KR102297217B1 - Method and apparatus for identifying object and object location equality between images - Google Patents

Abstract

A method and apparatus are disclosed for identifying the identity of an object and an object position between images. A method for identifying the identity of an object and an object position between images includes collecting a plurality of images based on crowdsourcing, separating the collected images into an object region and a background region, respectively, and separating the plurality of images into the object region and identifying the identity of an object by matching feature points extracted from the obtained images with each other, and identifying the identity of an object position by matching feature points extracted from images separated by the background area.

Description

METHOD AND APPARATUS FOR IDENTIFYING OBJECT AND OBJECT LOCATION EQUALITY BETWEEN IMAGES

The present invention relates to a method and apparatus for identifying the identity of an object and an object position between images, and more particularly, by determining the identity of the object as well as the identity of the object between images, the change state of the object of interest can be accurately determined It's about identifying skills.

With the development of image analysis technology, research is being actively conducted to analyze various images captured by a camera to observe changes in the surrounding environment including the object of interest. In particular, in order to build a safe and automated smart city, it is essential to analyze various CCTV images installed on roads or residential areas to quickly identify and respond to changes in the road environment or residential environment in the city.

As such, one of the prior art techniques for tracking changes in various objects included in images by analyzing images obtained from multiple cameras is to estimate the moving speed of an object of interest, and based on the estimated speed, it is possible to become an object. minimize candidates. Thereafter, the identity of the object is identified by analyzing the pose and feature points of the object with respect to the minimized candidates.

Another prior art, in order to recognize the presence or absence of the same object of interest regardless of the camera view, determines the optimal camera setting based on the relative directionality of the object, and the object through feature point extraction for the image taken with the determined camera setting Identify the identity of objects by classifying them.

However, the existing prior art concentrates on determining whether the object is identical, and does not consider whether the object is actually located where the absolute position of the object is identical. It's easy to figure out.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above problems, to provide a method for identifying the identity of an object and an object position between images.

Another object of the present invention to solve the above problems is to provide an apparatus for identifying the identity of an object and an object position between images.

Another object of the present invention to solve the above problems is to provide a road environment management system using crowdsourcing-based images.

The present invention for achieving the above object provides a method for identifying the identity of an object and an object position between images.

A method for identifying the identity of an object and an object position between images includes collecting a plurality of images based on crowdsourcing, separating the collected images into an object region and a background region, respectively, and separating the plurality of images into the object region The method may include identifying the identity of an object by matching feature points extracted from the obtained images with each other, and identifying the identity of an object location by matching feature points extracted from images separated by the background area with each other.

The separating into the object region and the background region may include classifying the plurality of images into one or more groups based on location information on the plurality of collected images, and dividing each of the plurality of images classified into the same group into an object region and Separation into a background region may be included.

The location information for the plurality of images may be GPS information obtained at the time of capturing an image by an imaging device that has captured each of the plurality of images.

The separating into the object region and the background region may include extracting an object from each of the plurality of images and generating a region separation mask for each of the plurality of images by using the extracted object.

After generating the region separation mask, the method may further include multiplying the generated region separation mask by the plurality of images in units of pixels.

The step of identifying the identity of the object may include calculating the number of matching feature points and identifying the identity of the object based on the calculated number of feature points.

The identifying of object identity based on the number of feature points may include determining that there is object identity when the number of feature points is greater than a preset threshold value.

The step of identifying the identity of the object location includes: classifying the plurality of images into one or more groups according to a result of identifying the identity of the object; identifying the identity of the object location with respect to each other.

Another aspect of the present invention for achieving the above object provides an apparatus for identifying the identity of an object and an object position between images.

An apparatus for identifying the identity of an object and an object position between images includes at least one processor and a memory for storing instructions instructing the at least one processor to perform at least one step. ) may be included.

The at least one step may include: collecting a plurality of images based on crowdsourcing; separating the plurality of collected images into an object region and a background region, respectively; matching feature points extracted from the images separated into the object region with each other and identifying the identity of the object and identifying the identity of the object position by matching feature points extracted from the images separated by the background area.

The separating into the object region and the background region may include classifying the plurality of images into one or more groups based on location information on the plurality of collected images, and dividing each of the plurality of images classified into the same group into an object region and Separation into a background region may be included.

The location information for the plurality of images may be GPS information obtained at the time of capturing an image by an imaging device that has captured each of the plurality of images.

The separating into the object region and the background region may include extracting an object from each of the plurality of images and generating a region separation mask for each of the plurality of images by using the extracted object.

After generating the region separation mask, the method may further include multiplying the generated region separation mask by the plurality of images in units of pixels.

The step of identifying the identity of the object may include calculating the number of matching feature points and identifying the identity of the object based on the calculated number of feature points.

The identifying of object identity based on the number of feature points may include determining that there is object identity when the number of feature points is greater than a preset threshold value.

The step of identifying the identity of the object location includes: classifying the plurality of images into one or more groups according to a result of identifying the identity of the object; and identifying the identity of the object location with respect to each other.

Another aspect of the present invention for achieving the above object provides a road environment management system using crowd-sourcing-based images.

A road environment management system using crowdsourcing-based images receives a plurality of images photographed from one or more user terminals and location information about the one or more user terminals, and identifies the plurality of images and the location information received from an image The plurality of images received from the road control server and the road control server are transmitted to the device to receive the analysis result, and the object included in the plurality of images and the position of the object are identical by using the learned data and the location information. and an image identification device that transmits the analysis result performed identification to the road control server.

The road control server may monitor a real-time road environment condition based on the analysis result.

The image identification apparatus may include at least one processor and a memory for storing instructions instructing the at least one processor to perform at least one step.

The at least one step may include: collecting the plurality of images; separating the plurality of collected images into an object region and a background region, respectively; The method may include identifying the identity and identifying the identity of the object position by matching feature points extracted from the images separated by the background region with each other.

The separating into the object region and the background region may include: classifying the plurality of images into one or more groups based on the location information; and separating each of the plurality of images classified into the same group into an object region and a background region may include.

The separating into the object region and the background region may include extracting an object from each of the plurality of images and generating a region separation mask for each of the plurality of images by using the extracted object.

When using the method and apparatus for identifying the identity of an object and an object position between images according to the present invention as described above, it is possible to accurately analyze whether the same object is in the same position as the object and the object position are simultaneously identified in the image. can

In addition, since the road environment can be managed based on crowdsourcing, it has the advantage of inducing citizen participation and preventing residents' inconvenience and various accidents.

1 is a conceptual diagram for explaining a method for identifying an object and an object position identical between images according to an embodiment of the present invention.
2 is an exemplary diagram for explaining a method for matching feature points between images according to an embodiment of the present invention.
3 is a representative flowchart of a method for identifying the identity of an object and an object position between images according to an embodiment of the present invention.
4 is a block diagram of an apparatus for identifying an object and a position of an object between images according to an embodiment of the present invention.
5 is a conceptual diagram for explaining a road environment management system using crowd-sourcing-based images to which a method and apparatus for identifying an object and an object position are identical between images according to an embodiment of the present invention.
6 is an exemplary diagram for explaining an effect of a method for identifying an object and a position of an object between images according to an embodiment of the present invention.

Since the present invention can have various changes and can have various embodiments, specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention. In describing each figure, like reference numerals have been used for like elements.

Terms such as first, second, A, and B may be used to describe various elements, but the elements should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component. and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

When a component is referred to as being “connected” or “connected” to another component, it is understood that the other component may be directly connected or connected to the other component, but other components may exist in between. it should be On the other hand, when it is said that a certain element is "directly connected" or "directly connected" to another element, it should be understood that no other element is present in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as “comprise” or “have” are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

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

1 is a conceptual diagram for explaining a method for identifying an object and an object position identical between images according to an embodiment of the present invention.

First, in an embodiment of the present invention, the first image 10 and the second image 20 for identity identification may be selected from among a plurality of images collected based on crowdsourcing. In this case, when images are collected based on crowdsourcing, even images shot in the same place are collected from multiple photographers, so the same object can be photographed from different angles. Therefore, in FIG. 1 , for convenience of explanation, an image obtained by photographing the rear side of a specific vehicle at one location is the first image 10 , and an image obtained by photographing the rear of the vehicle and the same location as the first image 10 . is illustrated as the second image 20 .

Next, region separation masks for the first image 10 and the second image 20 may be generated to separate the object region and the background region from each of the first image 10 and the second image 20 . (S100). An object of interest to be identified may be detected from the first image 10 and the second image 20 , and a region separation mask may be generated using the detected object of interest. Here, for object detection, learning of the size and location of a bounding box for an object, and learning of classification may be performed.

In this case, the region separation mask may be generated by setting the pixel value to white for the region detected as the object of interest and setting the pixel value to black for the region other than the object of interest. Referring to reference numeral 10a, a region separation mask for the first image 10 may be identified, and referring to reference numeral 20a, a region separation mask for the second image 20 may be identified.

When the region separation mask is generated, the first image 10 and the second image 20 can be separated into an object and a background using region separation masks for each of the first image 10 and the second image 20 . There is (S110). Specifically, the first image 10 may be divided into an object 10b and a background 10c by multiplying the first image 10 by the region separation mask 10a of the first image 10 in units of pixels. . In addition, the second image 20 may be divided into an object 20b and a background 20c by multiplying the second image 20 by the region separation mask 20a of the second image 20 in units of pixels.

When the object and the background are separated, the object 10b for the first image 10 and the object 20b for the second image 20 are compared with each other to perform a process for determining the identity of the object, , a process for determining the identity of an object position may be performed by comparing the background 10c of the first image 10 and the background 20c of the second image with each other.

In order to determine the identity of an object, a feature point (or feature block) is extracted for each of the object 10b for the first image 10 and the object 20b for the second image 20 ( S120 ), and the first The feature points extracted from the object 10b for the image 10 and the feature points extracted from the object 20b for the second image 20 may be matched with each other ( S122 ).

In addition, in order to determine the identity of the object position, a feature point is extracted from each of the background 10c of the first image 10 and the background 20c of the second image 20 ( S121 ), and the first image ( The feature points extracted from the object 10c for 10) and the feature points extracted from the object 20c for the second image 20 may be matched with each other ( S123 ).

As a method of extracting a feature point according to step S120 or step S121, a feature point strong against a geometric change of an object may be extracted by using a feature point extraction technique such as SIFT or MSER.

As a result of matching feature points in step S122 or step S123, positions and numbers of feature points corresponding to each other may be calculated. In this case, when the feature points correspond to each other, the sum of the absolute values of the pixel differences according to the two feature points to be compared (Sum of Absolute Difference, SAD operation value) is smaller than the threshold value, or when the two feature points are mutually exclusive It may be a case of being configured with the same pixel value.

Based on the result of matching the feature points between the object and the background according to steps S120 and S121, it is possible to determine the identity of the object and the position of the object (S124).

2 is an exemplary diagram for explaining a method for matching feature points between images according to an embodiment of the present invention.

Referring to FIG. 2 , a result of performing feature point matching between objects on two images 30 and 40 obtained by photographing the same object at different places can be confirmed.

Specifically, it can be confirmed that all five feature points are extracted from the object included in the image 30 shown on the left side of FIG. 2 , and three feature points among the five extracted feature points are included in the image 40 located on the right side. It can be confirmed that the three feature points extracted from the obtained object correspond to each other.

As shown in FIG. 2 , according to an embodiment of the present invention, the identity of the object may be determined based on the number of feature points corresponding to each other between the two images among the feature points extracted from the objects of the two images. For example, when the number of feature points corresponding to each other among feature points extracted from objects included in each of the two images is equal to or greater than a preset threshold, it may be determined that the two objects are identical to each other.

Similar to the determination of object identity, when the number of feature points corresponding to each other among the feature points extracted from the backgrounds of two images is equal to or greater than a preset threshold, it can be determined that the positions of the objects are the same. This is because, because the feature points on the background correspond to each other, it may mean that two images are images taken at the same or similar places, so that the positions of the objects may be determined to be the same or similar to each other.

As a specific example, among the feature points extracted from the object, the number of corresponding feature points between the two images is equal to or greater than (or exceeding) the threshold, and the number of corresponding feature points between the two images among the feature points extracted from the background is equal to or greater than the threshold (or exceeding), it can be determined that the same object is in the same location.

Also, among the feature points extracted from the object, the number of corresponding feature points between the two images is less than (or less than) the threshold, and the number of corresponding feature points between the two images among the feature points extracted from the background is greater than (or greater than) the threshold , it may be determined that different objects are located at the same location.

Also, among the feature points extracted from the object, the number of corresponding feature points between the two images is equal to or greater than the threshold value (or more than), and the number of corresponding feature points between the two images among the feature points extracted from the background is less than (or less than) the threshold value , it may be determined that the same object is located at different locations.

Also, among the feature points extracted from the object, the number of corresponding feature points between two images is less than (or less than) the threshold, and the number of corresponding feature points between the two images among the feature points extracted from the background is less than (or less than) the threshold. , it may be determined that different objects are located at different positions.

3 is a representative flowchart of a method for identifying the identity of an object and an object position between images according to an embodiment of the present invention.

Referring to FIG. 3 , a method for identifying the identity of an object and an object position between images includes collecting a plurality of images based on crowdsourcing, and separating the plurality of images into an object region and a background region, respectively. , identifying the identity of an object by matching feature points extracted from the images separated by the object region with each other, and identifying the identity of an object location by matching feature points extracted from images separated by the background region with each other. can

The separating into the object region and the background region may include classifying the plurality of images into one or more groups based on location information on the plurality of collected images, and dividing each of the plurality of images classified into the same group into an object region and Separation into a background region may be included.

The location information for the plurality of images may be GPS information obtained at the time of capturing an image by an imaging device that has captured each of the plurality of images.

Since the images collected based on crowdsourcing are images taken by various users in various places, a large computational load may occur in analyzing the entire collected image. In this case, since the images having the same or similar location information at which the images were captured were captured at locations similar to each other, the images have a high correlation with each other. Accordingly, according to an embodiment of the present invention, images having the same or similar locations at which the collected images were captured are clustered into the same group, and the object and object location are identical between images captured at the same (or similar) location. It is possible to maintain the accuracy of image analysis while reducing the computational load by identifying

The separating into the object region and the background region may include extracting an object from each of the plurality of images and generating a region separation mask for each of the plurality of images by using the extracted object.

After generating the region separation mask, the method may further include multiplying the generated region separation mask by the plurality of images in units of pixels.

The step of identifying the identity of the object may include calculating the number of matching feature points and identifying the identity of the object based on the calculated number of feature points.

The identifying of object identity based on the number of feature points may include determining that there is object identity when the number of feature points is greater than a preset threshold value.

The step of identifying the identity of the object location includes: classifying the plurality of images into one or more groups according to a result of identifying the identity of the object; and identifying the identity of the object location with respect to each other. That is, according to an embodiment of the present invention, unnecessary computational load can be reduced by clustering images having the same object into the same group and performing feature point matching between background regions on an image including the same object. .

4 is a block diagram of an apparatus for identifying an object and a position of an object between images according to an embodiment of the present invention.

Referring to FIG. 4 , the apparatus for identifying the identity of an object and an object position between images includes at least one processor and instructions for instructing the at least one processor to perform at least one step. It may include a memory (memory) for storing the.

The at least one step may include: collecting a plurality of images based on crowdsourcing; separating the plurality of collected images into an object region and a background region, respectively; matching feature points extracted from the images separated into the object region with each other and identifying the identity of the object and identifying the identity of the object position by matching feature points extracted from the images separated by the background area.

The separating into the object region and the background region may include classifying the plurality of images into one or more groups based on location information on the plurality of collected images, and dividing each of the plurality of images classified into the same group into an object region and Separation into a background region may be included.

The location information for the plurality of images may be GPS information obtained at the time of capturing an image by an imaging device that has captured each of the plurality of images.

The separating into the object region and the background region may include extracting an object from each of the plurality of images and generating a region separation mask for each of the plurality of images by using the extracted object.

After generating the region separation mask, the method may further include multiplying the generated region separation mask with the plurality of images in units of pixels.

The step of identifying the identity of the object may include calculating the number of matching feature points and identifying the identity of the object based on the calculated number of feature points.

The identifying of object identity based on the number of feature points may include determining that there is object identity when the number of feature points is greater than a preset threshold value.

The step of identifying the identity of the object location includes: classifying the plurality of images into one or more groups according to a result of identifying the identity of the object; identifying the identity of the object location with respect to each other.

5 is a conceptual diagram for explaining a road environment management system using crowd-sourcing-based images to which a method and apparatus for identifying an object and an object position are identical between images according to an embodiment of the present invention.

Referring to FIG. 5 , the road environment management system using crowdsourcing-based images collects a plurality of images captured from one or more user terminals 55 and 56 and location information on the one or more user terminals 55 and 56 . Receive and transmit the received plurality of images and the location information to the image identification device 60 to learn and learn the plurality of images received from the road control server 50 and the road control server to receive the analysis result and an image identification device 60 for transmitting the analysis result obtained by performing identification of the object and the position of the object included in the plurality of images to the road control server using one data and the location information.

The user terminals 55 and 56 are various devices that shoot or store images, collect location information, and collect sensor information (eg, vibration information), such as various cameras and smartphones used by the general public or citizens. can In more detail, for example, the user terminals 55 and 56, which can communicate, a desktop computer, a laptop computer, a notebook, a smart phone, a tablet PC, Mobile phone, smart watch, smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) player, digital audio recorder, digital audio player, digital video recorder, digital video player, PDA (Personal Digital Assistant), etc. can

Images collected by multiple users through the user terminals 55 and 56, location information of the user terminal at the time of shooting, and vibration information may be transmitted to the road control server 50 . For example, the user terminals 55 and 56 transmit images, report messages, sensor information, etc. related to road environment information (eg, illegal parking report, pothole status, road kill status, road obstacle, etc.) to the road control server 50 ) can be transmitted.

The road control server 50 may monitor the road environment state in real time based on the analysis result received from the image identification device 60 . Specifically, the road control server 50 analyzes the images received from the plurality of user terminals 55 and 56 in conjunction with the image identification device 60, monitors the real-time road environment based on the analysis result, and monitors the Various services, including a real-time road environment information provision service and a response action status provision service for the reported road environment information, may be provided to the user terminal through a wired/wireless network. For example, the road control server 50 sets an illegal parking area, and identifies the parking state of the illegal parking area through image analysis in connection with the image identification device 60 for the set illegal parking area, and A corresponding action state may be provided to the user terminal.

The image identification device 60 is a device with a built-in device for identifying the identity of an object and an object position between the images according to FIG. 4 , and collects various tipped images received from the road control server 50 as big data, and , by learning the collected images, it is possible to determine the identity of the object and the identity of the object position, and the illegal parking state, the road kill state, and the pothole state can be determined according to the determination result. However, in another embodiment, when the image identification device 60 transmits the analysis result of determining the identity of the object and the identity of the object position to the road control server 50, the road control server 50 based on the received analysis result It can also determine illegal parking status, road kill status, pothole status, etc.

The image identification device 60 uses the images collected through the road control server 50 as learning data, learning object identification learning, learning to identify the location of an object, learning to divide the image into various areas including objects and backgrounds, Learning about vibration information may be performed. In addition, the image identification apparatus 60 may be interpreted as including components and operations according to the method or apparatus described with reference to FIGS. 1 to 4 , and detailed descriptions will be omitted to prevent duplicate description.

Meanwhile, although not shown in FIG. 5 , there may be a separate data collection server that collects images, location information, and vibration information from the user terminals 55 and 56 , and the data collected from the data collection server is RTSP (real-time). Streaming Protocol) may be transmitted to the road control server 50 through a transmission/reception interface such as an API (Application Programming Interface), REST (Representational State Transfer) API, and the like.

6 is an exemplary diagram for explaining the effect of the method for identifying the object and the object location are identical between images according to an embodiment of the present invention.

The road environment management can be more accurately and easily performed by using the method, apparatus, and system described with reference to FIGS. 1 to 5 .

Specifically, images having location information corresponding to an illegal parking area may be selected from among the images reported by a plurality of users, and object identity and object location identity between the selected images may be determined. Here, when images having the object identity and object location identity are identified, the vehicle according to the corresponding object may determine that the vehicle parked illegally in a time interval in which images for which identity is recognized are taken.

In particular, according to an embodiment of the present invention, since the identity of the object as well as the identity of the object is determined, as shown in FIG. 6 , whether the image is taken of the same vehicle in the same location from different angles or the same vehicle is used in different places It is possible to accurately determine whether parking is illegal by checking whether it is a recorded video or not.

Here, the process of determining illegal parking has been described as an example, but it can be used to identify and quickly respond to potholes, road kills, and road obstacles that may affect the road environment.

The methods according to the present invention may be implemented in the form of program instructions that can be executed by various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer-readable medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software.

Examples of computer-readable media may include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions may 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. The hardware device described above may be configured to operate as at least one software module to perform the operations of the present invention, and vice versa.

In addition, the above-described method or apparatus may be implemented by combining all or part of the configuration or function, or may be implemented separately.

Although the above has been described with reference to preferred embodiments of the present invention, those skilled in the art can variously modify and change the present invention within the scope without departing from the spirit and scope of the present invention as set forth in the claims below. You will understand that it can be done.

Claims (20)

As a method for identifying the identity of an object and an object position between images,
collecting a plurality of images based on crowdsourcing;
separating the plurality of collected images into an object area and a background area, respectively;
identifying the identity of an object by matching feature points extracted from the images separated into the object region; and
identifying the identity of the object position by matching feature points extracted from the images separated by the background region with each other;
Identifying the identity of the object comprises:
calculating the number of matching feature points; and
Comprising the step of identifying the identity of the object based on the number of calculated feature points,
Calculating the number of matching feature points comprises:
When the sum of the absolute values of the pixel differences according to the compared feature points (Sum of Absolute Difference; SAD operation value) is less than the threshold value set in advance by the user, or when the feature points are composed of the same pixel values, a matching feature point A method for identifying identity of an object and object position between images comprising determining. In claim 1,
The step of separating the object area and the background area comprises:
classifying the plurality of images into one or more groups based on the collected location information on the plurality of images; and
A method for identifying the identity of an object and an object position between images, comprising dividing each of a plurality of images classified into the same group into an object region and a background region. In claim 2,
The location information for the plurality of images,
A method for identifying the identity of an object and an object position between images, which is GPS information obtained at the time of capturing an image by an image capturing device that has captured each of the plurality of images. In claim 1,
The step of separating the object area and the background area comprises:
extracting an object from each of the plurality of images; and
and generating a region separation mask for each of the plurality of images by using the extracted object. In claim 4,
After generating the region isolation mask,
The method for identifying the identity of an object and an object position between images, further comprising multiplying the generated region separation mask with the plurality of images on a pixel-by-pixel basis. delete In claim 1,
The step of identifying the identity of the object based on the number of the feature points,
and determining that the object is identical when the number of the feature points is greater than a preset threshold. In claim 1,
Identifying the identity of the object location comprises:
classifying the plurality of images into one or more groups according to a result of identifying the identity of the object; and
A method for identifying the identity of an object and an object position between images, comprising: identifying the identity of an object position with respect to images classified as the background region among a plurality of images classified into the same group. A device for identifying the identity of an object and an object position between images,
at least one processor; and
a memory for storing instructions instructing the at least one processor to perform at least one step;
The at least one step is
collecting a plurality of images based on crowdsourcing;
separating the plurality of collected images into an object area and a background area, respectively;
identifying the identity of an object by matching feature points extracted from the images separated into the object region; and
identifying the identity of the object position by matching feature points extracted from the images separated by the background region with each other;
Identifying the identity of the object comprises:
calculating the number of matching feature points; and
Comprising the step of identifying the identity of the object based on the number of calculated feature points,
Calculating the number of matching feature points comprises:
When the sum of the absolute values of the pixel differences according to the compared feature points (Sum of Absolute Difference; SAD operation value) is less than the threshold value set in advance by the user, or when the feature points are composed of the same pixel values, a matching feature point An apparatus for identifying identity of an object and object position between images, comprising determining. In claim 9,
The step of separating the object area and the background area comprises:
classifying the plurality of images into one or more groups based on the collected location information on the plurality of images; and
An apparatus for identifying the identity of an object and an object position between images, comprising dividing each of a plurality of images classified into the same group into an object region and a background region. In claim 10,
The location information for the plurality of images,
An apparatus for identifying the identity of an object and an object position between images, which is GPS information obtained at the time of capturing an image in an image capturing device that has captured each of the plurality of images. In claim 9,
The step of separating the object area and the background area comprises:
extracting an object from each of the plurality of images; and
and generating a region separation mask for each of the plurality of images by using the extracted object. In claim 12,
After generating the region isolation mask,
The apparatus for identifying the identity of an object and an object position between images, further comprising multiplying the generated region separation mask by the plurality of images in units of pixels. delete In claim 9,
The step of identifying the identity of the object based on the number of the feature points,
and determining that the object is identical when the number of the feature points is greater than a preset threshold. In claim 9,
Identifying the identity of the object location comprises:
classifying the plurality of images into one or more groups according to a result of identifying the identity of the object; and
An apparatus for identifying the identity of an object and an object position between images, comprising: identifying the identity of an object position with respect to the images classified as the background region from among a plurality of images classified into the same group. A road environment management system using crowdsourcing-based images,
a road control server that receives a plurality of images taken from one or more user terminals and location information on the one or more user terminals, and transmits the received plurality of images and the location information to an image identification device to receive an analysis result; and
It learns a plurality of images received from the road control server, and uses the learned data and the location information to transmit the analysis result of identifying the object included in the plurality of images and the location of the object to the road control server. Including a video identification device for transmitting,
The road control server monitors a real-time road environment condition based on the analysis result,
The image identification device,
at least one processor; and
a memory for storing instructions instructing the at least one processor to perform at least one step;
The at least one step is
collecting the plurality of images;
separating the plurality of collected images into an object area and a background area, respectively;
identifying the identity of an object by matching feature points extracted from the images separated into the object region; and
identifying the identity of the object position by matching feature points extracted from the images separated by the background region with each other;
Identifying the identity of the object comprises:
calculating the number of matching feature points; and
Comprising the step of identifying the identity of the object based on the number of calculated feature points,
Calculating the number of matching feature points comprises:
When the sum of the absolute values of the pixel differences according to the compared feature points (Sum of Absolute Difference; SAD operation value) is less than the threshold value set in advance by the user, or when the feature points are composed of the same pixel values, a matching feature point A road environment management system using crowd-sourced images, comprising the step of determining. delete In claim 17,
The step of separating the object area and the background area comprises:
classifying the plurality of images into one or more groups based on the location information; and
A road environment management system using crowd-sourcing-based images, comprising dividing each of a plurality of images classified into the same group into an object region and a background region. In claim 17,
The step of separating the object area and the background area comprises:
extracting an object from each of the plurality of images; and
A road environment management system using crowd-sourcing-based images, comprising generating a region separation mask for each of the plurality of images by using the extracted object.

Images