KR20190080142A - A system for displaying accident situation event in tunnel to panorama image and a method of displaying the same - Google Patents

Abstract

An objective of the present invention is to provide a system for displaying an accident situation event occurring in a tunnel in a panorama image, which collects real-time image data of a previously installed CCTV and generates and displays panorama image data formed by converting the real-time image data into an air view type, thereby enabling an operator to easily monitor an accident situation in a tunnel. To this end, according to the present invention, the system for displaying an accident situation event occurring in a tunnel in a panorama image comprises: a plurality of image data collection units (100) collecting images in the tunnel; a panorama image conversion unit (200) converting image data collected by the image data collection unit (100) into the panorama image data to generate and display the panorama image data; and an image monitoring unit (300) monitoring an accident situation in the tunnel based on the panorama image data converted by the panorama image conversion unit (200).

Description

TECHNICAL FIELD [0001] The present invention relates to a system and a method for displaying an event in a tunnel in a panoramic image,

The present invention relates to a system and a display method for displaying a temporal event in a tunnel on a panoramic image, and more particularly, to a system and a method for displaying real time image data of a CCTV installed on a panoramic image, The present invention relates to a system and a display method for displaying a temporal event in a tunnel to be generated and displayed on a panoramic image.

Generally, in many mountainous areas, there are many tunnels due to geographical characteristics, and the length of tunnels for overcoming these geographical characteristics is gradually increasing. Despite various efforts to reduce accidents in tunnels, accidents in tunnels have continued to increase.

If we look at the detection status of conventional accident automatic detection systems installed in tunnels (eg, CCTV, etc.) by year, the detection rate has been steadily increasing since 2012. This is because the currently operating control systems and software are very sensitive to environmental changes in the tunnel and the quality of the sensed image, and the automatic recognition rate for each image item in most tunnels is lower as time goes by the aging of the equipment to be.

In recent years, the illumination characteristics of the tunnels have been reflected in the software in consideration of the characteristics of the tunnels, and the detection characteristics of shadows and the like have been taken into consideration for each tunnel shape to improve the detection performance within about 100 m Technology is being developed.

In reality, however, it is not easy to optimize all the CCTV equipment in a large number of tunnels, and the optimization conditions can also vary depending on the time history and the tunnel situation.

Therefore, there is a continuing need for a technology capable of receiving a more optimized image of a real-time event in a tunnel and managing it easily.

Accordingly, the present invention has been proposed in order to solve the above-mentioned problems, and it is an object of the present invention to provide a method and apparatus for capturing real-time image data of a CCTV installed in the past and generating and displaying panorama image data converted into a bird's- And more particularly, to a system and a display method for displaying a temporal event in a tunnel, which can be easily controlled, on a panoramic image.

In order to achieve the above object, a system for displaying a temporal event in a tunnel according to the present invention on a panoramic image includes a plurality of image data collection units 100 for collecting images in a tunnel, The panorama image conversion unit 200 converts the image data collected by the panorama image conversion unit 200 into panorama image data and generates and displays the panorama image data. And a video control unit (300).

In the system for displaying a temporal event in a tunnel according to the present invention on a panoramic image, the panoramic image conversion unit 200 includes a video data input unit (not shown) for receiving the video data collected from the video data collection unit 100 An image data analyzer 220 for analyzing an image of a region of interest of the image data input to the image data input unit 210 by using a Deep Learning Algorithm and generating deep- A deep running image data inferring unit 230 for analyzing and inferring the deep data image data generated by the image data analyzing unit 220, And a converted image data output unit 240 for combining the multiple images of the deep running image data and outputting the combined images as the panorama image data.

The system for displaying the temporal event in the tunnel according to the present invention on the panoramic image includes a deep learning learning unit 250 for repeatedly learning a plurality of deep learning learning image data generated by the image data analysis unit 220, The deep learning learning unit 250 generates a plurality of deep learning learning image data that have been repeatedly learned and stores the deep learning learning image data as big data, and the deep learning abruption image inferring unit 230 Provides deep learning image data for analysis and inference without error.

In addition, in the system for displaying a temporal event in a tunnel according to the present invention on a panoramic image, the image data analyzer 220 shapes image frames of the region of interest, spreads them to the same size, and performs distortion correction.

In a system for displaying a temporal event in a tunnel according to the present invention on a panoramic image, the transformed image data output unit 240 combines multiple images of the deep-running image data, Blurring an unclear region in an image into a bright region and replacing the blurred region with a sharp region, comparing the overlapping regions of the multiple images, and combining them into a panoramic image.

A method of displaying a temporal event in a tunnel according to the present invention on a panoramic image includes a first step S100 of receiving image data collected from the image data collection unit 100 by the image data input unit 210, A second step S200 of analyzing an image of a region of interest input to the image data input unit 210 by the image data analyzing unit 220 using a deep learning algorithm to generate converted deep training image data, A third step S300 of analyzing and inferring the deep learning image data generated by the image data analysis unit 220 by the deep learning sudden image inferring unit 230, (S400) of converting the multiple images of the deep-running image data determined by the image data output unit (230) by the converted image data output unit (240) and outputting the combined images as the panorama image data.

The method for displaying a temporal event in a tunnel according to the present invention on a panoramic image includes a deep learning learning unit 250 for repeatedly learning a plurality of deep learning learning image data generated by the image data analysis unit 220, The deep learning learning unit 250 generates a plurality of deep learning learning image data that have been repeatedly learned and stores the deep learning learning image data as big data, and the deep learning abruption image inferring unit 230 Provides deep learning image data for analysis and inference without error.

In addition, in the method of displaying a temporal event in a tunnel according to the present invention on a panoramic image, the image data analyzer 220 shapes image frames of the region of interest, spreads them to the same size, and performs distortion correction.

In addition, in the method of displaying a temporal event in a tunnel on a panoramic image, the converted image data output unit 240 may be configured to combine multiple images of the deep-running image data so that the frame of the multiple images is sharp And displays the panorama image in a tunnel in which the panorama image is combined by connecting the stereoscopic images by comparing and analyzing the overlapping areas of the multiple images.

According to the present invention, there is an effect of collecting real-time image data of CCTV installed in the past and generating and displaying the panoramic image data converted into bird's-eye view form, thereby easily controlling the situation in the tunnel.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an overall configuration of a system for displaying a temporal event in a tunnel on a panoramic image according to an embodiment of the present invention; FIG.
BACKGROUND OF THE INVENTION Field of the Invention [0001] The present invention relates to a system and method for displaying a high-resolution event in a tunnel on a panoramic image according to an embodiment of the present invention.
FIG. 3 is an exemplary view conceptually illustrating a detailed function execution procedure of a sudden sudden sudden image deduction unit of a system for displaying a temporal event in a tunnel on a panoramic image according to an exemplary embodiment of the present invention. FIG.
FIG. 4 is a photograph illustrating a tunnel image captured by a video data collector in a system for displaying a temporal event in a tunnel on a panoramic image according to an embodiment of the present invention. FIG.
FIG. 5 is a photograph showing a video image determined by a sudden sudden sudden depression of a tunnel in a system for displaying a temporal event in a tunnel according to an embodiment of the present invention. FIG.
6 is a photograph showing a panoramic image of a bird's-eye view type output by a transformed image data output unit in a system for displaying a temporal event in a tunnel according to an embodiment of the present invention on a panoramic image.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

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.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, the terms "comprises" or "having" and the like refer to the presence of stated features, integers, steps, operations, elements, components, or combinations thereof, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an overall configuration of a system for displaying a temporal event in a tunnel on a panoramic image according to an embodiment of the present invention.

1, a system for displaying a temporal event in a tunnel according to the present invention on a panoramic image includes a video data collection unit 100, a panoramic image conversion unit 200, and a video control unit 300 do.

The image data collection unit 100 includes a plurality of images, and collects images in the tunnel. The image data collecting unit 100 may be, for example, CCTV.

The panoramic image conversion unit 200 converts the image data collected by the image data collection unit 100 into panoramic image data, and generates and displays the panoramic image data.

The technique for converting the collected image data into panoramic image data will be described later.

The video control unit 300 controls the remote control of the tunnel with the panorama image data converted by the panorama image conversion unit 200. The video control unit 300 generates the video based on a deep learning algorithm to be described later without any extra effort to improve the quality of the image in the tunnel reflecting the poor environment (for example, dust, low illumination, low resolution, etc.) It is possible to more effectively control the unexpected situation of a moving object in a series of tunnels on the basis of the corrected image in the tunnel having high reliability (that is, extremely low threshold ratio).

The panorama image converting unit 200 includes the image data input unit 210, the image data analyzing unit 220, the deep running image disturbance estimating unit 230, and the image data output unit 240 .

The image data input unit 210 receives the image data collected from the image data collection unit 100.

The image data analyzer 220 analyzes the image of the ROI of the image data input to the image data input unit 210 using a deep learning algorithm to generate the deep running image data.

The deep-running unexpected image inference unit 230 analyzes and infer the deep-running image data generated by the image data analysis unit 220 to determine the deep-running image data.

The converted image data output unit 240 combines the multiple images of the deep-running image data determined by the deep-running sudden image inferring unit 230 and outputs the combined images as panorama image data.

Meanwhile, a system for displaying a temporal event in a tunnel according to the present invention on a panoramic image includes a deep learning learning unit 250 for repeatedly learning a plurality of deep learning learning image data generated by the image data analysis unit 220 .

The deep learning learning unit 250 generates a plurality of deep learning learning image data that have been repeatedly learned and stores the deep learning learning image data as big data. The deep learning learning unit 250 stores the deep learning learning image data as deep data, And provides learning learning image data.

FIG. 2 is a conceptual diagram illustrating a detailed configuration of a sudden sudden sudden depression of a system for displaying a temporal event in a tunnel according to an exemplary embodiment of the present invention. FIG. FIG. 8 is a diagram conceptually illustrating a detailed function execution procedure of a sudden sudden sudden depression of a system for displaying a temporal event in a tunnel according to an exemplary embodiment of the present invention; FIG.

2, the deep running unintentional video inference unit 230 receives the video data from the video data collection unit 100 (for example, a CCTV system) via an interface module 231, a wired / wireless on- An operation information storage module 232, a tunnel image big data configuration module 233, a moving object position / size / type learning module 234, a moving object unexpected state processing module 234, A moving object location / size / type learning extension support module 236, and the like.

In this case, the operation information storage module 232 may store various operation information necessary for the in-tunnel remote situation management procedure, for example, registration information of the image data collection unit 100, registration information of the video control unit 300, The program source information necessary for the process progress of the modules, the session information necessary for the communication connection of the respective computation modules, and the address information of the information storage space of each computation module in the information storage space of its own, And to assist the management process to proceed normally without any problem.

The tunnel image big data configuration module 233 may transmit the image data to the image data collection unit 100 through the interface module 231 and a wired / wireless on-line network (not shown) ), The video control unit 300, and the like, and proceeds to acquire the source video 201 in the tunnel.

When the source image 201 in the tunnel is acquired / completed, the tunnel image big data construction module 233 advances a series of information extraction routines to acquire a region of interest (Region) from the source image 201 in the acquired / of interest is extracted in the form of a still image 203 (e.g., 30 frames per second).

When the image region 202 of the region of interest is extracted / completed in the form of the still image 203 (for example, 30 frames per second) from the source image 201 in the tunnel, On the side of the big data configuration module 233, a series of information extraction / creation routines are executed on the still image 203, and the position information of the moving objects included in the still image 203, the size information of the moving objects, (Eg, a general moving vehicle, a person, a fire, a work vehicle, a fallen object, etc.) included in the still image 203 is coded.

Thus, the location information of the moving objects included in the still image 203, the size information of the moving objects, the type information of the moving objects (for example, a general moving vehicle, a person, a fire generation, a work vehicle, The tunnel image big data configuration module 233 proceeds to a series of information generation routines for these to generate location information of each moving object in the still image 203 corresponding to the area of interest 202, Size information, and type information of each moving object form a pair of structured large amounts of tunnel image big data (BDT).

The location information of each moving object in the still image 203 corresponding to the region of interest 202 on the source image 201 side in the tunnel, the size information of each moving object, and the type information group of each moving object, Size / type learning module 234 side communicates with the tunnel image big data construction module 233 and generates a series of images (large image data) The position information of each moving object in the still image 203 corresponding to the region of interest 202 on the source image 201 side in the tunnel, size information of each moving object, The type information of the moving objects, and the like.

The location information of each moving object in the still image 203 corresponding to the input information (for example, the interest area 202 on the source image 201 in the tunnel) (E.g., size information of each moving object, type information of each moving object, and the like) are repeatedly calculated with a weight (e.g., moving object location weight, moving object size weight, moving object type weight, etc.) allocated in the artificial neural network, (E.g., optimum mobile object location weight, optimal mobile object size weight, optimal mobile object type weight, etc.) with minimized error between the location / size / type information of the moving image and the moving object location / size / (E.g., optimal mobile object location weight, optimal mobile object size weight, optimal mobile object type weight < RTI ID = 0.0 > Etc.) is reflected in the &quot; moving object position / size / type learning data &quot;

At this time, a poor environment (e.g., dust, low illuminance, low resolution, etc.) inside the tunnel is intentionally reflected in the source image 201 in the tunnel. In this case, the moving object position / size / type learning module 234 The source image 201 in the tunnel is directly reflected in the learning procedure of the tunnel without the filtering procedure of the tunnel.

On the other hand, through the above-described procedure, when the <moving body position / size / type learning data> reflecting the combination of the optimum weight (for example, the combination of the optimum moving object position weight, the optimal moving object size weight, and the optimal moving object type weight) The moving object unexpected situation processing module 235 communicates with the video control unit 300 and the video data collecting unit 100 via the interface module 231 and a wired / wireless on-line network , A process of acquiring the source image in the tunnel in which the current situation of the current tunnel is reflected in real time from the image controller 300, the image data collecting unit 100, and the like.

In this way, when the source image in the tunnel in which the current situation of the current tunnel is reflected is obtained from the video control unit 300, the image data collecting unit 100, and the like, the moving object unexpected situation processing module 235, Size / type learning data] calculated by the moving object position / size / type learning module 234, and outputs the selected interest region image to the moving object position / size / (E.g., vehicle, person, drop, hindrance, etc.) and type (e.g., a moving object) of the moving object, recognizing the moving objects in the region of interest image (Eg, driving, stopping, stopping, reversing, fire, etc.).

If the type of the moving objects in the ROI image (e.g., vehicle, person, drop, obstacle, etc.) and type (for example, fixed running, stopped state, reverse run, fire, etc.) are inferred through the above procedure, The processing module 235 performs a series of image conversion routines on the basis of the types (for example, vehicles, people, falling objects, obstacles and the like) of the inferred moving objects and types (for example, fixed running, stopped state, reverse running, , A correction image in the tunnel reflecting the type of the moving objects (e.g., a vehicle, a person, a fallen object, a obstacle, etc.) and a type (for example, a fixed running, a stopped state, a reverse run, a fire, etc.) is performed.

In this way, when a corrected image in a tunnel reflecting the type of moving objects (e.g., a vehicle, a person, a falling object, a obstacle, etc.) and type (for example, a fixed running, a stopped state, The module 235 communicates with the video data collecting unit 100 via the interface module 231 and a wired / wireless on-line network (not shown) and transmits the video data to the video control unit 300, On the subject side, the source image in the low-quality tunnel is checked and the corrected image in the high-quality tunnel is checked, thereby supporting the more effective tunnel control procedure.

In addition, the moving object unexpected situation processing module 235 reads the corrected image in the tunnel, checks whether a series of unexpected situations have occurred in the moving object in the tunnel, and if it is determined that a series of unexpected situations have occurred in the moving object in the tunnel , Immediately proceeds to the information generation routine to generate an alarm message corresponding to the unexpected situation and then transmits the generated alarm message to the video control unit 300 so that the unexpected situation in the tunnel can be rapidly And will help them cope quickly.

As described above, in the present invention, after the source image in the tunnel is acquired by communicating with the video control unit 300, the image data collecting unit 100, and the like, a stop corresponding to the image region of the ROI A computer module capable of extracting an image and configuring a tunnel image big data structured by pairing the location information of the moving objects included in the extracted still image, the size information of the moving objects, and the type information of the moving objects, Based on the tunnel image big data, the Faster regional convergence neutral network algorithm is performed to learn position information of each moving object in the still image, size information of each moving object, type information of each moving object, Size / type learning data that reflects the combination of the optimal moving object size weight, the optimal moving object size weight, and the optimal moving object type weight The source image in the tunnel in which the current situation of the tunnel is reflected in real time is obtained by communicating with the computable module, the video control unit 300 and the image data collecting unit 100, And recognizes the position and size of the moving objects in the ROI by comparing the extracted ROI image with the moving object position / size / type learning data, The corrected image in the tunnel is reflected to the video controller 300 and the corrected image in the tunnel is checked to check whether the corrected image in the tunnel is reflected in the moving object in the tunnel, If it is determined that an unexpected situation has occurred in the moving object in the tunnel, an alarm message corresponding to the unexpected situation is transmitted to the video control unit 300, (Tunneling) of the tunnels in which a poor environment (for example, dust, low illumination, low resolution, etc.) inside the tunnel is reflected on the tunnel control subject side under the implementation environment of the present invention. Without any extra effort to improve the quality of the source image, it is only possible to use a highly reliable (ie, low threshold sensitivity) tunnel image based on the Faster regional convergence neutral network algorithm So that it is possible to more effectively carry out a procedure for coping with an unexpected situation of a moving object in a series of tunnels.

On the other hand, on the side of the moving object location / size / type learning extension support module 236 which operates in cooperation with the respective computation modules described above, in the phase in which the in-tunnel corrected image is generated by performing the function of the moving object incursion processing module 235, And communicates with the outbreak situation processing module 235, and proceeds to acquire the intra-tunnel corrected image generated thereby.

In this way, when the corrected image in the tunnel is acquired, the moving object location / size / type learning extension support module 236 advances a series of information generation routines so as to process the intra-tunnel corrected image (for example, A process of structuring the location information of the moving objects / the size information of the moving objects / the type information of the moving objects, and the like) is performed to transform the corrected image in the tunnel into the same format as the tunnel image big data (BDT) .

If the corrected image in the tunnel is transformed into the same format as the tunnel image big data BDT through the above procedure, the moving object location / size / type learning extension support module 236 can acquire the moving object location / size / type learning module 234 And proceeds to a process of transmitting / inputting the corrected intra-tunnel corrected image in the same format as the tunnel image big data (BDT) to the moving object position / size / type learning module 234 side while communicating with the moving object position / size / type learning module 234.

Of course, under the situation that the corrected intra-tunnel image is inputted / provided in the same format as the tunnel image big data (BDT), the moving object position / size / type learning module 234 side, As a new learning material, it becomes possible to carry out the extended learning process repeatedly.

In the system for displaying the temporal event in the tunnel according to the present invention on the panoramic image, the image data analyzer 220 shapes the image frame of the region of interest, spreads the image frame to the same size, and performs distortion correction.

In addition, when the multiple images of the deep-running image data are combined, the converted image data output unit 240 blends the unclear areas of the images of the multiple images into a clear area and replaces them, And combines them into a panoramic image by connecting the stereoscopic images.

As described above, the method of displaying the temporal event in the tunnel according to the present invention on the panoramic image includes a first step (S100) of receiving the image data collected from the image data collection unit 100 by the image data input unit 210, A second step S200 of analyzing the image of the region of interest inputted to the image data input unit 210 by the image data analysis unit 220 using a deep learning algorithm and generating the deep running image data, A third step S300 of analyzing and inferring the deep learning image data generated by the image data analysis unit 220 by the deep learning image data inferring unit 230, And a fourth step (S400) of outputting the combined image of the deep-running image data determined by the image data output unit 230 to the converted image data output unit 240 as panorama image data.

Hereinafter, a method for displaying a temporal event in a tunnel according to the present invention on a panoramic image includes a deep learning learning unit 250 for repeatedly learning a plurality of deep learning learning image data generated by the image data analysis unit 220 The deep learning learning unit 250 generates and stores a plurality of deep learning learning image data repeatedly learned as big data, and the deep learning sudden image inference unit 230 analyzes and analyzes the deep learning image data without error. And provides deep learning learning image data for reasoning.

In the method of displaying a temporal event in a tunnel according to the present invention on a panoramic image, the image data analysis unit 220 shapes image frames of a region of interest, spreads the image frames to the same size, and performs distortion correction.

On the other hand, in the method of displaying the temporal event in the tunnel on the panoramic image, the converted image data output unit 240 outputs the combined image of the deep-running image data, Are displayed on the panoramic image by blending and replacing the panoramic images with the panoramic images by comparing and analyzing the overlapping areas of the multiple images and combining the stereoscopic images into the panoramic images.

FIG. 4 is a photograph illustrating a tunnel image captured by the image data collector in a system for displaying a temporal event in a tunnel according to an exemplary embodiment of the present invention. FIG. In the system for displaying the temporal situation event in the tunnel on the panoramic image, this is a picture showing the image determined by the sudden sudden sudden imaging inference unit.

Referring to FIG. 4, in a system for displaying a temporal event in a tunnel according to the present invention in a panoramic image, the in-tunnel image collected by the image data collection unit 100 can be confirmed with a photograph. The image of the region of interest among the images collected in the tunnel is analyzed by a deep learning algorithm by the image data analysis unit 220 of the panoramic image conversion unit 200 and analyzed by the deep learning sudden image inference unit 230 And inferred to be a photograph as shown in Fig.

FIG. 6 is a photograph showing a bird's-eye view panorama image output by the converted image data output unit in a system for displaying a temporal event in a tunnel on a panoramic image according to an embodiment of the present invention.

Referring to FIG. 6, it can be confirmed that the multiple images of the deep-running image data determined by the deep-running unexpected image inference unit 230 are combined by the converted image data output unit 240 and output as panoramic image data .

The foregoing description is merely illustrative of the technical idea of the present invention, and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

10: wired / wireless online network
100: image data collection unit
200: panoramic image conversion unit
201: source video in tunnel
202: image area of interest area
203: still image
210: video data input unit
220: image data analysis unit
230: Deep Learning Emergency Image Reasoning
231: Interface module
232: Operational information storage module
233: Tunnel image big data configuration module
234: Moving object position / size / type learning module
235: Moving object breakdown processing module
236: Mobile location / size / type learning extension support module
240: The converted video data output section
250: Deep Learning Learning Section
300: Video control unit

Claims (9)

A plurality of image data collection units 100 for collecting images in the tunnel,
A panorama image converting unit 200 for converting the image data collected by the image data collecting unit 100 into panorama image data to generate and display the panorama image data,
Wherein the panorama image conversion unit (200) displays the panorama image in the tunnel including the video control unit (300) for controlling the panorama image data in the tunnel with the panorama image data converted by the panorama image conversion unit (200). The method according to claim 1,
The panorama image converting unit 200 converts the panorama image,
A video data input unit 210 receiving video data collected from the video data collecting unit 100,
An image data analyzing unit 220 for analyzing an image of a region of interest input from the image data input unit 210 by using a Deep Learning Algorithm to generate converted deep-running image data,
A deep running unexpected image inferring unit 230 for analyzing and inferring the deep running image data generated by the image data analyzing unit 220,
And a transformed image data output unit 240 that combines the multiple images of the deep running image data determined by the deep running image rejection unit 230 and outputs the combined image data as panorama image data is referred to as a panorama image . 3. The method of claim 2,
And a deep learning learning unit (250) for repeatedly learning a plurality of deep learning learning image data generated by the image data analysis unit (220)
The deep learning learning unit 250 generates a plurality of deep learning learning image data that have been repeatedly learned and stores the deep learning learning image data as big data. The deep learning learning image generating unit 250 stores the deep learning learning image data, And the deep-learning learning image data is displayed on the panoramic image. 3. The method of claim 2,
Wherein the image data analyzing unit 220 displays a temporal temporal event in a tunnel in which the image frame of the region of interest is shaped and spreads to the same size and performs distortion correction on the panoramic image. 3. The method of claim 2,
The converted image data output unit 240 blends the unclear region in the image of the frame of the multiple images into a clear region and replaces it when the multiple images of the deep running image data are combined,
A system for displaying a high-definition event in a tunnel in which panorama images are combined by connecting and comparing formatted images by comparing and analyzing overlapping areas of the multiple images. A first step S100 of receiving the image data collected from the image data collecting unit 100 by the image data inputting unit 210,
A second step S200 of analyzing the image of the region of interest inputted to the image data input unit 210 by the image data analysis unit 220 using the deep learning algorithm and generating the converted deep-running image data, ,
A third step (S300) in which the deep learning image data inferring unit 230 analyzes and infer the deep learning image data generated by the image data analyzing unit 220,
And a fourth step (S400) of combining the multiple images of the deep-running image data determined by the deep-running unexpected image inference unit 230 with the converted image data output unit 240 and outputting the combined images as panorama image data How to display my resignation event on a panoramic image. The method according to claim 6,
And a deep learning learning unit (250) for repeatedly learning a plurality of deep learning learning image data generated by the image data analysis unit (220)
The deep learning learning unit 250 generates a plurality of deep learning learning image data that have been repeatedly learned and stores the deep learning learning image data as big data. The deep learning learning image generating unit 250 stores the deep learning learning image data, And displaying the temporal event in the tunnel providing deep learning learning image data on the panoramic image. The method according to claim 6,
Wherein the image data analyzer (220) displays a temporal temporal event in a tunnel that shapes image frames of the region of interest, spreads them to the same size, and performs distortion correction on the panoramic image. The method according to claim 6,
The converted image data output unit 240 blends the unclear region in the image of the frame of the multiple images into a clear region and replaces it when the multiple images of the deep running image data are combined,
And comparing the overlapping areas of the multiple images with each other to combine the formatted images into a panorama image.

Images