is a train when an object enters the track.

Referring to FIG. 4, image information generated by the camera module 110 is stored in image information D/B 140.

In addition, the image information generated by the camera module 110 is received and compared and analyzed with previously stored train image data using image reading technology to read whether or not a train is entering the track, and when the train is read as entering, a train with a specific frequency is detected. An object analysis module 130 is configured to generate and transmit an entry signal.

In addition, the train reading device 100 is equipped with an administrator terminal 150 that uploads train image data through an application.

In other words, train image data is constantly updated and stored in the photo information D/B 120 inside the train reading device 100 through the application installed on the manager terminal 150.

That is, through the smartphone application, the manager terminal 150 and the photo information D/B are communicated and linked, and the train image data uploaded from the manager terminal 150 is stored in the photo information D/B ( 120).

To further explain the train reading device 100, the train reading device 100 includes a camera module 110 that generates image information when an object enters the track, and a terminal application in communication with the manager terminal 150. A photo information D/B (120) is formed in which the train image data uploaded through is stored.

When an object enters the track, the object analysis module 130 compares the image information received from the camera module and stored in the image information D/B 140 with the train image information stored in the photo information D/B 120 to identify the train. Entry status is read, and when it is determined that the train is entering, a train entry signal is generated and transmitted to the outside through the antenna.

The object analysis module 130 photographs the image generated by the camera module 110 through an application installed inside and stored in the image information D/B 140, and the train image data previously stored in the photo information D/B 120. Each reading technology is compared and analyzed, and based on whether or not it matches, it is determined whether the train is currently entering.

In addition, in this embodiment, a repeater 500 is placed on the train track side and configured to receive a train entry signal with a specific frequency transmitted from the train reading device 100, and receives a signal transmitted from the repeater 500. When this happens, vibration or an alarm is generated, and the LCD display unit includes a communication alarm device 400 for workers configured to display information about the entry of the train.

To elaborate, when the object analysis module 130 of the train reading device 100 generates a train entry signal, it is configured to transmit an RF signal to the repeater 500 through an antenna.

When a signal is received from the repeater 500, the signal is transmitted to the communication alarm device 400 worn or held by workers.

Referring to the figure, the wrist-type communication alarm device 410 worn on the worker's wrist is configured in the form of a wristband to be worn by workers, and receives the signal output from the repeater 500.

The wrist-type communication alarm device 410 is comprised of a reception control module and a power unit, and immediately generates vibration or an alarm upon receiving a signal output from the repeater 500, and the LCD display also indicates the entry of a railroad vehicle. The content is configured to be displayed.

In addition, it constitutes a vest-type communication alarm device 420 that is worn on workers' tops. The vest-type communication alarm device 420 is comprised of a reception control module, a charger, and an LED inside. Similarly, upon receiving a signal output from the repeater 500, vibration or an alarm is immediately generated, and the LED is also configured to emit light. .

Figure 3 shows a state in which the train entry signal generated when the train reading device 100 determines that a train is entering is transmitted to the communication warning device 400 through the repeater 500.

That is, the object analysis module 130 compares the image information received from the camera module and stored in the image information D/B 140 with the train image information stored in the photo information D/B 120 when an object enters the track. This reads whether the train is entering, and when it is determined that the train is entering, a train entry signal is generated and transmitted to the outside through the antenna.

The present invention receives a signal transmitted from the repeater 500 through the communication alarm device 400 owned by the worker, thereby reducing the possibility of a collision with the worker's railroad vehicle (motor car, etc.) on the track, securing evacuation time, and It has the effect of preventing safety accidents, etc. from occurring by lowering the possibility of railroad car accidents caused by obstacles placed on the table, that is, work tools.

5 to 6 are diagrams showing the configuration of a worker safety support system when entering a train according to a second embodiment of the present invention.

When using the train reading device 100 in the first embodiment described above, there may be the following difficulties.

When the speed of the approaching train is high, the resolution of the image information generated by the camera module 110 decreases, making it difficult to read the train, and when filming a moving train through the camera module 110, the camera module 110 and the train The gap between is bound to narrow.

In this case, even if the train reading device 100 determines that a train is actually entering, even if the train entry is recognized in the work space, there may be a problem in that the response time is inevitably insufficient.

Accordingly, in the second embodiment, a supplemented configuration is presented to address this issue.

Referring to FIG. 5, the configuration of the second embodiment of the on-track train entry detection system using a camera module is shown.

Looking at the system of the second embodiment, it consists of a train detection device 200, an image control device 300, a communication alarm device 400, and a repeater.

The train detection device 200 is configured to read whether a train has entered the track from a built-in sensor when an object enters the track, and to expose a customized image object to the external output monitor 240 according to the train entry reading result. .

A detailed description of this will be provided later with reference to FIGS. 7 to 9.

In addition, the image control device 300 of this embodiment photographs an image object exposed to the external output monitor 240 of the train detection device 200, extracts a character from the photographed image object, and operates according to the extracted character. Each is configured to generate and transmit a train entry signal with a specific programmed frequency.

In addition, as in the first embodiment, a repeater 500 is placed on the track side and configured to receive a train entry signal with a specific frequency transmitted from the artificial intelligence object analysis module, and vibrates when receiving a signal transmitted from the repeater. An alarm is generated, and a communication alarm device 400 for workers is configured to display information about the entry of the train on the LCD display.

Figure 7 is a structural diagram of a train detection device according to a second embodiment of the present invention. When explaining the configuration of the train detection device 200 with reference to the figure, the train detection device 200 includes a smart device configured to adjust the brightness of the LED according to the speed of an object detected within the detection distance of the radar sensor module 210. A street light 250 is configured.

The radar sensor module 210 functions to detect the surrounding brightness and the movement and speed of objects or objects (trains, people, motorcycles, bicycles, etc.) in real time. In order to adjust the radar detection distance of the radar sensor module 210, the radar detection distance can be adjusted by adjusting the radar frequency or the antenna size.

The smart streetlight 250 is configured to automatically adjust the brightness of the streetlight LED based on the movement and speed of objects detected by the radar sensor module 210.

In addition, the train detection device 200 includes an illuminance sensor module 220 that measures and outputs the illuminance of the LED emitted from the smart streetlight.

Figure 8 is an internal configuration diagram of a train detection device according to a second embodiment of the present invention.

Referring to the figure, the illuminance sensor module 220 uses a CDS sensor, and the CDS sensor actually functions as a resistance. The value of the resistance varies variably depending on the amount of light, and the illuminance is determined by the resulting voltage difference. is distinguished.

In other words, when there is a lot of light, the resistance decreases, and when there is little light, the resistance increases. A sensor called cadmium sulfide (CDS) acts as a resistor, and the size of the output voltage changes.

The illuminance sensor module 220 uses a pull-up circuit to measure illuminance. When a pull-up resistor (R1) is used, voltage distribution occurs according to the resistance value (R2) of the illuminance sensor, and the brighter the surroundings, the greater the measured voltage. It gets smaller.

Accordingly, in this embodiment, the voltage detection module 230 is configured to select a signal to be output to the outside while switching according to a preset value based on the voltage output from the illuminance sensor module 200.

The voltage detection module 230 detects the output voltage of the illuminance sensor module 220, which is varied and output according to the brightness of the LED of the smart street light 250.

Referring to Figure 8, the size of the detected output voltage is configured to be divided into certain areas (A, B, C). For example, area "A" is an area with a relatively large output voltage, which is a smart streetlight ( 250) means that the brightness of the LED is small, and this level of output voltage can be defined as a level that cannot be recognized as the speed of train entry.

In addition, the "C" area is an area with a relatively low output voltage, meaning that the brightness of the LED of the smart street light 250 is high, and this level of output voltage is defined as a fast entry speed of the train.

Also, similarly, the “B” area can be seen as having a relatively intermediate output voltage between the “A” and “C” areas.

Accordingly, the voltage detection module 230 switches according to a preset value based on the voltage output from the illuminance sensor module 200 and selects a signal to be output to the outside.

To elaborate, when the magnitude of the voltage output from the illuminance sensor module 200 is included in the “A” area, a signal is set to be output through pin 1.

Similarly, when the magnitude of the voltage output from the illuminance sensor module 200 is included in the “B” area, a signal is set to be output through pin 2.

In addition, in this second embodiment, an external output monitor 240 is configured to selectively expose one of the customized image objects related to train entry according to the signal output from the voltage detection module 230.

Figure 9 is a state diagram of an external output monitor included in the train detection device according to the second embodiment of the present invention.

Referring to FIGS. 8 and 9, when an output signal is output from pin 1 of the voltage detection module 230, an image object written “not train entry situation” is displayed on the external output monitor 240.

In other words, judging by the magnitude of the voltage output from the illuminance sensor module 200, it is determined that “this is not a train entry situation” based on the magnitude of the voltage output, and the external output monitor 240 displays “ An image object that says “Not a train entry situation” is exposed and configured to be viewed from the outside.

As another example, when an output signal is output from pin 3 of the voltage detection module 230, an image object written on the external output monitor 240 that reads “fast train entry speed” is configured to be exposed for viewing from the outside.

In summary, the image object exposed for viewing from the outside in the external output monitor 240 consists of an image including a text area, and the text area contains characters such as high train entry speed, slow train entry speed, and no train entry situation. It is configured to include.

The configuration of the image control device 300 of the second embodiment of the present invention is as follows.

The image control device 300 includes a camera module 310 that photographs image objects exposed to the external output monitor 240 of the train detection device 200.

In addition, the image control device 300 includes an object analysis module 320. The object analysis module 320 recognizes text areas in image objects captured by the camera module 310, and extracts characters from the recognized text areas. An algorithm is configured to extract, generate and transmit a train entry signal with a specific programmed frequency according to the extracted characters.

The algorithm stored in the object analysis module 320 is a general character recognition algorithm or an artificial intelligence algorithm using a trained deep learning model.

Referring to FIG. 6, the image control device 300 recognizes a character from an image object exposed to the external output monitor 240, and sends a train entrance signal with a preset specific frequency according to the recognized character to the operator in the work space. It shows the status of being sent to the communication alarm device (400).

That is, the size of the vibration or alarm generated by the communication alarm device 400 varies depending on the train entry signal generated through character recognition of the object analysis module 320, and the LCD display unit of the communication alarm device 400 The information about the train's entry displayed is configured to change according to the situation.

In other words, when the characters recognized by the object analysis module 320 are “fast train entry speed,” “slow train entry speed,” and “not train entry situation,” train entry signals with different frequencies are generated. By configuring it to be sent to the communication warning device 400, it is differentiated to send a signal that workers can respond to appropriately for the train entry condition.

Therefore, the present invention accurately determines the entry state of a train (motor car, etc.) on the track and provides a warning signal tailored to the situation, thereby preventing work interference or noise generation due to unnecessary warning sounds in the work space.

In addition, in the second embodiment, unlike the first embodiment, the camera module 310 does not capture the moving image of a fast or moving train, but only captures a still image object exposed from the external output monitor 240. As you shoot, the clarity of the image quality increases.

In addition, the installation location of the video control device 300 can be placed at a considerable distance from the train detection device 200, so that when the train detection device 200 actually detects train entry, the video control device 300 detects the train. When generating and transmitting an entry signal, there is an advantage in that workers have more spare time in their workspace to respond.

Hereinafter, a method for detecting train entry into a track using a camera module according to a second embodiment of the present invention will be described.

First, when an object enters the track through the train detection device 200, it reads whether the train is entering based on the speed of the object, and a customized image object is exposed to the external output monitor 240 according to the train entry reading result. The ordering step (S100) is configured.

Next, a step (S200) of photographing an image object exposed to the external output monitor 240 of the train detection device through the camera module 310 of the image control device 300 is configured.

Next, the text recognition algorithm model stored in the internal object analysis module 320 of the image control device 300, specifically, a text recognition-based artificial intelligence algorithm, is used to recognize the text area in the image object captured by the camera module. A step (S300) consists of extracting characters from the text area and generating and transmitting a train entry signal with a specific programmed frequency according to the extracted characters.

The algorithm stored in the object analysis module 320 may be a general character recognition algorithm or an artificial intelligence algorithm using a trained deep learning model.

The character recognition method of the object analysis module 320 using the character recognition-based artificial intelligence algorithm is as follows.

A step (S310) of acquiring an original image object by photographing an image object exposed to the external output monitor 240 of the train detection device through the camera module 310 is comprised.

Next, a step (S320) is configured to remove noise from the obtained original image object, apply perspective transformation, and convert it to gray scale.

Next, a step 330 of binarizing the converted image object using an adaptive binarization algorithm is configured, and all edges of the image object are detected from the binarized image object using an edge detection algorithm, and the predicted character area is A step (S340) is performed to detect the rectangular area of the thick border line.

Next, a step (S350) is configured to extract the result of Tesseract OCR by using the image within the previously detected rectangular area as input data of the Tesseract OCR engine. Finally, a step (S360) of recognizing the result of the Tesseract OCR as text is performed.

In order to increase the reliability of character recognition, the present invention has the following additional configuration.

If character recognition in the image obtained in the previous S360 step fails, there are various reasons for the failure. Among them, the representative cause of failure is failure to detect text areas in the image due to insufficient preprocessing such as noise removal, the second cause is the use of characters of a significantly small size that cannot be recognized by Tesseract OCR, and the third cause is unusual This is the use of typefaces and fonts.

Accordingly, in order to increase the character recognition rate using the Tesseract OCR engine, the format of the image object exposed to the external output monitor 240 was configured as follows.

The image object exposed to the external output monitor 240 is relatively resistant to surrounding noise and uses black text on a white background to have a clear color contrast. Additionally, a thick rectangular border is created to make it easy to extract only the text area within a complex image.

Detecting only the part containing text in a complex image object requires a very large amount of computation time. Accordingly, the reason for using a thick line on the border of the external output monitor 240 of the present invention is that the area where letters exist in the image object of the external output monitor 240 of the train detection device can be easily detected through a simple algorithm, enabling character recognition. This is because the computational time required can be significantly shortened.

Ultimately, the present invention accurately determines the entry of a train (motor car, etc.) into the track, and at the same time takes into account the speed of the train entering the track and differentiates and transmits an alarm signal to the worker's communication alarm device 400, thereby creating unnecessary work space. A system configured to prevent work interference or noise generation due to warning sounds, etc. can be provided.

In the detailed description of the present invention, specific embodiments have been described above, but it will be obvious to those skilled in the art that various modifications are possible without departing from the scope of the present invention.

100: Train reading device
110: Camera module 120: Photo data D/B
130: Object analysis module 140: Transmission module
150: Administrator terminal
200: Train detection device
210: Radar sensor module 220: Illuminance sensor module
230: Voltage detection module 240: External output monitor
250: Smart street light
300: Video control device
310: Camera module 320: Object analysis module
330: Transmission module
400: Communication alarm device 410: Wrist-type communication alarm device
420: Vest-type communication alarm device 500: Repeater

Claims (8)

In the train entry detection system on the track,
A smart streetlight configured to adjust the brightness of the LED according to the speed of an object detected within the detection range of the radar sensor module, an illuminance sensor unit that measures and outputs the illuminance of the LED emitted from the smart streetlight, and an illuminance sensor unit that measures and outputs the illuminance of the LED emitted from the smart streetlight. An external device configured to selectively expose one of a voltage detection module that switches according to a preset value based on voltage and selects a signal to be output externally, and a customized image object regarding train entry according to the signal output from the voltage detection module. A train detection device including an output monitor;
A camera module that captures an image object exposed to the external output monitor of the train detection device and recognizes a text area in the image object captured by the camera module, extracts a text from the recognized text area, and performs programming according to the extracted text. An image control device comprising an object analysis module storing an algorithm trained to generate and transmit a train entrance signal with a specific frequency;
A repeater disposed on the track side and configured to receive a train entry signal with a specific frequency transmitted from the object analysis module; and
A radar sensor module and a camera module comprising a communication alarm device for workers configured to generate vibration or an alarm when receiving a signal transmitted from the repeater, and to display information about the entry of a train on the LCD display. Train entry detection system used on track. According to clause 1,
The image object exposed to the external output monitor of the train detection device consists of an image including a text area, and the text area is configured to include characters such as fast train entry speed, slow train entry speed, and no train entry situation. A train entry detection system on the track using a radar sensor module and camera module. According to clause 2,
The size of the vibration or alarm generated by the communication alarm device varies depending on the train entry signal generated through the character recognition of the object analysis module, and the contents of the train entry displayed on the LCD display are changed according to the situation. A train entry detection system on the track using a radar sensor module and a camera module, characterized in that it is configured to do so. In the method of detecting train entry on the track,
When an object enters the track through the radar sensor module of the train detection device, the speed of the object detected within the detection distance is detected, and based on the detected speed, the train detection device reads whether the train has entered, and the train entry reading result exposing a customized image object to an external output monitor;
Photographing an image object exposed to an external output monitor of the train detection device through a camera module of the video control device;
Through a character recognition-based artificial intelligence algorithm stored in the internal object analysis module of the video control device, character areas are recognized in image objects captured by the camera module, characters are extracted from the recognized character areas, and each program is programmed according to the extracted characters. It includes generating and transmitting a train entry signal with a specific frequency,
The character recognition method of the object analysis module using the character recognition-based artificial intelligence algorithm is,
Obtaining an original image object by photographing an image object exposed to an external output monitor of the train detection device through the camera module;
removing noise from the obtained original image object, applying perspective transformation, and converting it to gray scale;
Binarizing the converted image object using an adaptive binarization algorithm;
Detecting all edges of the image object from the binarized image object using an edge detection algorithm and detecting a rectangular area with a thick border line predicted to be a character area;
Extracting a Tesseract OCR result using the image within the detected rectangular area as input data for the Tesseract OCR engine; and
A method for detecting train entry on a track using a radar sensor module and a camera module, comprising the step of recognizing the result of the Tesseract OCR as text. According to clause 4,
The text in the text area of the image object is relatively strong against surrounding noise, uses black text on a white background to have a clear color contrast, and uses a thick rectangular border for the text area to make it easy to extract only the text area within a complex image object. A method for detecting train entry on a track using a radar sensor module and a camera module. delete delete delete

Images