KR101165810B1 - A method and an apparatus extracting a depth image information using a stereo camera - Google Patents

Abstract

PURPOSE: A method for extracting image depth information by a stereo camera and a device thereof are provided to process a high quality image in real time. CONSTITUTION: A first image and a second image about an object are photographed using a stereo camera(100). The sizes of the images are adjusted using a plurality of scalers(106). Image depth information is generated using the adjusted first and second images wherein the image depth information corresponds to the scalers(108). Optimized image depth information is extracted(110).

Description

A method and an apparatus extracting a depth image information using a stereo camera}

The present invention relates to extracting image depth information through a stereo camera. More specifically, in extracting depth information, the image scaler is changed into image information having different sizes before processing one image information. The present invention relates to a technique for extracting proper image depth information.

In general, the depth information extraction apparatus through a stereo camera uses the binocular difference, i.e., the right eye and the left eye, as in the human eye. Refers to extracting depth information of a target by calculating a difference between input images by creating a remote image input state. When the difference between the two cameras is obtained and the distances of the two cameras are known, the depth information of the corresponding images can be extracted by a calculation formula.

In general, if the two input images coincide with an epipolar line, the same feature point can be extracted by searching in the horizontal direction at the same vertical position of the two images, and the distance of the object is calculated based on this. For this purpose, the same object must be found in both images, so the distortion caused by the lens or camera's point of view must be eliminated as much as possible, so the depth information can be easily extracted by performing the distortion correction process before the distance information is extracted. .

After the distortion correction is performed, the joint reference lines of the two images are matched through the rectification of the images so that the common reference lines of the two images coincide. Then, the feature points are extracted and the position difference of the feature points is changed into depth information.

However, there are some things to consider when generating depth information from a stereo camera. FIG. 1 is a reference diagram for explaining a focal length and a field of view (FOV) of a lens, and FIG. 2 is a 50 [mm] and 105 [mm] lens mounted on a stereo camera, and an image of the same object is displayed. As a reference diagram comparing input cases, the size of the same object is changed according to the viewing angle, and the size of the overlapped area is also changed.

In general, in the case of the wide-angle lens, it is relatively advantageous when extracting the depth information of the near field, but it is difficult to extract the depth information of the far field because the far objects appear relatively small. In the case of a telephoto lens, since the objects in the near field appear relatively large, the range of operations to be performed to extract the depth information in the near field also becomes wider. However, when the range of operations is widened, problems such as the limitation of the speed of the operation apparatus for real-time processing and the delay of bus processing due to the memory required for the operation may occur.

For this reason, when extracting depth information based on a stereo camera, it is difficult to simultaneously extract near depth information and far depth information.

The problem to be solved by the present invention is to extract depth information using two input images, the image by the stereo camera to be able to extract the depth information more efficiently by adjusting the size of the input image using a plurality of scalers A depth information extraction method and apparatus are provided.

In order to solve the above problems, the image depth information extraction method by a stereo camera according to the present invention comprises the steps of photographing the first image and the second image of the object using a stereo camera, respectively; Adjusting a size of an image by using a plurality of scalers for each of the captured first image and the second image; Generating image depth information corresponding to the plurality of scalers, respectively, by using the first image and the second image whose size of the image is adjusted by the plurality of scalers; And extracting optimal image depth information from the generated image depth information.

Preferably, the method for extracting image depth information by the stereo camera further includes correcting image distortion for each of the obtained first and second images, and correcting the first and second images. It is characterized in that for adjusting the size of the image.

Preferably, the step of correcting the image distortion, characterized in that for correcting the brightness information for each of the obtained first image and the second image.

Preferably, the method for extracting image depth information by the stereo camera further includes correcting horizontal lines of the obtained first image and the second image, and correcting the first and second images. It is characterized in that for adjusting the size of the image.

Preferably, the step of adjusting the size of the image using the plurality of scalers is characterized by reducing the size of the image.

Preferably, the step of generating the image depth information corresponding to the plurality of scalers, respectively, by multiplying the inverse of the scale ratio of the plurality of scalers, characterized in that for correcting the image depth information.

Preferably, extracting the optimal image depth information from the generated image depth information, when the image depth information is a predetermined distance or more, the image depth corresponding to the scaler of the plurality of scalers relatively large scale ratio; And extracting information and extracting image depth information corresponding to a scaler having a relatively small scale ratio from among the plurality of scalers when the image depth information is below a predetermined short distance.

In order to solve the above problems, the image depth information extraction apparatus by a stereo camera according to the present invention comprises a stereo camera for capturing the first image and the second image for each object; A plurality of scalers for adjusting a size of an image for each of the captured first image and the second image; A plurality of depth information generators respectively generating image depth information corresponding to the plurality of scalers by using the first image and the second image whose size of the image is adjusted by the plurality of scalers; And a depth information extracting unit which extracts optimal image depth information from the generated image depth information.

Preferably, the apparatus for extracting image depth information by the stereo camera may further include first and second image correctors configured to correct image distortion with respect to each of the captured first and second images. .

The first and second image correctors may correct brightness information with respect to each of the acquired first and second images.

Preferably, the first and second image correctors correct the horizontal lines of each of the captured first image and the second image.

Preferably, the plurality of scalers is characterized in that to reduce the size of the image.

Preferably, the plurality of depth information generators may multiply the reciprocal of the scale ratio of the plurality of scalers to correct the image depth information.

Preferably, the depth information extracting unit extracts image depth information corresponding to a scaler having a relatively large scale ratio from among the plurality of scalers when the image depth information is greater than or equal to a predetermined distance, and the image depth information is a predetermined short range. In the following case, image depth information corresponding to a scaler having a relatively small scale ratio among the plurality of scalers may be extracted.

According to the present invention, by using a plurality of scalers to adjust the size of the input image by processing the extraction of the near and far depth information by dividing, to reduce the volume of the stereo camera, reducing the amount of computation required to extract depth information to less than half You can expect the effect. Accordingly, it is possible to process a high quality image in real time compared to the existing system.

That is, when the amount of computation is the same, the range for extracting depth information is wider than that of the conventional method for extracting image depth information. As a result, when the same range must be detected in the past, the number of installed systems is reduced, resulting in cost savings. In addition, since a higher resolution image may be processed in real time than an image that may be processed in real time, more detailed depth information calculation may be expected. In addition, an effect of increasing reliability through a plurality of depth information of the same image can be expected.

FIG. 1 is a reference diagram for explaining the focal length and field of view (FOV) of a lens.
FIG. 2 is a reference diagram comparing a case in which 50 [mm] and 105 [mm] lenses are mounted on a stereo camera and an image of the same object is input.
3 is a flowchart of an embodiment for explaining a method for extracting image depth information by a stereo camera according to the present invention.
4 is an equation for explaining a method for correcting image distortion and a reference diagram corresponding thereto.
5 is a reference diagram for describing image distortion due to lens distortion.
6 is a reference diagram for explaining image correction due to a mismatch between horizontal lines of first and second images.
7 is an equation and a reference diagram for describing generation of image depth information using a horizontal distance difference between two objects.
8 is a reference graph for explaining that the number of pixels monotonically decreases with distance.
9 is a reference diagram for explaining a dynamic programming method for finding a location where singularities of two images coincide.
10 is a flowchart of an exemplary embodiment for explaining a process of extracting optimal image depth information.
11 is a block diagram of an embodiment for explaining an apparatus for extracting image depth information by a stereo camera according to the present invention.

Hereinafter, a method for extracting image depth information by a stereo camera according to the present invention will be described in detail with the accompanying drawings.

3 is a flowchart of an embodiment for explaining a method for extracting image depth information by a stereo camera according to the present invention.

First, a first image and a second image of one object are respectively photographed using a stereo camera (step 100). A stereo camera refers to a camera that photographs the same object with two shooting lenses spaced apart at regular intervals. Two shooting lenses are used to capture the first and second images of the same object.

After operation 100, image distortion is corrected for each of the acquired first image and the second image (step 102). When the image is input through the two photographing lenses, the input image distortion caused by the camera, such as lens distortion of the camera, camera distortion, and incident angle of light, is corrected and brightness information is corrected in each of the first and second images. Depth information extraction using a stereo camera can be calculated through the difference between the positions of the two images, but the image input from the camera is obtained by reflecting the lens distortion, the shape of the CCD inside the camera, and the antiparallelity of the left and right cameras. .

If such image is used as it is, problems of accurate depth information calculation occur, so it is possible to obtain image correction parameters for each camera by acquiring several images that are easy to calculate mathematically through certain feature points such as chessboard.

4 is an equation for explaining a method for correcting image distortion and a reference diagram corresponding thereto. As shown in Fig. 4, the equation is an equation that clearly shows an example of the geometric transformation. After extending this concept to three-dimensional space, Equation 1 below can be used to fix the line at infinity.

According to Equation 1 above, by applying the parameter value to calculate the value obtained in the actual image, the infinite line of the input image is parallel and the Z-axis of the two camera images is parallelly corrected.

In addition, in the case of lens distortion, distortion occurs in the peripheral part depending on the lens, and convex or concave images are output. 5 is a reference diagram for describing image distortion due to lens distortion. As shown in Fig. 5, radial distortion is caused by spherical aberration of the lens, and this kind of distortion is symmetric about the optical axis of the lens. Therefore, in this distortion, the following equation (2) can be used to correct the curve as an original straight line when the shape is curved around the optical center.

It is possible to reconstruct an image before distortion by obtaining each coefficient through Equation 2. In addition, since the incident amount of light is reduced in a radial shape, brightness information is also restored to the image before distortion through shading correction.

After operation 102, the horizontal line of each of the first image and the second image whose image distortion is corrected is corrected (step 104). 6 is a reference diagram for explaining image correction due to a mismatch between horizontal lines of first and second images. As shown in FIG. 6, horizontal lines of the two images are coincident to facilitate extraction of depth information from the corrected first and second images, respectively. Problems that may occur in the first image A and the second image B may cause a difference in magnification due to the focal length of the two cameras and a mismatch of the reference line of the image when the same object is reflected. Here, adjust so that the center axes of the two images are parallel.

After operation 104, the size of the image is adjusted using a plurality of scalers for each of the captured first image and the second image (step 106). In this case, in the step of adjusting the size of the image using the plurality of scalers, the size of the image is adjusted to n / m size, for example, one-to-one, 1/2, 1/4, and 1/8. You can adjust the size to shrink. A plurality of scalers are used to reduce and adjust the size of each of the first and second images to one-to-one, 1/2, 1/4, and 1/8, and the like, so that the plurality of first and second images corresponding to the respective scalers are adjusted. Output them. Since the horizontal reference lines coincide with each other in the first and second input images, the same horizontal line may be searched to find the same point by searching the same horizontal line.

After operation 106, image depth information corresponding to the plurality of scalers is generated using the first image and the second image whose image size is adjusted by the plurality of scalers (step 108).

7 is an equation and a reference diagram for describing generation of image depth information using a horizontal distance difference between two objects. As shown in FIG. 7, Z in the equation means image depth information according to horizontal distance differences with respect to the first and second images adjusted by the scaler.

It is important to note that the feature of the image is that the closer the distance is, the greater the proportion of the entire line. For example, at 1 [m] distance 10 [cm] is 1/10, at 2 [m] distance 1/20, at 4 [m] distance 1/40 and at 1/80 distance 1/80 In this case, the number of pixels represented in an image having a horizontal line of 640 pixels is 64 pixels at 1 [m], 32 pixels at 2 [m] distance, 16 pixels at 4 [m] distance, and 8 at 8 [m] distance. As the pixel becomes, the number of pixels along the distance becomes the shape of the monotonic reduction function, as shown in FIG.

An example of a pixel and a distance in the graph of FIG. 8 may be represented by Table 1 below.

Pixel difference 128 64 32 16 8 4 2 One 0 Distance (m) 0.5 One 2 4 8 16 32 64 infinity

It can be seen from Table 1 that the closer the distance is, the wider the area between each distance. For example, the values between 16 [m] and 32 [m] and the values between 2 [m] and 4 [m] are shown in Tables 2 and 3 below.

Pixel difference 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 Distance (m) 2 2.06 2.13 2.2 2.28 2.37 2.46 2.56 2.67 2.78 2.9 3.04 3.2 3.37 3.55 3.77

Pixel difference 4 3 Distance (m) 16 21

As illustrated in Tables 2 and 3, the closer the distance to the object is, the more precisely the distance information is displayed. Due to this feature, a problem arises in that unnecessary information is displayed in detail in order to express far information in detail. In other words, as the distance between the camera and the object increases, the difference between the two images decreases and increases as the distance between the two objects decreases. If you use a scaler, you can expect a reduction in computation.

In the above example, to calculate distance information from 0.5 [m], 128 pixels must be compared. At this time, suppose that the configuration using a plurality of scalers proposed in the present invention. The scaler is assumed to use 1/4 and 1/1 scalers.

Using a 1/1 scaler and calculating the distance using 32 pixels in an image having a total horizontal resolution of 640 pixels, it is shown in Table 4 below.

Pixel difference 32 16 8 4 2 One 0 Distance (m) 2 4 8 16 32 64 infinity

On the other hand, when using a 1/4 scaler, the total horizontal resolution is 160 pixels, and the distances and pixel differences are shown in Table 5 below.

Pixel difference 32 16 8 4 2 One 0 Distance (m) 0.5 One 2 4 8 16 infinity

As shown in Table 5, even searching up to 32 pixels enables the information of the object at the shortest distance of 0.5 [m].

In the above example, depth information of 2 [m] to 64 [m] may be obtained as depth information corresponding to a 1/1 scaler. Further, depth information from 0.5 [m] to 16 [m] can be obtained as depth information corresponding to a quarter scaler.

By integrating the two pieces of information thus obtained, depth information from 0.5 [m] to 64 [m] can be obtained. In this case, the amount of calculation is reduced to 25 [%] because 32 pixels are searched from the previous 128 pixels. However, since two depth information detection modules corresponding to 1/1 and 1/4 scalers are required, the depth information can be calculated by performing a total of 50 [%] by performing 25 [%] * 2.

Meanwhile, in order to generate image depth information using the first image and the second image of which the size of the image is adjusted, the two input images are compared to find a point where the singularity matches. Although there are several processes in this method, as shown in FIG. 9, when using a dynamic programming method, distance can be detected using only one line for each image when calculating the distance of one line in a search process. In this method, the distance information is extracted through the dynamic programming method. In the dynamic programming method, the depth information due to the difference between two images can be obtained by obtaining a difference of each pixel in a line of two images and following a path that makes the smallest difference between pixels. In the present invention, a dynamic programming method is taken as an example, but other types of searching methods using lines, scanline optimization methods, and the like may also be used.

Thereafter, after generating the image depth information corresponding to the plurality of scalers using the first image and the second image, respectively, multiply the inverse of the scale ratio of the plurality of scalers and correct the image depth information. do. Since the extracted depth information is an image size adjusted by the scaler, the image depth information is corrected by multiplying the scaled ratio by the inverse to correct it to the original image size.

After operation 108, the optimal image depth information is extracted from the generated image depth information (operation 110). When the image depth information is greater than or equal to a predetermined distance, image depth information corresponding to a scaler having a relatively large scale ratio among the plurality of scalers is extracted. When the image depth information is less than or equal to a predetermined short distance, the scale of the plurality of scalers Image depth information corresponding to a scaler having a relatively small ratio is extracted.

10 is a flowchart of an exemplary embodiment for explaining a process of extracting optimal image depth information. For example, image depth information A obtained after the first and second images are scaled through the 1/1 scaler and information of image depth information B obtained after the first and second images are scaled through the 1/4 scaler. The range of is different. The image depth information A is input (step 200), and it is determined whether the image depth information A exceeds the maximum value Max B that the image depth information B can have (step 202). If the image depth information A exceeds the maximum value Max B that the image depth information B can have, A is extracted as the image depth information (step 204). The fact that the image depth information A exceeds the maximum value Max B that the image depth information B can have means that the object is at a distance. In this case, a scaler having a relatively large scale ratio among the plurality of scalers, that is, 1 / Image depth information A corresponding to one scaler is extracted. However, if the image depth information A does not exceed the maximum value Max B that the image depth information B can have, the image depth information B is input (step 206), and the image depth information B may have the image depth information B. It is determined whether the minimum value Min A is less than step S208. If the image depth information B is less than the minimum value Min A that the image depth information A can have, B is extracted as the image depth information (step 210). If the image depth information B is less than the minimum value Min A that the image depth information A can have, it means that the object is located below a certain short distance. In this case, a scaler having a relatively small scale ratio among the plurality of scalers, that is, 1 here Extract image depth information B corresponding to the / 4 scaler. On the other hand, if the image depth information B is not less than the minimum value Min A that the image depth information A can have, it is determined whether the absolute value of the difference between the image depth information A and the image depth information B is within a predetermined threshold (Threshold). Step 212). Here, the predetermined threshold value corresponds to the error threshold range of the image depth information. If the absolute value of the difference between the image depth information A and the image depth information B falls within a predetermined threshold value, A is extracted as the image depth information (step 214). However, if the absolute value of the difference between the image depth information A and the image depth information B does not fall within a predetermined threshold, since no meaningful value exists as the image depth information, error information is output (216). step).

Hereinafter, an apparatus for extracting image depth information by a stereo camera according to the present invention will be described in detail with the accompanying drawings.

FIG. 11 is a block diagram illustrating an image depth information extracting apparatus using a stereo camera according to the present invention. The stereo camera 300, the first image correcting unit 310, and the second image correcting unit 312 are illustrated in FIG. , Scaler 1-1 (320), scaler 1-2 (322), scaler 2-1 (330), scaler 2-2 (332), depth information generator 1 340, depth information generator 2 342 And a depth information extractor 350.

The stereo camera 300 uses the first camera lens 302 and the second camera lens 304 to photograph the first image and the second image of the object, respectively, and captures the photographed first image by the first image corrector. The image is output to the 310 and the photographed second image is output to the second image corrector 312.

The first image corrector 310 and the second image corrector 312 respectively correct image distortion with respect to each of the first image and the second image transmitted from the stereo camera 300, and scale the corrected result. Output to The first image corrector 310 and the second image corrector 312 calculate values obtained from the real image with respect to the distorted image by using Equation 1 described above, such that the depth axes of the input image are parallel. Correct.

In addition, as illustrated in FIG. 5, the first image corrector 310 and the second image corrector 312 may have the radial distortion described above with respect to the distortion generated by the spherical aberration of the lens. Equation (2) is used to reconstruct the image before the distortion for the curved shape around the optical center. At this time, since the incident amount of light is reduced in a radial shape, brightness information is also restored to the image before distortion through shading correction.

In addition, the first image corrector 310 and the second image corrector 312 correct horizontal lines with respect to the first image and the second image in which the image distortion is corrected. As illustrated in FIG. 6, the first image corrector 310 and the second image corrector 312 respectively adjust horizontal lines of the two images to facilitate extraction of depth information from the corrected first and second images. Match. That is, the first image corrector 310 and the second image corrector 312 adjust the center axis of the two images to facilitate extraction of depth information.

The scalers 1-1 320 and the scalers 2-1 330 correspond to a 1/1 scaler, that is, a scaler that maintains the size as an image corresponding to a one-to-one ratio with respect to the original image. On the other hand, the scalers 1-2 (322) and the scalers 2-2 (332) adjust to reduce the size of the image by 1/2, 1/4, 1/8, and the like. Although four scalers are shown in FIG. 11, the number of scalers can be increased or decreased according to the designer's choice. Here, the scalers 1-1 320 and the scalers 2-1 330 illustrate scaling to a 1/4 scaler, that is, an image that is reduced by a ratio of 1/4 to the original image.

The scaler 1-1 320 outputs an image having a size corresponding to a one-to-one ratio with respect to the first image to the depth information generator 1 340 as it is. Meanwhile, the scalers 1-2 322 output the reduced image by the ratio of 1/4 to the depth information generator 2 342 with respect to the first image.

In addition, the scaler 2-1 330 outputs the image having the size corresponding to the one-to-one ratio with respect to the second image to the depth information generator 1 340 as it is. Meanwhile, the scaler 2-2 332 outputs the reduced image by the ratio of 1/4 to the depth information generator 2 342 with respect to the second image.

The depth information generator 1 340 uses the first image provided from the scaler 1-1 320 and the scaled second image provided from the scaler 2-1 330 to determine an image corresponding to the size of the original image one-to-one. Create image depth information. In addition, the depth information generator 2 342 may use the first image provided from the scalers 1-2 322 and the scaled second image provided from the scalers 2-2 332 to be 1/4 of the size of the original image. Generates image depth information for the reduced image. The depth information generator 1 340 and the depth information generator 2 342 generate and generate the image depth information Z according to the horizontal distance difference with respect to the first and second images by using the above-described formula shown in FIG. 7. Each depth information is output to the depth information extracting unit 350.

The depth information generator 1 340 and the depth information generator 2 342 compare the two input images to generate image depth information, respectively, by using the first image and the second image of which the image size is adjusted. To find out where the singularities match. As shown in FIG. 9, when the depth information generator 1 340 and the depth information generator 2 342 use a dynamic programming method, the depth information generator 1 340 and the depth information generator 2 342 calculate the distance of one line in the search process. Since distance detection is possible with only one line per image, the method extracts distance information through a dynamic programming method. In addition, the depth information generator 1 340 and the depth information generator 2 342 may use a method such as Scanline Optimazation in addition to the dynamic programming method.

Thereafter, after the depth information generator 1 340 and the depth information generator 2 342 generate the image depth information corresponding to the plurality of scalers using the first image and the second image, respectively, The image depth information is corrected by multiplying the inverse of the scale ratio of the plurality of scalers. Since the extracted depth information is an image size adjusted by the scaler, the image depth information is corrected by multiplying the scaled ratio by the inverse to correct it to the original image size.

In the present invention, the depth information generating unit 1 340 and two depth information generating unit 2 342 are illustrated, but as the number of scalers increases or decreases, the depth information generating unit corresponds to the number of depth information generating units. You can also increase or decrease.

The depth information extractor 350 extracts optimal image depth information from the image depth information generated by the depth information generator 1 340 and the depth information generator 2 342. The depth information extractor 350 extracts image depth information corresponding to a scaler having a relatively large scale ratio among the plurality of scalers when the image depth information is greater than or equal to a predetermined distance, and when the image depth information is less than or equal to a predetermined short distance. Extracts image depth information corresponding to a scaler having a relatively small scale ratio among the plurality of scalers. As illustrated in FIG. 10, the depth information extractor 350 extracts A as image depth information if the image depth information A exceeds the maximum value Max B that the image depth information B can have, and the image depth information B If B is less than the minimum value Min A that the image depth information A can have, B is extracted as the image depth information. In addition, the depth information extractor 350 determines whether the absolute value of the difference between the image depth information A and the image depth information B is within a predetermined threshold, and the difference between the image depth information A and the image depth information B is determined. If the absolute value of is within a certain threshold, A is extracted as the image depth information. Also, if the absolute value of the difference between the image depth information A and the image depth information B does not fall within a predetermined threshold value, the depth information extractor 350 does not have a meaningful value as the image depth information. Print error information.

Meanwhile, the method inventions of the present invention described above can be implemented as computer readable codes / instructions / programs. For example, it may be implemented in a general-purpose digital computer for operating the code / instructions / program using a computer-readable recording medium. The computer-readable recording medium includes storage media such as magnetic storage media (eg, ROM, floppy disk, hard disk, magnetic tape, etc.), optical reading media (eg, CD-ROM, DVD, etc.) .

The method and apparatus for extracting image depth information by a stereo camera according to the present invention have been described with reference to the embodiment shown in the drawings for clarity, but this is merely an example, and those skilled in the art will appreciate It will be appreciated that variations and other equivalent embodiments are possible. Accordingly, the true scope of the present invention should be determined by the appended claims.

300: stereo camera
310: first image correction unit
312: second image correction unit
320: scaler 1-1
322: scaler 1-2
330: scaler 2-1
332: scaler 2-2
340: depth information generator 1
342: depth information generator 2
350: depth information extraction unit

Claims (14)

Photographing a first image and a second image of the object using a stereo camera, respectively;
Adjusting a size of an image by using a plurality of scalers for each of the captured first image and the second image;
Generating image depth information corresponding to the plurality of scalers, respectively, by using the first image and the second image whose size of the image is adjusted by the plurality of scalers; And
And extracting the optimum image depth information from the generated image depth information. The method of claim 1, wherein the image depth information extraction method by the stereo camera is performed.
The method may further include correcting image distortion for each of the captured first and second images.
The image depth information extraction method by the stereo camera, characterized in that for adjusting the size of the image for the corrected first and second images. The method of claim 2, wherein the correcting of the image distortion
The image depth information extraction method of the stereo camera, characterized in that for correcting the brightness information for each of the first image and the second image. The method of claim 1, wherein the image depth information extraction method by the stereo camera is performed.
Correcting a horizontal line with respect to each of the captured first image and the second image;
The image depth information extraction method by the stereo camera, characterized in that for adjusting the size of the image for the corrected first and second images. The method of claim 1, wherein adjusting the size of the image using the plurality of scalers
Image depth information extraction method by a stereo camera, characterized in that to reduce the size of the image. The method of claim 1, wherein generating image depth information corresponding to the plurality of scalers, respectively,
And multiplying the inverse of the scale ratios of the plurality of scalers, thereby correcting the image depth information. The method of claim 1, wherein the extracting optimal image depth information from the generated image depth information comprises:
When the image depth information is greater than or equal to a predetermined distance, image depth information corresponding to a scaler that is greater than or equal to a first scale ratio among the plurality of scalers is extracted. When the image depth information is less than or equal to a predetermined short distance, a second of the plurality of scalers is extracted. And extracting image depth information corresponding to a scaler that is less than or equal to a scale ratio. A stereo camera for capturing the first image and the second image with respect to the object;
A plurality of scalers for adjusting a size of an image for each of the captured first image and the second image;
A plurality of depth information generators respectively generating image depth information corresponding to the plurality of scalers by using the first image and the second image whose size of the image is adjusted by the plurality of scalers; And
And a depth information extracting unit for extracting optimal image depth information from the generated image depth information. The apparatus of claim 8, wherein the image depth information extraction apparatus by the stereo camera is used.
And a first and second image correcting unit configured to correct image distortion with respect to each of the captured first image and the second image. The display apparatus of claim 9, wherein the first and second image correctors are provided.
Image depth information extraction apparatus by a stereo camera, characterized in that for correcting the brightness information for each of the first image and the second image photographed. The display apparatus of claim 9, wherein the first and second image correctors are provided.
Image depth information extraction apparatus by a stereo camera, characterized in that for correcting the horizontal line for each of the first image and the second image photographed. The method of claim 8, wherein the plurality of scalers
Image depth information extraction apparatus by a stereo camera, characterized in that for reducing the size of the image. The method of claim 8, wherein the plurality of depth information generation unit
And extracting the depth information by multiplying the inverse of the scale ratios of the plurality of scalers to correct the depth information. The method of claim 8, wherein the depth information extracting unit
When the image depth information is greater than or equal to a predetermined distance, image depth information corresponding to a scaler that is greater than or equal to a first scale ratio among the plurality of scalers is extracted. When the image depth information is less than or equal to a predetermined short distance, a second of the plurality of scalers is extracted. An apparatus for extracting image depth information by a stereo camera, characterized in that extracting image depth information corresponding to a scaler that is equal to or less than a scale ratio.

Images