KR101868017B1 - Method for stereo matching and apparatus thereof - Google Patents

Abstract

A stereo matching method according to the present invention includes the steps of determining a plurality of search target blocks in a search target image corresponding to a reference block in a reference image based on a maximum deviation value, Determining a maximum correlation value among the block-based correlation values, determining a search target pixel corresponding to the maximum correlation value as a matching pixel for the reference pixel, And determining a corresponding variation value as a variation value of the reference pixel. According to the present invention, the stereo matching efficiency can be improved.

Description

≪ Desc / Clms Page number 1 > METHOD FOR STEREO MATCHING AND APPARATUS THEREOF &

The present invention relates to image processing, and more particularly, to a stereo matching method and apparatus thereof.

Digital broadcasting service using 3D video is attracting attention as next generation broadcasting service following HDTV along with UDTV service. Based on the development of related technology such as the launch of high-resolution commercial 3D display, it is possible to enjoy 3D video in each home 3DTV services are expected to be available within the next few years.

The 3D broadcasting service currently being used as a commercial service or a trial service is a service using stereoscopic video mainly composed of left and right images. Unlike conventional 2D video, 3D video may have parallax between the left and right images.

SUMMARY OF THE INVENTION The present invention provides a stereo matching method and apparatus capable of improving stereo matching efficiency.

It is another object of the present invention to provide a method and apparatus for deriving a variation value that can improve stereo matching efficiency.

It is another object of the present invention to provide a depth information generating method and apparatus capable of improving stereo matching efficiency.

One embodiment of the present invention is a stereo matching method. The method includes: determining a plurality of search target blocks in a search target image corresponding to a reference block in a reference image, based on a maximum disparity value; determining, for each of the plurality of search target blocks, Determining a maximum correlation value among the block correlation values and determining a search target pixel corresponding to the maximum correlation value as a matching pixel for the reference pixel; Determining a variation value corresponding to a matching pixel as a variation value of the reference pixel, wherein the reference block includes the reference pixel, each of the plurality of search target blocks includes a search target pixel corresponding thereto , In the step of deriving the block unit correlation value, the first sub-block for the reference block and the plurality of search target blocks Block for each of the first sub-blocks, derives a partial correlation value of the second sub-block with respect to the first sub-block, and calculates a partial correlation value for the second sub- sum), and the first sub-block and the second sub-block are each a column.

According to the stereo matching method of the present invention, the stereo matching efficiency can be improved.

According to the mutation value derivation method of the present invention, the stereo matching efficiency can be improved.

According to the depth information generating method of the present invention, the stereo matching efficiency can be improved.

FIG. 1 is a diagram schematically showing an embodiment of a stereo matching method based on a block comparison method.
2 is a diagram schematically showing another embodiment of a stereo matching method based on a block comparison method.
3 is a diagram for explaining a relationship between degrees of correlation for two adjacent pixels.
4 is a diagram schematically showing another embodiment of a stereo matching method based on a block comparison method.
5 is a flow chart schematically illustrating an embodiment of a stereo matching method according to an embodiment of the present invention.
6 is a block diagram schematically illustrating an embodiment of a stereo matching apparatus according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following description of the embodiments of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present disclosure rather unclear.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . In addition, the description of "including" a specific configuration in the present invention does not exclude a configuration other than the configuration, and means that additional configurations can be included in the practice of the present invention or the technical scope of the present invention.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

In addition, the components shown in the embodiments of the present invention are shown independently to represent different characteristic functions, which does not mean that each component is composed of separate hardware or software constituent units. That is, each constituent unit is included in each constituent unit for convenience of explanation, and at least two constituent units of the constituent units may be combined to form one constituent unit, or one constituent unit may be divided into a plurality of constituent units to perform a function. The integrated embodiments and separate embodiments of the components are also included within the scope of the present invention, unless they depart from the essence of the present invention.

In addition, some of the components are not essential components to perform essential functions in the present invention, but may be optional components only to improve performance. The present invention can be implemented only with components essential for realizing the essence of the present invention except for the components used for performance improvement, Are also included in the scope of the present invention.

FIG. 1 is a diagram schematically showing an embodiment of a stereo matching method based on a block comparison method.

In the 3D image, the same scene can be photographed simultaneously using two or more cameras. For example, a stereo image can be acquired through two cameras including a left camera and a right camera. Here, the image obtained through the left camera may be referred to as a left image, and the image obtained through the right camera may be referred to as a right image.

The angle of view of the left camera and the right camera may be different from each other. Therefore, there may be a viewpoint difference between the left camera and the right camera, and parallax may occur between the left and right images due to the viewpoint difference. Due to the parallax, the position of the pixel representing one point in the left image and the position of the pixel representing the same point on the right side may be different from each other. Accordingly, the parallax can be represented by the disparity value between the left image pixel indicating the same point and the right image pixel. Hereinafter, the pixels of the left image indicating the same point and the pixels of the right image are referred to as the same pixel.

The variation value for each pixel can be derived from the stereo image. Once the disparity value for each pixel is derived, the three-dimensional depth information can be calculated based on the derived disparity value, the focal length of the camera, and the baseline distance of the camera. Here, the baseline distance may indicate the distance between the left camera and the right camera on the same axis when the left camera and the right camera exist on the same axis. Accordingly, in deriving a depth image from a stereo image or deriving an image at an arbitrary point in time using the depth image, it is very important to derive a variation value for the same pixel.

The above-described variation value can be derived by finding a pixel corresponding to each of all the pixels in the left image and / or right image. Here, a pixel corresponding to an arbitrary pixel may mean a pixel representing the same point as the point indicated by the arbitrary pixel. Hereinafter, a pixel representing the same point as the point indicated by any pixel is referred to as a matching pixel for the arbitrary pixel. The process of finding pixels corresponding to each of all the pixels in the left and / or right image may be referred to as 'stereo matching'.

There are various types of stereo matching methods. For example, in the stereo matching method, there may be a local matching method in which matching pixels are considered only considering an area around a current pixel, and a global matching method in which a matching pixel is searched considering an entire image. Hereinafter, a stereo matching method based on a block comparison method among local matching methods will be described.

Hereinafter, an image including a reference pixel from a left image and a right image is referred to as a reference image, and an image other than a reference image is referred to as a search target image. Here, the reference pixel may mean a pixel currently being processed to find a matching pixel in the search target image.

The stereo matching apparatus can find a matching pixel corresponding to each of all the pixels in the reference image in the search target image.

When the reference pixel is determined, the stereo matching apparatus can determine a reference block including the reference pixel as a center. In addition, the stereo image may have a predetermined maximum deviation value, and a pixel positioned at the same position as the reference pixel may exist in the search target image. Hereinafter, a pixel at a position spatially the same as the reference pixel in the search target image is referred to as a co-located pixel. The stereo matching apparatus searches all the pixels existing in the search target image from the '-maximum mutation' position (for example, left) to the '+ maximum mutation' position (for example, right) Position pixel) can be used as a matching pixel candidate for the reference pixel. Hereinafter, the matching pixel candidate is referred to as a search target pixel. Here, each of the pixels to be searched may correspond to one variation value. Hereinafter, a variation value corresponding to a search target pixel is referred to as a variation value with respect to a reference pixel. At this time, for each search target pixel, there may be a block including the search target pixel at the center, which is called a search target block. Each search target block may have the same size as the reference block.

Referring to FIG. 1, search target blocks in a search target image are shown. Reference numeral 110 in FIG. 1 denotes a search target block existing in the same position as the reference block in the search target image, and 115 in FIG. 1 may indicate a search target pixel existing at the same position as the reference pixel in the search target image. In FIG. 1, reference numeral 120 denotes a search target block including a search target pixel existing at a position of '-maximum mutation' on the basis of a co-located pixel, and 130 of FIG. Quot; + < / RTI > maximum mutation " position.

For the sake of convenience of explanation, FIG. 1 shows the search target blocks between the search target block of FIG. 110 and the search target block of FIG. 120, and some of the search target blocks between the search target block of FIG. Are omitted. Also in the embodiments of FIGS. 2 to 4 described later, a block or a part of a column may be omitted in the same or similar manner as in FIG.

The stereo matching apparatus can calculate the correlation with the reference block for every search target block. The degree of correlation may indicate a degree of similarity between the search target block and the reference block. For example, the correlation may be obtained by calculating a cross-correlation value between a search target block and a reference block and / or a sum of absolute difference (SAD) value.

In order to calculate the correlation, the stereo matching apparatus can search all the pixels in the reference block and the pixels in the search object block corresponding to the respective pixels. For example, when SAD is used, the stereo matching apparatus can obtain the absolute value of the difference with respect to all the pixels in the reference block, from the pixel in the search object block corresponding to each pixel. At this time, the stereo matching apparatus can derive the SAD value by adding all the obtained absolute values. The correlation can be obtained in the same way for all search target blocks. When the maximum variation value is defined as in the embodiment of FIG. 1, a correlation value of '2 * maximum variation value + 1' can be obtained for one reference pixel.

If the degree of correlation is obtained for all the search target blocks, the stereo matching device can determine the search target pixel existing in the center of the search target block having the highest correlation value as the matching pixel for the reference pixel. As described above, since each of the pixels to be searched can correspond to one variation value, the determined matching pixel can also correspond to one variation value. At this time, the stereo matching apparatus can determine or select the variation value of the determined matching pixel as the variation value with respect to the reference block.

However, according to the above-described method, the stereo matching apparatus can determine a matching pixel for a reference block by considering only the degree of correlation on a block-by-block basis. Since the matching pixel may not be determined accurately when only the block-by-block correlation is considered, the stereo matching apparatus further considers the similarity between the center pixel of the reference block (reference pixel) and the center pixel of the search target block To determine a matching pixel for the reference block.

In one embodiment, the stereo matching apparatus determines a matching pixel by the above-described method, and the search target pixel whose absolute value of the difference value from the reference pixel is higher than the predetermined threshold value is determined as a matching pixel for the reference pixel . That is, when the absolute value of the difference between the reference pixel and the search target pixel is higher than a predetermined threshold value, the stereo matching apparatus can determine that the search target pixel is not a matching pixel of the reference pixel.

In another embodiment, the stereo matching apparatus may calculate the final correlation by taking into account both the degree of correlation in units of whole blocks and the degree of correlation in units of pixels.

The degree of correlation in block units can be obtained in the same manner as in the above embodiment. For example, the stereo matching apparatus can search or compare all the pixels in the reference block and the pixels in the search target block corresponding to the respective pixels to obtain correlation on a block-by-block basis. At this time, for example, the degree of correlation on a block-by-block basis can be obtained by obtaining the cross-correlation value and / or the SAD value. The degree of correlation on a pixel basis can be derived based on the center pixel (reference pixel) of the reference block and the center pixel (search target pixel) of the search target block. In one example, the stereo matching apparatus can derive the pixel-by-pixel correlation by obtaining the absolute value of the difference between the reference pixel and the search target pixel.

When the degree of correlation in units of blocks and the degree of correlation in units of pixels are obtained, the stereo matching apparatus can derive the final correlation value of the search object block by the weighted sum of the degree of correlation and the degree of correlation in units of pixels. At this time, the stereo matching apparatus can apply different weights to the block-by-block correlation and the pixel-by-pixel correlation.

When the final correlation value is obtained for all the search target blocks, the stereo matching apparatus can determine the search target pixel existing in the center of the search target block having the highest final correlation value as the matching pixel for the reference pixel. As described above, since each of the pixels to be searched can correspond to one variation value, the determined matching pixel can also correspond to one variation value. At this time, the stereo matching apparatus can determine or select the variation value of the determined matching pixel as the variation value with respect to the reference block.

2 is a diagram schematically showing another embodiment of a stereo matching method based on a block comparison method. The embodiment of FIG. 2 may correspond to a digital hardware implementation structure according to the present invention.

In the present invention, in order to derive the correspondence between the pixels, the correlation values for the reference block of the search target blocks may be compared with each other. As described above, each of the search target blocks corresponds to one search target pixel, and each search target pixel corresponds to one variation candidate. Therefore, the correlation of the search object block may be viewed as correlation with the mutation candidate corresponding to the search object block.

On the other hand, each of the pixels in the reference image may have a variation value candidate of a predetermined number (for example, '2 * maximum variation value + 1'). Thus, for each of the pixels in the reference image, the stereo matching device may store correlation values for all the variation candidate values that each pixel can have. However, in this case, since the amount of data to be stored may be excessively large, a stereo matching method that can solve this problem is needed.

In order to solve the above-described problem, a stereo matching method for obtaining the maximum value of the degree of correlation for each pixel clock can be provided. To this end, the stereo matching apparatus calculates a correlation for all the variation candidate values of the reference pixel for each pixel clock, and calculates a maximum correlation value (for example, The minimum value of the SAD). In this case, a variation value corresponding to each pixel can be derived.

Referring to FIG. 2, the left and right image data may be input to the stereo matching device in a scanning order. At this time, the stereo matching apparatus can calculate the degree of correlation with respect to all the variation candidate values that the reference pixel 210 of the left image and / or the reference pixel 220 of the right image can have for each pixel clock. To this end, the stereo matching apparatus may perform a shift operation on the left image data and / or the right image data in the correlation calculation process.

Correlation values for the reference pixel 210 of the left image may be derived based on the reference block corresponding to the reference pixel 210 and the search target blocks in the right image corresponding to the reference block of the left image. Likewise, the correlation values for the reference pixel 220 of the right image may be derived based on the reference block corresponding to the reference pixel 220 and the search target blocks in the left image corresponding to the reference block of the right image have. In this case, for example, when the maximum variation value is N (N is an arbitrary natural number), the number of search target blocks in the right image corresponding to the reference block of the left image and / May be 2N + 1. A specific example of the method for obtaining the correlation is described in Fig. 1, and thus will not be described here.

Assuming that the size of the reference block is Lr * Lc (where Lr is the size of the row of the reference block and Lc is the size of the column of the reference block), in order to calculate the correlation between one reference block and one search target block, A correlation calculation module having an input size of 2 * Lr * Lc is required. At this time, assuming that the maximum variation value is N, 2 * N + 1 correlation calculation modules are required to simultaneously calculate correlation values for one reference block. In FIG. 2, correlation values for the reference block of the left image and the reference block of the right image may be calculated for each pixel clock, and thus, for example, 4 * N + 2 correlation calculation modules may be used.

Reference numeral 230 in FIG. 2 may indicate correlation values for all the transition value candidates corresponding to the reference pixel. In one embodiment, assuming that the maximum transition value is N in FIG. 2, 4 * N + 2 correlation values can be derived for each pixel clock by 4 * N + 2 correlation calculation modules.

When 4 * N + 2 correlation values are derived for each pixel clock, the stereo matching device obtains the maximum correlation value for the reference pixel of the left image and the maximum correlation value for the reference pixel of the left image through the binary search 240 based on the pipeline structure. And / or the maximum correlation value for the reference pixel of the right image can be obtained. At this time, the stereo matching apparatus can determine the disparity value corresponding to the maximum correlation value as the disparity value with respect to the reference pixel. Therefore, in the stereo matching method of FIG. 2, a variation value can be derived for each pixel clock. The disparity values derived by the stereo matching method of FIG. 2 may constitute a disparity value stream. Assuming that the N value is 64, for example, a transition value may be generated after 9 pixel clocks. Further, since a pipeline structure is used, a variation value can be generated every clock after a certain delay.

Once the disparity value is derived, the stereo matching device may generate depth information based on the derived disparity value and the camera parameters f, b (250). Here, f denotes the focal length of the camera, and b denotes the baseline distance of the camera.

3 is a diagram for explaining a relationship between degrees of correlation for two adjacent pixels.

3 shows search target blocks in a search target image. 3, reference numeral 310 denotes a search target pixel located at the same position as the first reference pixel of the reference image in the search target image, and 320 of FIG. 3 shows the search target pixel at the same position as the second reference pixel of the reference image Represents an existing search target pixel. Here, it is assumed that the second reference pixel is a pixel positioned adjacent to the first reference pixel. In addition, reference numerals 313, 316, and 319 in FIG. 3 denote search target blocks for the first reference pixel, and reference numerals 323, 326, and 329 in FIG. 3 denote search target blocks for the second reference pixel.

As described above, the stereo matching apparatus can calculate the degree of correlation between the search target block and the reference block for all the variation candidate values of each pixel in the reference image. Referring to FIG. 3, since the first reference pixel and the second reference pixel are adjacent to each other, the search target block for the first reference pixel and the search target block for the second reference pixel can share many areas within the block have. Thus, the correlation to the first reference pixel and the correlation to the second reference pixel may share the same or similar partial values with respect to the shared portion in the search target block. Therefore, a stereo matching method using the above-described characteristics can be provided, which will be described later in the embodiment of FIG.

4 is a diagram schematically showing another embodiment of a stereo matching method based on a block comparison method. The embodiment of FIG. 4 may correspond to a digital hardware implementation structure according to the present invention.

Referring to FIG. 4, the left image data and the right image data may be input to the stereo matching device in a scanning order. At this time, the stereo matching apparatus can derive the correlation for all of the variation candidate values that the reference pixel 410 of the left image and / or the reference pixel 420 of the right image can have for each pixel clock. In the embodiment of FIG. 4, it is assumed that the stereo matching apparatus derives the correlation by SAD calculation.

As described above, the correlation value for each of the variation value candidates can be derived based on the reference block and the search target block in the search target image corresponding to the reference block. Therefore, in the following embodiments, the SAD finally derived for each variation value candidate is referred to as a block SAD.

As described above with reference to FIGS. 1 and 2, the stereo matching apparatus can derive a block-based SAD value by directly comparing a reference block and a search target block. However, the stereo matching apparatus may derive a block-based SAD value by calculating and accumulating partial SAD values based on a sub-block for a reference block and a sub-block for a search target block.

In Fig. 4, an embodiment in which a sub-block for a reference block and a sub-block for a search target block correspond to one column is described. Here, the subblock for the reference block in this specification may be referred to as a reference column, and the subblock for the search target block may be referred to as a search target column.

For example, the stereo matching device may determine a reference column that includes a reference pixel at the center. Further, the stereo matching apparatus can determine a search target column including a search target pixel as a center. At this time, the stereo matching apparatus can calculate the partial SAD value with respect to the reference column for every search target column. 4, reference numeral 430 denotes partial SAD values corresponding to a reference pixel.

Once the partial SAD values are derived, the stereo matching device may derive a block-by-block SAD value by performing a SAD running sum 440 based on the derived partial SAD values. For example, it is assumed that a plurality of partial SAD values are accumulated to derive a block-based SAD value for one search target pixel. In this case, the block unit SAD value of another search target pixel adjacent to the search target pixel is obtained by adding the partial SAD value of the new search target column to the previous block SAD value, and adding the SAD value of the new search target column Can be derived by subtracting the SAD value. The partial SAD value of the initially accumulated search target column can be obtained, for example, by using a shift register.

Assuming that the size of the reference column is Lr, a correlation calculation module having an input size of 2 * Lr is required to calculate the partial SAD value between one reference column and one search target column. At this time, the stereo matching apparatus has a function of adding a new partial SAD value and subtracting a first accumulated partial SAD value, thereby obtaining a block-based SAD value for the variation candidate. Therefore, in this case, the same or similar result can be obtained with less hardware resources than the embodiment of FIG.

The contents of 450 and 460 of FIG. 4 are similar to those of the embodiment of FIG. 2, and will not be described here.

5 is a flow chart schematically illustrating an embodiment of a stereo matching method according to an embodiment of the present invention.

Referring to FIG. 5, the stereo matching apparatus may calculate a degree of correlation for each of the variance candidates corresponding to the reference pixel (S510).

For example, the stereo matching apparatus can calculate the correlation with the reference block for each of the search target blocks. In this case, if only the block-by-block correlation is considered, the matching pixel may not be correctly determined. Therefore, the stereo matching apparatus may further consider the similarity between the reference pixel and the search target pixel as described above with reference to FIG. In addition, as described above with reference to FIG. 4, the stereo matching apparatus may derive the final SAD value by calculating and accumulating partial SAD values in units of subblocks in obtaining correlation.

Referring again to FIG. 5, the stereo matching apparatus may derive a variation value of a reference pixel (S520).

If the degree of correlation is obtained for all the search target blocks, the stereo matching device can determine the search target pixel existing in the center of the search target block having the highest correlation value as the matching pixel for the reference pixel. At this time, the stereo matching apparatus can determine the disparity value candidate corresponding to the determined matching pixel as the disparity value of the reference pixel.

The stereo matching device may store correlation values for all the variation candidate values for each pixel in the reference image to derive the variation value of the reference pixel. However, in this case, the amount of data to be stored may be excessively large. 2, the stereo matching apparatus calculates a correlation for all the variation candidate values of the reference pixel for each pixel clock, and calculates a variation value of the reference pixel for each pixel clock based on the calculated correlation, .

Once the disparity value is derived, the stereo matching device may generate depth information based on the derived disparity value and camera parameters (S530). Here, the camera parameters may include the focal length of the camera and the baseline distance of the camera.

6 is a block diagram schematically illustrating an embodiment of a stereo matching apparatus according to an embodiment of the present invention. The stereo matching apparatus according to the embodiment of FIG. 6 may include a disparity value calculation unit 610 and a depth information calculation unit 620.

The disparity value calculation unit 610 may derive a disparity value for each pixel in the stereo image based on the inputted left image data and right image data. For example, the variation value calculation unit 610 may calculate a correlation for each of the variation candidate values corresponding to the reference pixel, and derive the variation value of the reference pixel based on the calculated correlation. Since specific embodiments of the deviation value derivation method have been described above, they will not be described here.

The depth information calculation unit 620 can generate depth information based on the derived disparity value and camera parameters. Here, the camera parameters may include the focal length of the camera and the baseline distance of the camera.

The present invention can be applied to both moving images and still images. According to the present invention, the mutation value and / or depth information can be obtained in real time for each pixel clock, and when the function according to the present invention is implemented in hardware logic, less hardware resources can be used. Thus, the efficiency and / or performance of stereo matching can be improved.

In the above-described embodiments, the methods are described on the basis of a flowchart as a series of steps or blocks, but the present invention is not limited to the order of steps, and some steps may occur in different orders or in a different order than the steps described above have. It will also be understood by those skilled in the art that the steps depicted in the flowchart illustrations are not exclusive and that other steps may be included or that one or more steps in the flowchart may be deleted without affecting the scope of the invention You will understand.

The above-described embodiments include examples of various aspects. While it is not possible to describe every possible combination for expressing various aspects, one of ordinary skill in the art will recognize that other combinations are possible. Accordingly, it is intended that the invention include all alternatives, modifications and variations that fall within the scope of the following claims.

Claims (10)

Determining a plurality of search target blocks in a search target image corresponding to a reference block in a reference image based on a maximum disparity value;
Deriving a block unit correlation value with the reference block for each of the plurality of search target blocks;
Determining a maximum correlation value among the block-based correlation values, and determining a search target pixel corresponding to the maximum correlation value as a matching pixel for a reference pixel; And
And determining a variation value corresponding to the matching pixel as a variation value of the reference pixel,
Wherein deriving the blockwise correlation value comprises:
Deriving absolute values of differences between all pixels included in each of the plurality of search object blocks and all pixels included in the reference block;
Adding the absolute values to derive the correlation value of each of the plurality of search target blocks; And
Deriving absolute values of differences between center pixels included in each of the plurality of search object blocks and center pixels included in the reference block and deriving a pixel unit correlation value for each of the plurality of search object blocks; ≪ / RTI > delete delete The method according to claim 1,
Wherein the maximum correlation value is determined by summing a weighted sum of the block-based correlation value and the pixel-based correlation value. The method according to claim 1,
And determining a maximum correlation value among the block-based correlation values at each pixel clock. Determining a plurality of search target blocks in a search target image corresponding to a reference block in a reference image based on a maximum disparity value;
Deriving a block unit correlation value with the reference block for each of the plurality of search target blocks;
Determining a maximum correlation value among the block-based correlation values, and determining a search target pixel corresponding to the maximum correlation value as a matching pixel for a reference pixel; And
And determining a variation value corresponding to the matching pixel as a variation value of the reference pixel,
Wherein deriving the blockwise correlation value comprises:
A first sub-block for the reference block and a second sub-block for each of the plurality of search target blocks are determined, and a partial correlation degree value with respect to the first sub-block is determined for each of the second sub- ; And
And performing a running sum on the derived partial correlation values. The method of claim 6, wherein
Wherein the first sub-block and the second sub-block are each a column. The method according to claim 6,
Wherein the matching pixel is a search target pixel located at the center of the search target block having the maximum correlation value. The method according to claim 6,
And generating depth information based on the determined variation value and the camera parameter. 10. The method of claim 9,
Wherein the camera parameters include a focal length of the camera and a baseline distance of the camera.

Images