KR102586175B1 - Apparatus for visualizing the result of detecting pothole in aerial image based on deep-learning and method thereof - Google Patents

Abstract

The present invention relates to a device and method for visualizing pothole detection results in deep learning-based aerial images.
According to the present invention, in a method for visualizing porthole detection results using a pothole detection result visualization device, the steps include receiving an aerial image captured by a drone, inputting the aerial image into a previously learned detection module, and detecting a porthole from the aerial image. Detecting, analyzing the aerial image in which the pothole was detected, measuring the size and coordinates of the pothole in the aerial image, and mapping the coordinates of the pothole on a map to visualize the detection point of the pothole with a marker. A step of visualizing detection results including the size and coordinates of the corresponding pothole and an aerial image in which the pothole was captured when the marker is selected on the map.
According to the present invention, the location of potholes can be detected by deep learning analysis of aerial images, the detection results can be visualized and mapped on a map, and the status and management of potholes in each region can be easily tracked.

Description

Apparatus for visualizing the result of detecting pothole in aerial image based on deep-learning and method thereof}

The present invention relates to a device and method for visualizing the results of pothole detection in aerial images based on deep learning. More specifically, it can detect potholes in the image by deep learning analysis of aerial images and provide mapping of the detection results on an actual map. This relates to a device and method for visualizing pothole detection results in aerial images.

Pothole refers to a localized small hole or crack that appears on an asphalt or concrete road surface. There are various causes of potholes, but they are mainly caused by aging of the road surface, poor asphalt mixture and construction, poor drainage structure, use of calcium chloride for snow removal, and rainy weather.

Since potholes are a major threat to vehicles traveling on the road, so much so that they are called landmines on the road, technology that can analyze and detect potholes on the road is required.

In relation to this, a method for detecting portholes using vibration, 3D restoration, laser, etc. has been disclosed in the past, but there are problems in that it is high cost, takes a long time, and has low detection accuracy, making it less effective.

In addition, most of the existing technologies are for detecting potholes while driving a vehicle, and there is a lack of technology that can comprehensively manage potholes scattered in various regions.

Additionally, detection technology using drones has been actively introduced recently. Therefore, a new technology is required that can efficiently detect and manage potholes in each region using images from these drones.

The technology behind the present invention is disclosed in Korean Patent Publication No. 10-2018-0136601 (published on December 26, 2018).

The purpose of the present invention is to provide a device and method for visualizing pothole detection results in deep learning-based aerial images, which can detect potholes by deep learning analysis of aerial images and provide mapping of the detection results on a map.

The present invention relates to a method for visualizing porthole detection results using a pothole detection result visualization device, comprising the steps of receiving an aerial image captured by a drone, and inputting the aerial image into a previously learned detection module to identify a pothole from the aerial image. Detecting, analyzing the aerial image in which the pothole was detected, measuring the size and coordinates of the pothole in the aerial image, and mapping the coordinates of the pothole on a map to visualize the detection point of the pothole with a marker. It provides a method for visualizing pothole detection results, including the step of selecting the marker on the map and visualizing detection results including the size and coordinates of the corresponding pothole and an aerial image in which the pothole was captured.

In addition, the detection module can detect potholes in the road by separating and detecting pothole objects and road objects from the aerial image based on the R-CNN technique and performing an AND operation on the detection results.

In addition, the measuring step can calculate the area, horizontal and vertical diameter, and latitude and longitude coordinate values of the pothole area detected in the aerial image based on metadata including EXIF and XMP recorded in the aerial image. there is.

Additionally, the step of providing the visualization may be provided by inserting a compass icon guiding the true north direction at a designated location in the aerial image and boxing the porthole area in the aerial image.

In addition, the step of providing the visualization includes providing an enlarged image of the box area when the box is clicked, and at least one of the area, horizontal and vertical diameter, and latitude and longitude coordinates of the pothole area around the enlarged image according to a user selection option. It can be output in text format.

In addition, the step of providing the visualization includes, when selecting the first option, the horizontal and vertical diameter values of the porthole area in a state in which horizontal and vertical scales of a set scale size are overlaid along the outline of the porthole area in the enlarged image. Output, and when the second option is selected, the area and latitude and longitude coordinates of the pothole area can be output while the pothole area in the enlarged image is masked and displayed with a set color.

In addition, the present invention includes an input unit that receives aerial images captured by a drone, a detection unit that inputs the aerial images into a previously learned detection module to detect potholes from the aerial images, and an aerial image in which the potholes were detected. a measuring unit that analyzes the size and coordinates of the pothole in the aerial image, a display unit that maps the coordinates of the pothole on a map and visualizes the detection point of the pothole as a marker, and when selecting the marker on the map, Provided is a pothole detection result visualization device that includes a control unit that visualizes detection results including the size and coordinates of the pothole and an aerial image in which the pothole was captured.

According to the present invention, the location of potholes can be detected by deep learning analysis of aerial images, and the detection results can be visualized and mapped on a map.

In addition, in the case of the present invention, the detection location, shape, measured size and area, and captured images of potholes in each region detected from aerial images are visualized and provided along with the true north direction of the road, and management of potholes observed in each region is provided. and enables history tracking.

Figure 1 is a diagram showing the configuration of a system for visualizing pothole detection results in aerial images according to an embodiment of the present invention.
Figure 2 is a diagram showing the configuration of the device for visualizing the porthole detection results of Figure 1.
Figure 3 is a diagram illustrating a method for visualizing porthole detection results using the device of Figure 2.
Figure 4 is a diagram illustrating the principle of detecting potholes in a road in an aerial image based on deep learning analysis in an embodiment of the present invention.
Figure 5 is a diagram explaining the principle of calculating the area of a pothole by analyzing aerial images in an embodiment of the present invention.
Figure 6 is a diagram explaining the principle of measuring the coordinates of a pothole by analyzing aerial images in an embodiment of the present invention.
Figure 7 is a diagram illustrating the visualization of a pothole detection location on a map in an embodiment of the present invention.
Figure 8 is a diagram explaining a porthole detection result visualization page according to an embodiment of the present invention.
Figure 10 is a diagram illustrating a screen that provides pothole detection results when a marker in a map is clicked according to an embodiment of the present invention.
Figure 11 is a diagram explaining the porthole location and coordinate output function in an aerial image according to an embodiment of the present invention.
Figure 12 is a diagram showing visualization of the area, horizontal, and vertical size of the pothole in an embodiment of the present invention.

Then, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention in the drawings, parts that are not related to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected," but also the case where it is "electrically connected" with another element in between. . Additionally, when a part "includes" a certain component, this means that it may further include other components rather than excluding other components, unless specifically stated to the contrary.

Figure 1 is a diagram showing the configuration of a porthole detection result visualization system according to an embodiment of the present invention.

As shown in FIG. 1, the porthole detection result visualization system includes a porthole detection result visualization device 100 and a user terminal 200.

The porthole detection result visualization device 100 may be connected to at least one user terminal 200 through a wired network, a wireless network, or a combined wired and wireless network to communicate with each other and transmit and receive data.

Here, the pothole detection result visualization device 100 may correspond to a pothole management server (hereinafter referred to as management server) that analyzes, detects, and manages potholes from aerial images. The management server can detect potholes from the aerial image by deep learning analysis of the aerial image input through the input interface, output the detection result through the output interface, and provide it to the display or user terminal 200.

The porthole detection result visualization device 100 is connected to the user terminal 200 through a wireless, wired, or wired/wireless combined network and can provide porthole detection results, and a service platform implemented as an application or web ( platform) can be provided to the user terminal 200.

Here, the porthole detection result visualization device 100 may be a management server itself that services the porthole detection results, but may take the form of an application program, such as an application that runs on the user terminal 200 and provides related services. . In this case, the user terminal 200 can receive related services by connecting to a management server and a wired or wireless network while an application running in an app or web environment is running.

Applications can be operated or provided on servers. At this time, the application can perform the entire process, including the upload process of aerial images, the pothole detection process through image analysis, and the detection information visualization process. Additionally, the application can receive and process multiple aerial video images as input and provide a list of detection results for the multiple aerial images.

The user terminal 200 may include devices that can exchange information by accessing a management server through a wired or wireless network, such as a PC, desktop, tablet, laptop, smartphone, or smart pad. In the case of devices with built-in wireless functions (smartphones, laptops, pads, etc.), the functions of this system can be provided in the form of a mobile application on the device. The user terminal 200 may correspond to an administrator's terminal or a general user's terminal.

In an embodiment of the present invention, the porthole detection result visualization device 100 can provide porthole occurrence status, history tracking, and search functions in each region by executing a process for porthole detection result service and providing the result.

The device 100 for visualizing the porthole detection results may include a processor, memory, user interface input/output device, storage device, etc. At this time, the processor can process and analyze data, and control the operation and data flow of each component. Input/output devices are responsible for functions such as information or data input, inquiry, and output.

Figure 2 is a diagram showing the configuration of the device for visualizing the porthole detection results of Figure 1.

Referring to Figure 2, the porthole detection result visualization device 100 according to an embodiment of the present invention includes an input unit 110, a detection unit 120, a measurement unit 130, a display unit 140, and a control unit 150. do. Here, the operation of each unit 110, 120, and 130 and data flow control between each unit may be performed by the control unit 140.

The input unit 110 receives aerial images captured by a drone. Then, the input aerial image is transmitted to the detection unit 120.

The detection unit 120 detects potholes from the aerial image by inputting the aerial image into a previously learned detection module. The detection unit 120 detects a pothole area distinguished from a normal area from an aerial image based on deep learning, and provides the detection result to the measurement unit 130.

The measuring unit 130 analyzes the aerial image in which the pothole is detected and measures the actual size and coordinates of the pothole detected in the aerial image. At this time, the measurement unit 130 can measure the actual size and coordinates of the pothole area detected in the aerial image using metadata recorded in the aerial image.

At this time, the size of the pothole may include the porthole area and the horizontal and vertical sizes (diameter) of the pothole, and the coordinates of the pothole may include latitude and longitude coordinates. The measurement unit 130 may transmit the results of measuring the size and coordinates of the pothole to the display unit 140.

The display unit 140 maps the coordinates of the pothole on a map (e.g., Google Map, etc.), visualizes the detection point of the pothole through a marker, and displays it on the driving screen. According to this, a map screen with markers corresponding to the detection spots of each pothole can be provided for each region.

The control unit 150 can control the driving screen displayed by the display unit 140 according to a user manipulation signal and may have a built-in processor to perform the corresponding operation.

In addition, when a predetermined marker is selected (clicked) by the user on the map visualized on the screen through the display unit 140, the control unit 150 detects the pothole including the size, coordinates, and captured aerial image of the pothole corresponding to the marker. The results can be processed and visualized on the screen. At this time, the control unit 150 processes the detection result into a combination of at least one of an image, table, and text and provides it, enabling the user to quickly and intuitively check the porthole status and improving management efficiency.

Below, based on FIG. 3, a method for visualizing pothole detection results in deep learning-based aerial images according to an embodiment of the present invention will be described in more detail.

Figure 3 is a diagram illustrating a method of visualizing porthole detection results using the device of Figure 2.

First, the input unit 110 receives an aerial image captured by a drone (S310). The input unit 110 can load and receive an aerial image stored in storage from a network-connected user terminal 200 or a USB storage device connected to a communication port.

At this time, aerial video is video captured through a drone, and metadata including EXIF (Exchangeable Image File Format) and XMP (Extensible Metadata Platform from Adobe) are recorded in the captured image file. Accordingly, metadata includes information such as camera type and settings, shooting conditions, time, and location.

Using the metadata recorded in the aerial video image in this way, it is possible to calculate the actual area, horizontal/vertical size (diameter), latitude and longitude coordinate location, etc. of the pothole area detected in the aerial video image.

The input unit 110 transmits the input aerial image to the detection unit 120. Then, the detection unit 120 inputs the aerial image received from the input unit 110 into a previously learned detection module and detects the pothole from the aerial image (S320).

In an embodiment of the present invention, the detection unit 120 may be configured to include a detection module trained to detect potholes in the road.

The detection module can be pre-trained through multiple aerial video images. For example, a neural network is trained by deep learning analysis of the features of the pothole area labeled in multiple aerial video images, and the learned neural network is used to detect pothole areas that are distinct from normal areas in the aerial image. You can. Of course, the detection module can be continuously modified and updated like a typical neural network model.

In an embodiment of the present invention, the detection module can be trained to detect pothole areas in the road from aerial images based on the R-CNN technique.

Specifically, the detection module can be implemented to first separate and detect pothole objects and road objects from aerial images based on the Mask R-CNN technique, and perform an AND operation on the detection results to detect potholes in the road. The Mask R-CNN technique is a deep learning model that performs both object detection and object segmentation and can detect segmentation by masking multi-class objects.

Figure 4 is a diagram illustrating the principle of detecting potholes in a road in an aerial image based on deep learning analysis in an embodiment of the present invention.

As shown in FIG. 4, the detection unit 120 inputs the aerial image to a previously learned detection module (S410). Then, the detection module performs a deep learning analysis of the input aerial image to detect pothole objects and road objects in the aerial image (S420), and then performs an AND operation on the separately detected pothole detection results and the road detection results to detect pothole objects in the road. Only the part is detected (S430) and the detection result is output (S440). Here, the portion of the pothole area detected in the aerial image may be masked and output.

Accordingly, the detection unit 120 provides pothole detection results in the aerial image, and in particular, provides only the pothole detection results that exist on the road based on the above-described process. Afterwards, the detection unit 120 provides the results of pothole detection in the aerial image to the measurement unit 130.

Referring again to FIG. 3, the measuring unit 130 analyzes the aerial image in which the pothole is detected and measures the size and coordinates of the pothole in the aerial image (S330).

Here, the measurement unit 130 calculates the area, horizontal and vertical diameters, and latitude and longitude coordinate values of the pothole area detected in the aerial image based on metadata including EXIF and XMP recorded in the aerial image. The specific calculation method will be explained in detail through FIGS. 5 and 6 below.

Figure 5 is a diagram explaining the principle of calculating the area of a pothole by analyzing aerial images in an embodiment of the present invention.

As shown in FIG. 5, the measurement unit 130 receives an aerial image in which a pothole is detected (S510) and analyzes EXIF and XMP data, which are metadata recorded in the aerial image, (S520). The diagonal length (d), focal length (f), and altitude (c) of the image sensor are analyzed from metadata.

Table 1 below shows the factors used to calculate the pothole area, where d, f, and c are values obtained from metadata of aerial images.

sign explanation unit d Image sensor diagonal length mm f focal length mm c Altitude m C Diagonal of shooting area m A Angle of view/2 m a Complementary angle of A ˚

First, to calculate the actual area of the pothole area (porthole object) detected in the aerial image, the value A is calculated. Specifically, the value A is obtained by applying the diagonal length (d) and focal length (f) of the image sensor to Equation 1.

The value of A obtained in this way is 1/2 of the angle of view. In other words, A is the diagonal value of the image, and dividing it by 2 becomes the distance from the center of the image to the edge, so it is 1/2 of the angle of view, not the actual angle of view. Therefore, the A value is substantially proportional to the angle of view (proportionality value 0.5).

Afterwards, the complementary angle (a) and altitude (c) of A are applied to Equation 2 to obtain the diagonal length (C) of the actual shooting area of the aerial image.

And, the horizontal and vertical lengths of the actual shooting area of the aerial image are calculated through a proportional equation between the diagonal length (d) of the image sensor and the diagonal length (C) of the actual shooting area obtained by Equation 2.

Afterwards, the actual captured area (area per pixel) per pixel in the aerial image can be calculated using the proportional formula between the resolution of the aerial image obtained from the metadata and the horizontal and vertical sizes of the actual captured area (S530). Of course, in the process, you can also find out the horizontal and vertical lengths in meters per pixel.

The measurement unit 130 measures the area of the mask as a result of pothole detection using the unit area of the pixel (pixel unit area) in the aerial image (S540). Specifically, the actual area of the pothole detection area is calculated using the total number of pixels in the pothole area (porthole detection area) in the aerial image and the pixel unit area value obtained previously (S540). And the calculation result is output (S550).

Here, the measuring unit 130 may additionally obtain the porthole horizontal diameter and vertical diameter corresponding to the maximum horizontal size and maximum vertical size of the porthole, as well as the area of the porthole area.

Figure 6 is a diagram explaining the principle of measuring the coordinates of a pothole by analyzing aerial images in an embodiment of the present invention.

As shown in FIG. 6, the measurement unit 130 receives an aerial image in which a pothole is detected (S610) and analyzes EXIF data among the metadata of the aerial image (S620). At this time, EXIF data is analyzed to obtain the latitude (L) and longitude (K) coordinates of the aerial image.

Table 2 below shows the factors used to calculate the coordinates of potholes in seconds, of which L and K are values obtained from metadata of aerial images.

sign explanation unit x latitude radius km y equatorial radius km z polar radius km n Values based on equatorial standards
(North latitude = 1, South latitude = -1) - L Latitude do K Hardness do

Then, the x, y, z, and n values in Table 2 are applied to Equation 3 below to convert the degree-unit values of the latitude (L) and longitude (K) of the aerial image into second-unit values (S630).

Next, calculate the horizontal and vertical distances from the midpoint of the aerial image to the pothole area (S640). Using the pixel unit area obtained in the pothole area measurement process, the horizontal and vertical distances from the midpoint coordinates of the aerial image to the pothole can be calculated. For example, the horizontal and vertical distances to the reference point (center) within the pothole area can be calculated.

Then, the value in seconds is divided by the horizontal and vertical distances from the midpoint coordinate of the aerial image to the pothole, and the coordinate difference value (distance) in seconds from the midpoint coordinate to the pothole is calculated (S650). Then, the horizontal and vertical distances from the midpoint coordinates of the image to the porthole converted to seconds are added to calculate and output the final location of the pothole in seconds (S660).

In this way, the embodiment of the present invention calculates the exact coordinates of the pothole in the aerial image and provides convenience to practitioners through this. The coordinate value of the pothole area detected in the aerial image is transmitted to the display unit 140.

Referring again to FIG. 3, the display unit 140 maps the coordinates of the pothole on the map and visualizes the detection point of the pothole as a marker (S340).

Then, when a predetermined marker is selected by the user on the map, the control unit 150 visualizes and provides the detection result of the corresponding pothole on the screen (S350). At this time, the detection result includes the size and coordinates of the pothole and the captured aerial image.

Here, the control unit 150 can control the driving screen of the display unit 140, and when selecting a marker, the pothole area in the visualized aerial image can be boxed and provided, and can be provided in a true north direction at a designated location in the aerial image. You can insert a compass icon to guide you. In addition, the control unit 150 may output at least one of the area, horizontal and vertical diameters, and latitude and longitude coordinates of the corresponding pothole area in text form.

Figure 7 is a diagram illustrating the visualization of a pothole detection location on a map in an embodiment of the present invention, and Figure 8 is a diagram illustrating a visualization page of pothole detection results according to an embodiment of the present invention.

As shown in FIG. 7, the display unit 140 can visualize the pothole detection location as a marker on Google Maps and provide map enlargement and reduction functions. Through this, pothole detection spots in each region can be confirmed at a glance on a single map screen, and it is possible to conveniently view and manage potholes by region.

Additionally, referring to FIG. 8, the display unit 140 provides an aerial image list, an automatic pothole detection result output function, and a pothole automatic detection result metadata download function. At this time, the automatic pothole detection result output function includes a true north direction display function to determine the up and down direction of the road, a function to output the pothole location and coordinates in the aerial image, and a function to visualize the area, horizontal and vertical size of the pothole.

Figure 9 is a diagram illustrating an image upload and processing process for automatic porthole detection in an embodiment of the present invention. As shown in Figure 9, in the case of the present invention, a plurality of aerial video image files can be easily uploaded at once through an application, and the image upload process and the list of images being processed can be provided in real time.

Figure 10 is a diagram illustrating a screen that provides pothole detection results when a marker in a map is clicked according to an embodiment of the present invention.

As shown in Figure 10, in the embodiment of the present invention, the display unit 140 provides visualization of the pothole detection location in the current map through a marker, and when the user clicks on the marker, an aerial image of the location corresponding to the marker and the corresponding aerial image are displayed. A list of porthole detections and detection results can be provided. Here, the driving screen of the display unit 140 can be controlled by the control unit 150.

Figure 11 is a diagram explaining the porthole location and coordinate output function in an aerial image according to an embodiment of the present invention.

As shown in FIG. 11, the control unit 150 inserts a compass icon guiding the true north direction at a designated location in the aerial image (e.g., upper left corner of the image) on the screen that provides the porthole detection results for the selected marker point. can do. In addition, the detected pothole area in the aerial image can be boxed in a specific color for guidance.

In addition, when clicking on the corresponding box, the control unit 150 can provide an enlarged image of the box area through a separate window, and according to the user-selected option, the area, horizontal and vertical diameters of the porthole area around the enlarged image, At least one of the latitude and longitude coordinates can be provided in text form.

Figure 12 is a diagram showing visualization of the area, horizontal, and vertical size of the pothole in an embodiment of the present invention.

As shown in Figure 12, when the first option (porthole horizontal and vertical output button) is selected, the control unit 150 overlays horizontal and vertical scales of the set scale size along the outline of the porthole area in the enlarged image and displays the porthole. The horizontal and vertical diameter values of the area can be output.

In addition, when the second option (porthole area and coordinate output button) is selected, the control unit 150 displays the pothole area in the enlarged image by masking it with a set color, and displays the area of the pothole area and the latitude/longitude coordinate values. can be printed and provided. Additionally, in the case of the present invention, a button to download metadata for potholes detected in aerial images can be provided.

As described above, according to the present invention, the location of potholes can be accurately detected by deep learning analysis of aerial images, and the detection results can be visualized and mapped on a map.

In the case of this invention, the detection location, shape, measured size and area, and captured images of potholes in each region detected from aerial images are visualized and provided along with the true north direction of the road, thereby managing potholes observed in each region. and history tracking and can increase user convenience.

The present invention has been described with reference to the embodiments shown in the drawings, but these are merely illustrative, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true scope of technical protection of the present invention should be determined by the technical spirit of the appended claims.

100: Porthole detection result visualization device 110: Input unit
120: detection unit 130: measurement unit
140: display unit 150: control unit
200: user terminal

Claims (12)

In a method of visualizing porthole detection results using a porthole detection result visualization device,
Receiving an aerial image captured by a drone;
Inputting the aerial image into a previously learned detection module to detect potholes from the aerial image;
Analyzing the aerial image in which the pothole is detected and measuring the size and coordinates of the pothole in the aerial image;
Mapping the coordinates of the pothole on a map and visualizing the detection point of the pothole with a marker; and
When selecting the marker on the map, it includes visualizing and providing detection results including the size and coordinates of the pothole and an aerial image in which the pothole was captured,
The step of providing the visualization is,
A compass icon guiding the true north direction is inserted at a designated location in the aerial image, and the pothole area in the aerial image is boxed and provided,
When the box is clicked, an enlarged image of the box area is provided, and when the user selects the first option according to the user-selected option, horizontal and vertical scales of the set scale size are overlaid along the outline of the porthole area in the enlarged image. In one state, the horizontal and vertical diameter values of the pothole area are output in text form around the enlarged image,
When the user selects the second option, the porthole area in the enlarged image is masked and displayed with a set color, and the area and latitude and longitude coordinates of the porthole area are output in text form around the enlarged image. How to visualize detection results. In claim 1,
The detection module is,
A method for visualizing pothole detection results that detects potholes in the road by separating and detecting pothole objects and road objects from the aerial image based on the R-CNN technique and performing an AND operation on the detection results. In claim 1,
The measuring step is,
A pothole detection result visualization method that calculates the area, horizontal and vertical diameters, and latitude and longitude coordinate values of the detected pothole area in the aerial image based on metadata including EXIF and XMP recorded in the aerial image. delete delete delete An input unit that receives aerial images captured by a drone;
A detection unit that detects potholes from the aerial image by inputting the aerial image into a previously learned detection module;
A measuring unit that measures the size and coordinates of the pothole by analyzing the aerial image;
A display unit that maps the coordinates of the pothole on a map and visualizes the detection point of the pothole with a marker; and
When the marker is selected on the map, it includes a control unit that visualizes detection results including the size and coordinates of the pothole and an aerial image in which the pothole was captured,
The display unit,
A compass icon guiding the true north direction is inserted at a designated location in the aerial image, and the porthole area in the aerial image is boxed and provided,
The control unit,
When the box is clicked, an enlarged image of the box area is provided, and when the user selects the first option according to the user-selected option, horizontal and vertical scales of the set scale size are overlaid along the outline of the porthole area in the enlarged image. In one state, the horizontal and vertical diameter values of the pothole area are output in text form around the enlarged image,
When the user selects the second option, the porthole area in the enlarged image is masked and displayed with a set color, and the area and latitude and longitude coordinates of the porthole area are output in text form around the enlarged image. Detection result visualizer. In claim 7,
The detection module is,
A pothole detection result visualization device that separates and detects pothole objects and road objects from the aerial image based on the R-CNN technique and detects potholes in the road by ANDing the detection results. In claim 7,
The measuring unit,
A pothole detection result visualization device that calculates the area, horizontal and vertical diameters, and latitude and longitude coordinate values of the pothole area detected in the aerial image based on metadata including EXIF and XMP recorded in the aerial image. delete delete delete