KR20150115335A - Distance measuring apparatus using Multi-view point image and method therefor - Google Patents

Abstract

The present invention relates to a device and a method for measuring the distance between objects in a three-dimensional space using multi-view images capable of reducing the complexity of processing while more accurately processing even in a section with no characteristics by using multiple multi-view collecting devices. The device for measuring the distance includes an image combining unit and a distance calculating unit and enables a stereo vision system applicable to a mobile device with a limit on processing performance by reducing the complexity of processing while more accurately processing even in the section with no characteristics by using the image combining unit and the distance calculating unit. Accordingly, the purpose of the present invention is to provide high accuracy in an urban environment with lots of walls and roads with little characteristics due to repeated similar colors.

Description

BACKGROUND OF THE INVENTION Field of the Invention The present invention relates to a distance measuring apparatus and method therefor,

The present invention relates to an apparatus and method for distance measurement in a three-dimensional space between objects, and more particularly, to an apparatus and method for measuring distance in a three-dimensional space between objects. More specifically, The present invention relates to an apparatus and method for measuring distance in a three-dimensional space between objects using an image.

In a non-marker based augmented reality technique utilizing a three-dimensional structure, it is required to acquire three-dimensional coordinates from the image.

For this purpose, we can use depth camera technology which can measure stereo vision, such as stereoscopic vision using a plurality of image cameras and stereo vision using a single image camera but using images at various points of view.

However, the Depth camera can measure the distance accurately, but it is impossible to measure the distance to a long distance object due to the restriction of the distance. Therefore, the stereo vision technique using plural video cameras is commonly used Increase in processing complexity and lack of accuracy occur

The present invention relates to an apparatus and method for measuring distances in a three-dimensional space between objects using multi-view images. More specifically, the present invention relates to an apparatus and method for measuring distances in three- The present invention aims to provide a stereoscopic vision system that can be applied to a mobile device having a limited number of pixels, thereby providing high accuracy in a city environment in which many similar colors are continuous and many features are not present.

According to an embodiment of the present invention, an apparatus for measuring distances in an inter-object space using multi-viewpoint images is characterized in that a feature for recognizing an object included in the image is extracted by receiving an image acquired at a plurality of viewpoints for the same object An image integrating unit for identifying correspondence points corresponding to the images by using the consistency of the relative positions of the extracted features in the images at each viewpoint and verifying the identity of the identified corresponding points; And

And a distance calculating unit for calculating a distance between the objects in the image through the corresponding points whose identities are verified.

According to an embodiment of the present invention, the image integrating unit may include a feature extracting unit for extracting a feature for receiving the image and identifying an object included in each view-point image; And a feature corresponding unit that identifies and verifies corresponding points corresponding to the correspondence by associating the extracted feature with each view-point image.

According to the embodiment of the present invention, when the feature extracted from the image obtained at a specific time point of the received image is not extracted from the image at the different view point, the feature corresponding unit uses the image at the extracted point of time, A feature correcting unit for correcting feature extraction from an image at a point in time that is not extracted; An area setting unit for generating a polygon based on a plurality of adjacent features among the extracted features to generate an area in which the extracted feature is a vertex and a side of the polygon in each viewpoint image; Wherein each of the viewpoint regions is aligned so that the region corresponding to a specific time point of the generated region can be corresponded to the region generated from another image, and whether or not the feature included in the aligned region corresponds to a relatively same position, A corresponding point identifying unit that identifies whether the corresponding point corresponds to the corresponding point; And an equivalence verifying unit for verifying whether or not the corresponding points in the respective images of the identified corresponding points have the same color value or not by comparing whether the corresponding points in the respective images have the same color value.

According to an embodiment of the present invention, the distance calculating unit may include: a distance calculating unit that calculates disparity-based distances in a stereo vision using all of the identified corresponding points with respect to the point-by-point image; And a candidate determining unit that corrects the distance measurement error by averaging the distances calculated for each corresponding point on all the viewpoint-specific images using the identified corresponding points.

According to an embodiment of the present invention, a method for measuring distances in a three-dimensional space between objects using multi-viewpoint images is performed by receiving images acquired at a plurality of viewpoints for the same object to extract features for the viewpoint- A step of correcting a feature extracted from a feature point extracted from a point-in-time image; And generating regions having a plurality of features located at adjacent distances from the extracted features as vertexes and sides, and generating regions of the generated images at a specific time point from the images of the other viewpoints Aligning the respective viewpoint areas; Determining whether the features included in the aligned region are corresponding points corresponding to the respective identical positions, and verifying whether the determined corresponding points are the same feature among the aligned images; And performing distance calculation through all corresponding points whose identities are verified.

According to an embodiment of the present invention, the step of performing the distance calculation includes a step of correcting the error of the distance measurement by averaging the distances calculated for the corresponding points in each viewpoint-by-view image have.

According to an embodiment of the present invention, the step of aligning the regions according to the respective viewpoints may include polygons based on a plurality of features located at adjacent distances among the extracted features, Creating regions; And aligning the respective viewpoint regions by performing the shear transformation and the parallel movement so that the regions of the image at a specific point in time of the generated regions can be corresponded to the regions generated from the viewpoint-specific images, respectively can do.

According to an embodiment of the present invention, the step of verifying whether the feature is the same may include: identifying whether the feature included in the aligned region corresponds to the corresponding point so that the feature included in the aligned region is relatively the same position with respect to the image at each viewpoint; And verifying whether the corresponding points in the respective images among the identified corresponding points have the same characteristic through whether or not corresponding points have similar color values.

1 is a block diagram of a distance measuring apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a difference in image according to a change in a point of time collected from the multi-view image collecting apparatus.
3 is a detailed configuration diagram of the image integration unit shown in FIG.
Fig. 4 is a detailed configuration diagram of the feature corresponding portion shown in Fig. 3; Fig.
5 is a diagram illustrating extracted features of a multi-view image according to an exemplary embodiment of the present invention.
6 is a detailed configuration diagram of the distance calculating unit shown in Fig.
7 is a diagram illustrating corresponding point identification according to an embodiment of the present invention.
8 is a flowchart illustrating a distance measuring method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when an element is referred to as "comprising ", it means that it can include other elements as well, without excluding other elements unless specifically stated otherwise.

Hereinafter, an apparatus and method for distance measurement in a three-dimensional space between objects using multi-view images according to an embodiment of the present invention will be described with reference to the drawings.

1 is a block diagram showing a distance measuring apparatus 1000 according to an embodiment of the present invention.

The distance measuring apparatus 1000 may include an image integrating unit 100 and a distance calculating unit 200.

The distance measuring apparatus 1000 receives the multi-view image from the multi-view image collecting apparatus 2000, derives the characteristics of the objects included in the image with respect to each view point image, It is possible to calculate the distance between the objects included in the image in the three-dimensional space based on this verification.

The image integrating unit 100 receives an image acquired at a plurality of viewpoints with respect to the same object, extracts a feature for recognizing the object included in the image, and utilizes the consistency of the relative positions of the extracted features with respect to each viewpoint The corresponding points corresponding to each image are identified, and the identities of the corresponding points can be verified.

Among the extracted features, it is possible to identify the correspondence points at each point in the image and verify the identity by using the coherence of the relative positions with respect to the vertexes of the respective points located in the polygons constituting the adjacent features in each viewpoint image.

The image integration unit 100 will be described in more detail with reference to FIG.

The distance calculator 200 may perform disparity-based distance calculations in normal stereoscopic vision for all matched equivalents and may calculate distance calculations based on left, By centering the distances computed for each of the center and center images, it is possible to reduce the distance measurement error between objects included in the image.

The distance calculating unit 200 will be described in more detail with reference to FIG.

FIG. 2 is a diagram showing a difference of images according to a change of a point of time acquired from a multi-view image collecting apparatus.

The multi-view image capturing device is an image capturing device or an image capturing device capable of acquiring an image with a plurality of focal points and acquiring an image of a different view point for the same object, different from an existing camera for acquiring an image with one focus, May be installed in different positions to acquire images of the same object at different viewpoints.

According to an embodiment of the present invention, a light field camera may be used as a multi-view image capturing device, but the present invention is not limited thereto.

The light field camera includes several microlenses on the focal plane of the camera main lens, and images can be acquired with a plurality of foci through the images captured by the lens, and information about the depth of the corresponding image can be obtained by analyzing the images .

Unlike the conventional camera, the light field camera senses light of various angles reaching a two-dimensional plane, so that it is possible to generate images at each viewpoint by one shot as shown in FIG. 2

An image obtained by such a multi-view image collecting apparatus can be shown in FIG.

According to the embodiment of the present invention, the two multi-viewpoint image collecting apparatuses are installed with a certain distance from the left and right in a state in which the optical axes are aligned, and images of left, right and middle viewpoints are acquired from each multi- Thus, a total of six viewpoint images can be obtained for one object.

Referring to FIG. 2, differences in left, right, and middle view images obtained from the multi-view image collection device installed on the left side according to the above-described embodiment can be known.

Even in the case of an image acquired at the same position, the size, color, depth, etc. of the object can be slightly differentiated as shown in FIG. 2 according to the position of the viewpoint.

However, since multi-view images captured by one multi-view image capture device are acquired at the same position, the same features are identified for a common view.

Also, since the features of the left and right viewpoint images of the multi-viewpoint image capture device are the same distance from the camera, they can be basically symmetrical based on the feature of the central viewpoint image.

3 is a detailed configuration diagram of the image integration unit 100 shown in FIG.

The image integration unit 100 may include a feature extraction unit 110 and a feature corresponding unit 120.

The feature extracting unit 110 may receive images of a plurality of viewpoints of the same object from the multi-view image collecting apparatus 2000 and may extract features for identifying the objects included in the received images.

Here, the feature means a singular point that can be recognized so as to recognize and identify an object included in the image. According to one embodiment, the point feature may be an algorithm such as Harris Corner Detector or Scale Invariant Feature Transform (SIFT) The line feature can be extracted using a Line Detector or the like, and the texture characteristics of a face can be extracted using a DoG (Difference of Gaussian) or the like.

Also, point, line, and surface features can be extracted and used according to the application environment. As a result, positions of points, lines, and features of objects included in each image can be identified.

The feature corresponding unit 120 uses the features of the images obtained from the multi-view image collecting apparatus 2000 to be symmetric with respect to the images of the central view, The corresponding point corresponding to the correspondence can be identified and verified.

The features extracted from the multi-view image acquired by the multi-view image capturing apparatus 2000 are used to restore the missing features by using the coherence of the relative positions with respect to the vertexes of the respective points located in the polygons constituting the adjacent features, The distance calculating unit 200 can generate information necessary to calculate the distance between the objects included in the image.

A method of identifying a corresponding point to a feature will be described in more detail with reference to FIG.

Fig. 4 is a detailed configuration diagram of the feature corresponding portion shown in Fig. 3; Fig.

The feature correspondence unit 120 may include a feature correction unit 121, an area setting unit 122, a correspondence point discrimination unit 123, and a consistency verification unit 124.

The feature correction unit 121 uses the symmetry of the multi-view image acquired by the multi-view image acquisition apparatus 200 to determine whether the feature extracted from the image acquired at a specific point in the received image is extracted , It is possible to correct the feature extracted from the image at the point in time when the feature is not extracted by using the image at the extracted point of time.

According to one embodiment, in the case of a multi-view image having three viewpoints at left, right, and center, when the feature is not identified at the central viewpoint image, When the feature is not identified in the left or right view image, the features can be set and corrected so that the features of the left and right images are symmetrical with respect to the feature of the central image.

That is, the features extracted from the images at the left, center, and right viewpoints acquired in the multi-view image capturing apparatus 2000 are common and only the distances between the features generated by the difference in viewpoints must be changed, The characteristic of the image can be corrected.

The area setting unit 122 may generate a polygon based on a plurality of adjacent features among the extracted features, and may generate a feature extracted from each view-point image to be a vertex and sides of the polygon.

According to the embodiment of the present invention, when there are three or more adjacent features, a polygonal area can be created.

The corresponding point identifying unit 123 arranges the respective viewpoint regions through the shear transformation and the parallel movement so that the region corresponding to the specific point in the generated region can be corresponded to the region generated from the other image, Whether the feature corresponds to a relatively same position or not can be identified.

Here, a shear transformation means a coordinate transformation that fixes one of coordinates and transforms several other coordinates.

As a result, one feature in the extracted region can exist at the same position relative to the corresponding region of the other image, and six images can be aligned at the same position relative to the corresponding region as in the embodiment of the present invention Can be identified.

This correspondence identification can be easily realized using an algorithm such as RANSAC, which matches the largest number of features in different images, but is not limited thereto.

The identity verification unit 124 can verify whether or not the corresponding points in each image among the identified corresponding points have the same color value or not.

In order to verify whether the same features are present in the same region, if points located at the same position in the same region are located on the same plane, the corresponding points in the respective images correspond to each other at the same point. .

More specifically, in the case of using six viewpoint images as in the embodiment of the present invention, if corresponding points in six images are similar color values, they can be regarded as corresponding points, and if they are not similar color values, .

For the sake of judgment, the corresponding points in six images can compare similar color values, which can be the way in which six images vote for identity.

As for the criterion of the determination of the identity, the average of the color values may be obtained for four or more neighboring pixels, and the threshold value may be set to And if the difference in the number of pixels defined beforehand in the image does not exceed the threshold value, the method can be used.

At this time, the point where the identity is not guaranteed can be determined as the area where the square is generated by the viewpoint of the camera.

5 is a diagram illustrating extracted features of a multi-view image according to an exemplary embodiment of the present invention.

Referring to FIG. 5, two multi-viewpoint image collecting apparatuses 2000 may be arranged in a state in which the optical axes are aligned and spaced apart from each other by a predetermined distance, as in the embodiment of the present invention.

Six images at the center, left, and right viewpoints can be generated through two multi-viewpoint image collecting apparatuses 2000 disposed on the left and right sides, respectively.

The features of the objects included in the image can be extracted for the six images thus generated.

According to an exemplary embodiment, point features and line features may be extracted and used together, but the present invention is not limited thereto.

As shown in FIG. 3, the feature points and the line features are extracted and used together, so that the feature points and lines can be displayed in red.

Thus, various corners and the like are identified as point features, and columns, doors, ceiling linear structures, etc. can be identified as line features.

6 is a detailed configuration diagram of the distance calculating unit 200 shown in FIG.

Referring to FIG. 6, the distance calculating unit 200 may include a distance calculating unit 210 and a candidate determining unit 220.

The distance calculator 210 calculates a distance based on the disparity based on the normal stereo vision with respect to the multi-view image acquired by the multi-view image collecting apparatus 2000 through all correspondence points whose identities are verified in the identity verifying unit 124 Distance calculation can be performed.

로 나타낼 수 있으며, 이 공식은 왼쪽 영상 한 점의 위치에서 대응되는 오른쪽 영상 한 점의 위치 차이는 그 점의 3차원 공간 상의 거리 z에 반비례하며, 촛점 거리 f와 두 영상을 촬영할 때 거리차 B(베이스라인)에 비례한다는 것을 나타낸다.Here, the disparity-based distance calculation in the conventional stereo vision

This formula shows that the positional difference of a right image corresponding to one point of the left image is inversely proportional to the distance z on the three-dimensional space of the point, and when the focal distance f and the distance difference B (Baseline).

According to the embodiment of the present invention, it is preferable to use a central image of the left and right multi-view image collecting apparatus 2000, but it is not limited thereto.

The candidate determining unit 220 may correct the error of the distance measurement by averaging the calculated distances for each corresponding point on the images of all the viewpoints using the identified corresponding points.

According to the embodiment of the present invention, by using the corresponding points determined in the feature coping unit 120, the distance calculated for each corresponding point on the left, right, center, and center images is averaged, .

However, this postulation can be applied selectively because it becomes a burden on the calculation amount.

7 is a diagram illustrating corresponding point identification according to an embodiment of the present invention.

Referring to FIG. 7, according to an embodiment of the present invention, a corresponding point is identified from a feature obtained from six images obtained through a multi-view image capturing apparatus 2000 that is spaced apart by a predetermined distance in a state in which optical axes are aligned, Can be simplified and expressed.

Fig. 7 shows an embodiment in which a rectangular region is formed by using the edges constituting the wall.

The orange color region corresponds to the left side wall in Fig. 5, and the yellow color region corresponds to the front side wall.

As the viewpoint of the multi-view image collecting apparatus 2000 moves from the left to the right, the viewpoint becomes closer to the vertical direction of the left wall, and the left wall area may be enlarged to a wider range as shown in FIG. have.

At this time, a point on the left wall indicated by green may be maintained in the same relative position in the region even if the viewpoint changes.

By using this characteristic, the corresponding point identification unit 123 can identify the correspondence between the images of the inner region having no characteristic in the region unit made up of the features in each viewpoint image.

As shown in FIG. 7, the relative positions to the vertices of an area of a point of the inner area are the same with respect to the areas of the left wall, so that the respective areas are aligned so that the relative positions of one point (feature) of the inner area are the same , The coordinates of the feature in the image can be obtained through the proportional relation of the aligned regions.

And we can measure the distance between the objects included in the feature through these coordinates.

8 is a flowchart illustrating a distance measuring method according to an embodiment of the present invention.

Referring to FIG. 8, first, a plurality of viewpoint images for a common object are acquired (310).

Unlike conventional cameras that acquire images with a single focus, the multi-view image acquisition apparatus 2000 acquires images through a plurality of focus lenses and acquires images of the same object at different points of view The device or the plurality of cameras may be installed in different positions to acquire images of the same object at different viewpoints.

According to an embodiment of the present invention, a light field camera may be used as a multi-view image capturing device, but the present invention is not limited thereto.

Each feature is extracted from the obtained point-in-time image (320).

Here, the feature means a singular point that can be recognized so as to recognize and identify an object included in the image. According to one embodiment, the point feature may be an algorithm such as Harris Corner Detector or Scale Invariant Feature Transform (SIFT) The line feature can be extracted using a Line Detector or the like, and the texture characteristics of a face can be extracted using a DoG (Difference of Gaussian) or the like.

Also, point, line, and surface features can be extracted and used according to the application environment. As a result, positions of points, lines, and features of objects included in each image can be identified.

(330) so that the feature can be extracted from the image at the point in time when the feature is not extracted.

According to one embodiment, in the case of a multi-view image having three viewpoints at left, right, and center, when the feature is not identified at the central viewpoint image, When the feature is not identified in the left or right view image, the features can be set and corrected such that the features of the left and right images are symmetric with respect to the feature of the center image.

That is, the features extracted from the images at the left, center, and right viewpoints acquired in the multi-view image capturing apparatus 2000 are common, and only the distances between the features generated due to the difference in viewpoints must be changed, The characteristic of the image can be corrected.

(340) a plurality of adjacent extracted features as vertices and sides.

According to the embodiment of the present invention, when there are three or more adjacent features, a polygon can be generated.

The respective viewpoint regions are aligned 350 by performing the shear transformation and the parallel movement so that the region for the specific time point among the generated regions can be corresponded to the region generated from the other images.

By aligning images at each viewpoint, each point within the region can maintain a relative positional relationship with respect to the features constituting the region.

(360) whether the feature included in the aligned region corresponds to a position relatively to the image at each viewpoint and corresponds to the corresponding point.

This correspondence identification can be easily realized using an algorithm such as RANSAC, which matches the largest number of features in different images, but is not limited thereto.

(370) whether each identified corresponding point among the aligned images is the same feature.

In order to verify whether the same feature exists at the same point, if corresponding points in the same position are located on the same plane, the corresponding points in each image may have similar color values.

In the case of using six viewpoint images as in the embodiment of the present invention, if corresponding points in six images are similar color values, they can be regarded as corresponding points, and if they are not similar color values, it can be determined that they are not the same point.

For determination, corresponding points in six images can compare similar color values, which can be the way in which six images vote for identity.

The method of determining the identity may differ slightly depending on the angle of illumination, the camera angle, and the characteristics of the camera, even at the same point in one embodiment. Therefore, an average of color values is obtained for four or more adjacent pixels, And if the difference in the number of pixels defined in advance in the image does not exceed the threshold value, the method can be used.

A distance calculation is performed (380) through all corresponding points whose identities have been verified.

로 나타낼 수 있으며, 이 공식은 왼쪽 영상 한 점의 위치에서 대응되는 오른쪽 영상 한 점의 위치 차이는 그 점의 3차원 공간 상의 거리 z에 반비례하며, 촛점 거리 f와 두 영상을 촬영할 때 거리차 B(베이스라인)에 비례한다는 것을 나타낸다.Here, the disparity-based distance calculation in the conventional stereo vision

This formula shows that the positional difference of a right image corresponding to one point of the left image is inversely proportional to the distance z on the three-dimensional space of the point, and when the focal distance f and the distance difference B (Baseline).

(390) so as to perform the same distance calculation in all the images.

According to the embodiment of the present invention, by using the corresponding points determined in the feature coping unit 120, the distance calculated for each corresponding point on the left, right, center, and center images is averaged, .

However, this postulation can be applied selectively because it becomes a burden on the calculation amount.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, Various modifications and improvements of those skilled in the art using the basic concept of the present invention are also within the scope of the present invention.

100: image integration unit 110: feature extraction unit
120: Feature correspondence unit 121:
122: area setting unit 123: corresponding point identification unit
124: identity verification unit 200: distance calculation unit
210: distance calculation unit 220:

Claims (8)

The method includes extracting a feature for recognizing an object included in the image by receiving an image acquired at a plurality of viewpoints with respect to the same object, An image integration unit for identifying the corresponding points and verifying the identity of the identified corresponding points; And
And a distance calculating unit for calculating a distance between the objects in the image through the corresponding points whose identities are verified. The image processing apparatus according to claim 1,
A feature extraction unit for receiving the image and extracting a feature for identifying an object included in each view-point image; And
Further comprising a feature corresponding unit which identifies and verifies corresponding points corresponding to the correspondence by associating the extracted feature with each image at each view point. . 3. The apparatus according to claim 2,
If a feature extracted from an image obtained at a specific point in time of the received image is not extracted from an image at another point in time, a feature can be extracted from an image at a point in time when the feature is not extracted using the extracted point- A feature correcting unit for correcting the feature correcting unit;
An area setting unit for generating a polygon based on a plurality of adjacent features among the extracted features to generate an area in which the extracted feature is a vertex and a side of the polygon in each viewpoint image;
Wherein each of the viewpoint regions is aligned so that the region corresponding to a specific time point of the generated region can be corresponded to the region generated from another image, and whether or not the feature included in the aligned region corresponds to a relatively same position, A corresponding point identifying unit that identifies whether the corresponding point corresponds to the corresponding point; And
Further comprising an equality verifying unit for verifying whether or not the corresponding points in the respective corresponding images corresponding to the identified corresponding points have a similar color value. Device. The apparatus according to claim 1,
A distance calculation unit for performing disparity-based distance calculation in stereo vision using all of the identified corresponding points with respect to the viewpoint-by-view image; And
Further comprising a candidate unit for correcting the error of the distance measurement by averaging the distances computed for each of the corresponding points with respect to the images at every viewpoint using the identified corresponding points, Distance measuring device in three dimensional space. A feature extracting unit that extracts features of the viewpoint-by-view image by receiving an image acquired at a plurality of viewpoints with respect to the same object, and corrects the features of the viewpoint- ;
And generating regions having a plurality of features located at adjacent distances from the extracted features as vertexes and sides, and generating regions of the generated images at a specific time point from the images of the other viewpoints Aligning the respective viewpoint areas;
Determining whether the features included in the aligned region are corresponding points corresponding to the respective identical positions, and verifying whether the determined corresponding points are the same feature among the aligned images;
And performing distance calculation through all corresponding points of which the identities have been verified. 6. The method of claim 5, wherein performing the distance calculation comprises:
And correcting a distance measurement error by averaging the distance calculated for each corresponding point in each viewpoint image by using the multi-view image. 6. The method of claim 5,
Generating regions having polygons based on a plurality of features located at adjacent distances among the extracted features and having areas of vertexes and sides of the extracted features in each viewpoint image; And
Further comprising the step of aligning the respective viewpoint regions through the shear transformation and the parallel movement so that the region for the image at a specific point in time among the generated regions can correspond to the region generated from the viewpoint- A Method of Distance Measurement in 3 - D Space between Objects Using Multi - view Image. 6. The method of claim 5,
Identifying whether the feature included in the aligned region corresponds to a corresponding point so as to be at a position that is relatively the same as an image at each view point; And
Further comprising verifying whether or not the corresponding points in the respective images of the identified corresponding points have the same color value by checking whether the corresponding points have similar color values. Way.

Images