KR102405937B1 - Method for estimating building height in a user-participating recognition platform using aritificial intelligence and apparatus therefore - Google Patents

Abstract

A method and apparatus for estimating the height of a building using an aerial photograph and street image analysis based on artificial intelligence object recognition are disclosed. A building height estimation method according to an embodiment of the present invention is an artificial intelligence object recognition-based building height estimation method, comprising: analyzing an aerial photograph of a target area and detecting at least one target building included in the aerial photograph; receiving a street image corresponding to the location information based on the detected location information of the target building; analyzing the street image to detect the target building from the street image; and detecting a preset object in the target building, and classifying the relative height of each target building based on coordinate information and quantity information of the detected object.

Description

A method and device for estimating the height of a building in a user-participatory recognition platform using artificial intelligence

The present invention relates to a technology for estimating the height of a building using aerial photo analysis and distance image analysis, and more specifically, it is possible to estimate the relative height of a building by analyzing aerial photos and street images based on artificial intelligence object recognition. It relates to a method for estimating the height of a building and an apparatus therefor.

In general, a geographic information system (GIS) refers to a computer-based system that uses and manages geographic data to solve problems related to space. When GIS was first developed in the early 1960s, GIS was nothing more than a computer-based program for processing map data. However, today, GIS is not only positioned as an important academic research field by itself, but is being actively applied to various sectors of society as it is combined with the latest information technology.

The basic characteristic of GIS is that it can handle attribute information related to the location of objects in spatial, ie, geographic space. Maps are the most efficient way to represent the locations and properties of spatially distributed objects. Therefore, the most basic data in GIS construction is a digital map. Numerical maps can be said to be indexed by interpreting and numerically editing various topographical data obtained from survey maps, aerial photographs, and satellite images, unlike classical paper maps.

The numerical map construction process is completed through several processes as follows. First, the paper map is digitized or scanned to form a numerical map, and then various input errors are corrected. Then, after transforming into an actual coordinate system to meet the user's purpose through coordinate transformation, a topological structure is established for understanding the mutual location and correlation between spatial objects. Thereafter, attribute data related to each figure data is input to the numerical map that has undergone the establishment of the topological structure.

Numerical maps produced in this way enable faster and more accurate map search than paper maps, and are superior in information management and usability to support various planning and decision-making more effectively.

Embodiments of the present invention provide a building height estimation method and apparatus capable of estimating the relative height of a building by analyzing an aerial photograph and a street image based on artificial intelligence object recognition.

A building height estimation method according to an embodiment of the present invention is an artificial intelligence object recognition-based building height estimation method, comprising: analyzing an aerial photograph of a target area and detecting at least one target building included in the aerial photograph; receiving a street image corresponding to the location information based on the detected location information of the target building; analyzing the street image to detect the target building from the street image; and detecting a preset object in the target building, and classifying the relative height of each target building based on coordinate information and quantity information of the detected object.

The detecting of the target building from the street image may include: removing a portion other than the target building from the street image using a preset segmentation technique; and detecting each of the target buildings by obtaining center coordinate information and vertex coordinate information of each of the target buildings detected from the distance image.

The classifying of the relative heights of the target buildings may include classifying the relative heights of the target buildings by reflecting center coordinate information and vertex coordinate information of each of the target buildings.

Classifying the relative heights of the target buildings may include classifying the relative heights of the target buildings by reflecting distance information between the target buildings on the aerial photograph.

Building height estimation apparatus according to an embodiment of the present invention is an artificial intelligence object recognition-based building height estimation apparatus, by analyzing an aerial photograph of a target area to detect at least one target building included in the aerial photograph. analysis department; a distance image analyzer configured to receive a street image corresponding to the location information based on the detected location information of the target building, and analyze the street image to detect the target building from the street image; and an estimator for detecting a preset object in the target building and classifying the relative height of each of the target buildings based on coordinate information and quantity information of the detected object.

The distance image analyzer removes the remaining parts other than the target building from the distance image using a preset segmentation technique, and obtains center coordinate information and vertex coordinate information of each target building detected in the distance image. , each of the target buildings may be detected.

The estimator may classify the relative height of each of the target buildings by reflecting the center coordinate information and the vertex coordinate information of each of the target buildings.

The estimator may classify the relative height of each of the target buildings by reflecting distance information between the target buildings on the aerial photograph.

According to embodiments of the present invention, since the relative height of a building can be estimated by analyzing an aerial photograph and a street image based on artificial intelligence object recognition, the relative height of a target building without an expensive 3D Lidar measurement dataset can be easily estimated.

According to embodiments of the present invention, the height of a specific building may be calculated, and heights between buildings may be compared.

The present invention can be used for various services such as urban planning and real estate commercial area analysis, and the information identified through the present invention can be provided to government offices and specific private companies.

1 is a flowchart showing an operation of a method for estimating the height of a building based on artificial intelligence object recognition according to an embodiment of the present invention.
2 shows an exemplary view for explaining the method of the present invention.
3 shows the configuration of an apparatus for estimating the height of a building based on artificial intelligence object recognition according to an embodiment of the present invention.

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be embodied in various different forms, and only these embodiments allow the disclosure of the present invention to be complete, and common knowledge in the art to which the present invention pertains It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of describing the embodiments, and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, "comprises" and/or "comprising" refers to the presence of one or more other components, steps, operations and/or elements mentioned. or addition is not excluded.

Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used with the meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless specifically defined explicitly.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the accompanying drawings. The same reference numerals are used for the same components in the drawings, and repeated descriptions of the same components are omitted.

To obtain the height of a building from an aerial photograph, an expensive 3D lidar measurement dataset is required, which is expensive.

Embodiments of the present invention analyze the aerial photograph and street image based on artificial intelligence object recognition to estimate the height of a building, for example, the relative height, thereby estimating the relative height of the target building without an expensive 3D Lidar measurement dataset. The gist is to estimate the height easily.

At this time, the present invention detects a target building set by a user or manager through aerial photo analysis by receiving or downloading an aerial photo of the target survey area, and receives a street image corresponding to the location information of the target building to analyze the street image After detecting the target building from the street image through Height can be estimated or classified.

Furthermore, the present invention can estimate the height of the target building detected from the street image by reflecting the distance information between the target buildings on the aerial photograph. For example, according to the present invention, height information or relative height information of each target building can be estimated using the above-described coordinate information and object information as well as distance information in an aerial photograph and distance information in a distance image.

According to the present invention, processes such as aerial photo analysis, distance image analysis, and object recognition can be performed based on artificial intelligence object recognition. That is, the present invention can generate each learning model or an integrated learning model that detects a target building or detects an object such as a window in each target building by training artificial intelligence using a preset training dataset, The learning model generated in this way may be updated using the results derived through aerial photograph analysis and distance image analysis, or may be updated using additionally collected training datasets. Artificial intelligence may be implemented by a neural network such as a convolutional neural network (CNN), a deep learning neural network (DNN), and the like, and the detailed description thereof is obvious to those skilled in the art of the present invention. is omitted.

1 is a flowchart showing an operation flow for a method for estimating the height of a building based on artificial intelligence object recognition according to an embodiment of the present invention, and may be performed by an apparatus in which an algorithm for the method of the present invention is installed.

Referring to FIG. 1 , the method for estimating the height of a building according to an embodiment of the present invention analyzes an aerial photograph of an area to be investigated to detect a target building included in the aerial photograph (S110).

Here, in step S110, by receiving the latitude and longitude of the coordinates of the investigation target area from the user or the manager, the investigation target area is set, and an aerial photograph of the set investigation target area is received from the database and the aerial picture is an artificial intelligence object. Specific buildings included in the aerial photo, for example, at least one preset target building may be detected by analysis based on recognition.

Here, in step S110, the research target area may be set by dragging the survey target area from the map data displayed on the screen according to an input of a user or an administrator.

In the present invention, the aerial photograph of the survey target area may be an aerial photograph downloaded only the survey target area from a database of a specific server, or an aerial photo corresponding to the survey target area from among the aerial photos previously downloaded and stored in the device providing the method of the present invention It could also be an aerial photo from which the photo was retrieved.

The target building in step S110 may include a home building, a commercial building, a building of a certain size or more, and the target building may be determined by a business providing this technology or determined through selection of an item that can set the target building. can At this time, in step S110, information on the detected building is retrieved from the database in which information such as the purpose and use of the building according to the latitude and longitude of the aerial photograph and whether it is for home use or commercial use is stored, so that the building is a target in the present invention. It can be determined whether it is a target building to be used, and through this, a desired target building can be detected. Of course, the present invention may detect a target building of a certain size or larger without considering the use or purpose of the building depending on the situation.

Furthermore, step S110 detects preset target buildings by analyzing the aerial photo of the investigation target area, and the remaining parts other than the detected target building, for example, the remaining parts other than buildings including roads, parking lots, rivers, etc. in advance It is removed using a set segmentation technique, and a target building may be detected from an aerial photograph from which this part has been removed.

In this case, in step S110, the target building may be detected from the aerial photograph by recognizing the target building from the aerial photograph of the survey target area using the artificial intelligence of the pre-trained learning model.

When the target building is detected by step S110, based on the detected location information or coordinate information of the target building, a street image corresponding to the location information is searched for and downloaded from a specific server, or a street image corresponding to the location information from the street image stored in the device is downloaded. By searching for a distance image, a distance image corresponding to the location information is received (S120).

In this case, in step S120, location information of the target building, that is, a distance image of a certain range based on latitude and longitude may be received.

When the street image including the target building is received by step S120, the received street image is analyzed to detect the target building from the street image (S130).

For example, in step S130 , as shown in FIG. 2 , target buildings indicated by a red box are detected within a region of interest (ROI) indicated by a yellow dotted line in the distance image.

Here, in step S130, in order to detect a target building in the street image, unnecessary parts such as electric wires and trees included in the street image are removed using a preset segmentation technique, and each of the target buildings in the street image from which the unnecessary parts are removed. By acquiring center coordinate information and vertex coordinate information, each target building may be detected.

As an example, step S120 may detect a target building to include a target building having the largest center coordinate and coordinate information of an upper vertex and a target building having the smallest coordinate information of a lower vertex in the distance image. Here, it can be determined that the building having the relatively highest upper vertex coordinates is the tallest building in the street image, and the building having the relatively smallest lower vertex coordinates of the building can be determined as the closest building in the street image.

Here, in step S130, the target building may be detected from the street image by using the artificial intelligence of the pre-trained learning model.

When a target building is detected from the street image by step S130, a preset object, for example, a window is detected in the detected target building, and the height of the target building using coordinate information and quantity information of the window detected in each target building to estimate (S140).

For example, as shown in FIG. 2 , a window indicated by a blue box may be detected in each target building, and a relative height between target buildings may be estimated using coordinate information of each detected window and the number of windows. Here, the number of windows may be used to estimate the number of floors of the target building.

Furthermore, in step S140, the relative height of each target building may be classified or estimated by reflecting the center coordinate information and vertex coordinate information of each target building detected in step S130.

Furthermore, step S140 may additionally reflect distance information in the aerial photograph. That is, since the target building detected in the aerial photograph can acquire distance information between the target buildings using the scale of the aerial photograph, the accuracy of the estimated height of the target building may be improved by reflecting the distance information between the target buildings. have. For example, in step S140, as shown in FIG. 2 , the distance from the street image to the left target building, the distance to the right target building, center coordinate information and vertex coordinate information of the left target building, and the center coordinate of the right target building The distance information between the target building on the left and the target building on the right can be estimated through the information and the vertex coordinate information. You can estimate the relative height of Similarly, in the case of the intermediate target building located behind the left target building, by reflecting the above-described distance information, the approximate height of the target building can be estimated, and the relative height with other target buildings can be accurately estimated or classified.

And, in step S140, distance information between target buildings in the street image can be calculated using the surrounding landscape, for example, road information, and distance information between target buildings in the street image is distance information of aerial photography through road information of the street image. It can be calculated by reflecting

Step S140 may also detect an object in the target building using the artificial intelligence of the pre-trained learning model.

When the height of the target building is estimated or the relative height between the target buildings is estimated through the above-described process, location information of the target building, for example, latitude and longitude, may be provided along with classified height information between the target buildings.

Here, the present invention may provide information including the estimated height or relative height of the target building in the form of a specific file, and the provided information may include latitude and longitude of the target building, estimated height information, estimated relative height information, etc. can

As described above, in the method according to an embodiment of the present invention, it is possible to estimate the relative height of a building by analyzing an aerial photograph and a street image based on artificial intelligence object recognition. The relative height of a building can be easily estimated.

In addition, the method according to an embodiment of the present invention may calculate the height of a specific building or may compare the heights between buildings.

The present invention can be used for various services such as urban planning and real estate commercial area analysis, and the information identified through the present invention can be provided to government offices and specific private companies.

3 shows a configuration of an apparatus for estimating a building height based on artificial intelligence object recognition according to an embodiment of the present invention, and shows a conceptual configuration of an apparatus for performing the method of FIGS. 1 to 2 .

Referring to FIG. 3 , an apparatus 300 according to an embodiment of the present invention includes a receiving unit 310 , an aerial photograph analyzing unit 320 , a distance image analyzing unit 330 , an estimating unit 340 , and a providing unit 350 . ) and a database (DB) 360 .

The DB 360 is a means for storing data for carrying out the present invention, and various data such as aerial photos, map data, street images, artificial intelligence object recognition-based algorithms, learning models, and estimation height data providing algorithms of target buildings In addition to these data, all kinds of data required for carrying out the present invention may be stored.

Here, data stored in the DB 360 may be updated through a wired/wireless network, and the data update period may be determined by a business operator or an individual providing the technology of the present invention.

The receiving unit 310 receives the latitude and longitude of the coordinates of the survey target area from the user or manager, sets the survey target area, and receives an aerial photograph of the survey target area.

Here, the receiver 310 may set the research target area by dragging the survey target area from the map data by a user or administrator's input.

The aerial photograph analysis unit 320 detects buildings included in the aerial photograph, for example, preset target buildings by analyzing the aerial photograph of the survey target area based on artificial intelligence object recognition.

At this time, the aerial photograph analysis unit 320 searches for information on the detected building from the database in which information such as the purpose and use of the building according to the latitude and longitude of the aerial photograph and whether it is for home use or commercial use is stored, and the building is In the present invention, it is possible to determine whether the building is a target, and through this, it is possible to detect a desired target building. In this case, the aerial photo analysis unit 320 may detect a target building included in the selected predetermined area when a target building or a predetermined area including the target building is selected through the user's input.

Depending on the situation, the aerial photo analysis unit 320 removes the parts other than the detected target building, for example, the non-building parts including roads, parking lots, rivers, etc. using a preset segmentation technique, and , it is also possible to detect the target buildings from the aerial photograph from which this part has been removed.

The distance image analysis unit 330 searches for and downloads a street image corresponding to the location information from a specific server on the basis of the location information or coordinate information of the target building detected by the aerial photo analysis unit 320, or the distance stored in the device. By searching for a street image corresponding to the location information in the image, a street image corresponding to the location information is received, and the received street image is analyzed to detect a target building from the street image.

At this time, the distance image analysis unit 330 removes unnecessary parts such as electric wires and trees included in the distance image using a preset segmentation technique, and the center coordinate information of each target building from the distance image from which the unnecessary parts are removed. By acquiring vertex coordinate information, each target building may be detected.

The estimator 340 detects a preset object, for example, a window in the detected target building, and estimates the height of the target building using coordinate information and quantity information of the windows detected in each target building.

In this case, the estimator 340 may classify or estimate the relative height of each target building by reflecting the detected center coordinate information and vertex coordinate information of each target building.

In this case, the estimator 340 may estimate the height or the relative height of the target building by additionally reflecting distance information in the aerial photograph. For example, since the estimator 340 can obtain distance information between target buildings using the scale of an aerial photograph, by reflecting the distance information between the target buildings, the estimation unit 340 estimates the height of the target building detected in the street image, and This can improve the accuracy of the estimated height of the target building.

When the height of the target building is estimated or the relative height between the target buildings is estimated, the providing unit 350 provides location information of the target building, for example, latitude and longitude, along with classified height information between the target buildings.

Although the description is omitted in the apparatus of FIG. 3, each component constituting the FIG. 3 may include all the contents described in the method of FIGS. 1 to 2, which is obvious to those skilled in the art. do.

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It may be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

Software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or apparatus, to be interpreted by or to provide instructions or data to the processing device. may be embodied in The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible from the above description by those skilled in the art. For example, the described techniques are performed in an order different from the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (8)

In the method for estimating the height of a building based on artificial intelligence object recognition,
analyzing an aerial photograph of a target area to detect at least one target building included in the aerial photograph;
receiving a street image corresponding to the location information based on the detected location information of the target building;
analyzing the street image to detect the target building from the street image; and
Detecting a preset object in the target building, and classifying the relative height of each target building based on coordinate information and quantity information of the detected object,
The step of detecting the target building in the street image is
removing a portion other than the target building from the street image using a preset segmentation technique; and
Detecting each of the target buildings by obtaining center coordinate information and vertex coordinate information of each of the target buildings detected in the distance image,
The step of classifying the relative height of each of the target buildings is
When the target building is detected in the distance image, the height of the target building is estimated using coordinate information and quantity information of a window that is a preset object in the detected target building, and distance information between the target buildings in the aerial photograph A method for estimating the height of a building, characterized in that the relative height of each of the target buildings is classified by reflecting the .
delete According to claim 1,
The step of classifying the relative height of each of the target buildings is
Building height estimation method, characterized in that for classifying the relative height of each of the target buildings by reflecting the center coordinate information and the vertex coordinate information of each of the target buildings.
delete In the apparatus for estimating the height of a building based on artificial intelligence object recognition,
An aerial photograph analysis unit for analyzing the aerial photograph of the target area to detect at least one or more target buildings included in the aerial photograph;
a distance image analyzer configured to receive a street image corresponding to the location information based on the detected location information of the target building, and analyze the street image to detect the target building from the street image; and
Including an estimator for detecting a preset object in the target building, and classifying the relative height of each target building based on the coordinate information and quantity information of the detected object,
The distance image analysis unit
By removing the remaining parts other than the target building from the street image using a preset segmentation technique, and obtaining center coordinate information and vertex coordinate information of each target building detected in the street image, each of the target buildings to detect,
The estimator
When the target building is detected in the distance image, the height of the target building is estimated using coordinate information and quantity information of a window that is a preset object in the detected target building, and distance information between the target buildings in the aerial photograph Building height estimation device, characterized in that for classifying the relative height of each of the target buildings by reflecting.
delete 6. The method of claim 5,
The estimator
Building height estimation apparatus, characterized in that for classifying the relative height of each of the target buildings by reflecting the center coordinate information and the vertex coordinate information of each of the target buildings.
delete

Images