KR100911814B1 - Stereo Image Matching Error Elimination Device and Removal Method Using the Same - Google Patents

Abstract

The present invention relates to a stereo image matching error cancellation apparatus and a method of eliminating the same, and more particularly, to any one of pixels included in a stereo image as a center pixel, and the center pixel and a plurality of peripheral pixels adjacent to the center pixel. An image processor for generating an included window; A difference value calculator for calculating a difference value between the pixel value of the center pixel and the pixel value of the peripheral pixel according to a color in the window; A comparison unit for dividing the peripheral pixel into a first peripheral pixel having the difference value less than or equal to the comparison reference value and a second peripheral pixel having the difference value exceeding the comparison reference value; A summing unit for summing up any one of the first and second peripheral pixels; And an output determination unit that compares a predetermined number of reference pixels with one of the first and second peripheral pixels, and determines whether the center pixel is variable according to a comparison result. Accordingly, the present invention eliminates the matching error of the stereo image by generating a window in the stereo image and determining whether to output the center pixel invariably according to the difference in the pixel value according to the color between the center pixel and the peripheral pixel of the window. Provided are a stereo image matching error elimination device and a elimination method using the same.

Description

Stereo Image Matching Error Elimination Device and Removal Method Using It {STEREO-IMAGE MATCHING ERROR REMOVAL APPARATUS AND REMOVAL METHORD USING THE SAME}

The present invention can eliminate a matching error of a stereo image by generating a window in the stereo image and determining whether to output the center pixel invariably according to the difference in pixel values according to the color between the center pixel and the peripheral pixel of the window. The present invention relates to a stereo image matching error cancellation device and a removal method using the same.

In general, a stereo image refers to a stereoscopic image in which two images input from two cameras are matched. The basic principle is to find out where the pattern at one location in one image is located in the other image, in the same way that two people obtain distance information to a particular object or image pattern. In other words, the image matched by the binocular difference will have a stereoscopic.

Accordingly, since the stereo image can obtain distance information of all objects in the image, the stereo image can be applied to various fields as compared to the conventional art using only an image captured by one camera.

In particular, it is already used in various fields such as car number recognition, car collision avoidance devices, games, and intelligent robot vision, or it is expected to be applied in the future. In addition, new fields that require distance information may occur. have.

As shown in FIG. 10, the stereo image is a distance information image indicating a distance between a camera and an object by displaying a difference of gray levels, which are pixel values, according to a distance of an object located in the image.

Since stereo images may not accurately identify objects or patterns in some cases due to noise (hereinafter, referred to as a “matching error”) as shown in FIG. 10C, it is difficult to accurately determine the distance difference between objects.

Matching errors in stereo images are caused by various causes.

One of the causes is that the characteristics of the two cameras do not match completely. The reality is that the characteristics of the digital image sensor and lens within the camera are not the same for all cameras. This means that certain objects are taken differently by the two cameras.

The other is that the amount of light input is not the same because the two cameras are not in the same position. In particular, when an object reflects light at a certain angle, the images of the two cameras may be significantly different.

As such, the matching error is necessarily generated because a cause that occurs from the time when the image is taken by the camera is provided. Therefore, a method of eliminating a matching error has been required to extract more accurate stereo images.

Accordingly, the present invention creates a window in the stereo image and determines whether to output the center pixel invariably according to the difference in pixel values between the center pixel and the surrounding pixels of the window, thereby eliminating a matching error of the stereo image. It is an object of the present invention to provide a matching error elimination device and a removal method using the same.

According to an aspect of the present invention, there is provided an image processing unit configured to generate a window including one of pixels included in a stereo image as a central pixel and a plurality of peripheral pixels adjacent to the central pixel and the central pixel; A difference value calculator for calculating a difference value between the pixel value of the center pixel and the pixel value of the peripheral pixel according to a color in the window; A comparison unit for dividing the peripheral pixel into a first peripheral pixel having the difference value less than or equal to the comparison reference value and a second peripheral pixel having the difference value exceeding the comparison reference value; A summation unit which counts any one of the first and second peripheral pixels; And an output determination unit for comparing a preset reference pixel number with a counted number of any one of the counted first and second peripheral pixels, and determining whether the pixel value of the center pixel varies according to a comparison result. It is achieved by a stereo image matching error elimination device characterized in that.

The summation unit counts the number of the first peripheral pixels, the pixel value is provided with a gray level, and the output determining unit determines whether the center pixel is variable according to a comparison result. Compare the counted number with the reference pixel number, and output the center pixel unchanged when the counted number of the first peripheral pixels is equal to or greater than the reference pixel number; The center pixel may be regarded as a matching error, and the gray level of the center pixel may be lowest and output.

The summation unit counts the number of the second peripheral pixels, the pixel value is provided at a gray level, and the output determining unit determines whether the center pixel is variable according to a comparison result. Compares the counted number with the reference pixel number, outputs the center pixel unchanged when the counted number of the first peripheral pixels is less than the reference pixel number, and outputs the center pixel without change; The center pixel may be regarded as a matching error, and the gray level of the center pixel may be lowest and output.

In addition, after the window is generated by the image processor, the gray level and coordinates of the center pixel and the peripheral pixel are determined during the time when the window is determined whether the window is output through the difference value calculator, the comparator, or the adder. The apparatus may further include a storage configured to store corresponding pixel data, and the output determiner may output the center pixel based on the pixel data stored in the storage of the center pixel of the window in which the output is determined.

The storage unit may include: a first storage unit storing the pixel data for a time period during which the window generated by the image processor is calculated by the difference value calculator; A second storage unit storing the pixel data stored in the first storage unit for a time when the peripheral pixel of the window is divided into the first peripheral pixel and the second peripheral pixel by the comparing unit; And a third storage unit configured to store the pixel data stored in the second storage unit during a time when the number of the first peripheral pixels is counted by the summation unit, and the output determination unit stores the pixels stored in the third storage unit. Based on the data, it is possible to cause the center pixel for which the output is determined to be arranged at the coordinates of the center pixel and to output the center pixel with the gray level.

In addition, the adding unit counting the number of the first peripheral pixels may be divided into a plurality of first clusters in which the pixels of the window are arranged in rows, and counting the number of the first peripheral pixels in each of the first clusters. And extracting the number of the pixels included in each row by counting the counted number of the first peripheral pixels between the first clusters, and arranging the number of the first peripheral pixels included in the row in columns. And dividing into two second clusters, counting the number of the first peripheral pixels in the second cluster, and counting the sum of the sum of the first peripheral pixels between the second clusters.

The output determination unit may output the center pixel unchanged when the pixel value is the same between the center pixel and the corner pixel positioned at the corner of the window based on the pixel data.

Meanwhile, in the stereo image matching error elimination method according to the present invention, a center pixel is used as one of the pixels included in a stereo image, and the window includes the center pixel and a plurality of peripheral pixels adjacent to the center pixel. Generating a; Calculating a difference value between the pixel value of the center pixel and the pixel value of the peripheral pixel according to a color in the window; Setting a comparison reference value and comparing the comparison reference value with the difference value; Setting the peripheral pixel having the difference value equal to or less than the comparison reference value as a first peripheral pixel, and setting the peripheral pixel having the difference value exceeding the comparison reference value as a second peripheral pixel; Counting any one of the first peripheral pixel and the second peripheral pixel; Setting a reference pixel number and comparing the reference pixel number with a counted number of any one of the first peripheral pixel and the second peripheral pixel; And determining whether the pixel value of the center pixel is variable according to a result of comparing the number of the reference pixels with the counted number of any one of the first peripheral pixel and the second peripheral pixel. It is also achieved by a method of eliminating matching errors.

The counting of any one of the first peripheral pixel and the second peripheral pixel is a step of counting the number of the first peripheral pixel, the pixel value is provided as a gray level, and the reference pixel number. And comparing the counted number of any one of the first peripheral pixel and the second peripheral pixel is a step of comparing the number of the first peripheral pixel and the reference pixel number, and whether the pixel value of the center pixel is variable. The determining may include outputting the center pixel unchanged when the number of the first peripheral pixels is greater than or equal to the reference pixel, and when the number of the first peripheral pixels is less than the reference pixel, the center pixel is regarded as a matching error. The gray level of the center pixel is the lowest and can be output.

The counting of any one of the first peripheral pixel and the second peripheral pixel is a step of counting the number of the second peripheral pixel, the pixel value being provided at gray level, and the reference pixel number. And comparing the counted number of any one of the first peripheral pixel and the second peripheral pixel is a step of comparing the number of the second peripheral pixel and the reference pixel number, and whether the pixel value of the center pixel is variable. The determining may include outputting the center pixel unchanged when the number of the second peripheral pixels is less than the number of reference pixels, and when the number of the first peripheral pixels is greater than or equal to the reference pixel number, the center pixel is regarded as a matching error. The gray level of the center pixel is the lowest and can be output.

On the other hand, the pixel corresponding to the pixel value and coordinates of the center pixel and the peripheral pixel during the time from the step of generating the window to counting any one of the first peripheral pixel and the second peripheral pixel And storing the data, wherein the center pixel is invariably output by determining one of the output of the center pixel unchanged and the output of the gray level of the center pixel being lowest. In this case, the center pixel is output based on the stored pixel data, and when the center pixel is regarded as a matching error, the center pixel is based on the data corresponding to the coordinates of the center pixel stored and the gray level becomes the lowest. Can be output.

The storing of the pixel data may include providing a storage unit including a first storage unit, a second storage unit, and a third storage unit in which the pixel data is stored; Storing the pixel data in the first storage unit for a time of calculating a difference value of the gray level based on the generated window and comparing the difference value with the comparison reference value; Storing the pixel data stored in the first storage unit in the second storage unit during a time of setting the first peripheral pixel and the second peripheral pixel; Counting the number of any one of the first peripheral pixel and the second peripheral pixel, and comparing the number of the reference pixel number and the number of the first peripheral pixel counted to the second storage unit during the time The stored pixel data may be stored in the third storage unit.

The counting of the number of the first peripheral pixels may include dividing the pixels of the window into a plurality of first clusters arranged in rows; Counting the number of the first peripheral pixels in each of the first clusters; Counting the counted number of the first peripheral pixels between the first clusters and extracting the number of the first peripheral pixels included in each row; Dividing the number of pixels included in the row into a plurality of second clusters arranged in columns; Counting the number of the first peripheral pixels in the second cluster; And counting the number of the first peripheral pixels counted between the second clusters.

In addition, the determining of whether the center pixel is continuously output or the gray level of the center pixel is output as the lowest is based on the pixel data and is located at the corner of the center pixel and the window. When the pixel values are the same between the edge pixels, the center pixel may be output without change.

The stereo image matching error elimination apparatus and the elimination method using the same according to the present invention generate a window in the stereo image, and change the corresponding center pixel according to the gray level difference, which is the pixel value between the center pixel and the surrounding pixels of the window. By determining whether to output, there is an excellent effect that can provide a more accurate stereo image by eliminating the matching error of the stereo image.

In addition, the stereo image matching error elimination apparatus and the elimination method using the same according to the present invention is stored in the gray level data of the center pixel of the window continuously generated in the stereo image, so that the stored gray level is applied to the output center pixel In addition, the gray level of the center pixel has an excellent effect of obtaining an undamaged stereo image.

In addition, the stereo image matching error elimination apparatus and the elimination method using the same according to the present invention is stored in the coordinates of the center pixel of the window continuously generated in the stereo image, the stored coordinates are applied to the output center pixel, Even if each step of successively performing each step of determining whether or not is sequentially performed, the window is output to the original coordinates after a certain time has an excellent effect of obtaining a stereo image without the coordinates moved.

In addition, the stereo image matching error elimination apparatus and the elimination method using the same according to the present invention, since the summing up a plurality of clusters when summing the first peripheral pixels, the processing time is shortened and requires a stereo image in which matching errors are removed in real time. It can be applied in various fields such as intelligent robots and industrial automation devices.

Hereinafter, a stereo image matching error elimination apparatus and a elimination method using the same will be described with reference to the accompanying drawings.

As shown in FIG. 1, the apparatus 1 for canceling stereo image matching errors according to the present invention includes an image processor 10, a difference value calculator 20, a comparator 30, an adder 40, and a storage 60. ) And an output determination unit 70.

The image processor 10 generates a window in the stereo image A including the matching error. Here, the window is composed of a plurality of peripheral pixels adjacent to the center pixel by using any one of the pixels included in the stereo image as the center pixel. As shown in FIG. 12, the center pixel of the window 70 is sequentially selected from the pixels disposed on the left side of the stereo image A in the right row direction, and the rightmost pixel becomes the center pixel. Afterwards, the center pixel is selected from the left most of the pixels listed in the layer immediately below. Here, the window may be provided as an N × N window having the same number of pixels arranged in rows and columns, and the description is an 11 × 11 window. However, the window is not limited to the present invention.

3, when the window 70a is an 11X11 window composed of 121 pixels, 11X11 when the pixel disposed at the position (6,6) in the stereo image A is selected as the center pixel. Windows 70a are formed.

The difference calculator 20 calculates a difference between the pixel value of the center pixel and the pixel value of the peripheral pixel in the window. Here, the pixel value may be provided as a gray level.

For example, as shown in FIG. 3, when the gray level of the center pixel 71a of the window 70a is 150 and the gray level of the peripheral pixels 72a, 73a, 74a is 150, 140, 10, each peripheral pixel 72a is shown. The gray level difference between (73a, 74a) and the center pixel (71a) is 0, 10, 140. Accordingly, the difference values corresponding to the peripheral pixels 72a, 73a, and 74a are 0, 10, and 140, respectively.

The comparison unit 30 compares the reference value with the difference value, and based on the comparison result, the peripheral pixels 72a, 73a and 74a are shown in FIG. 4, and the first peripheral pixels 72a and 73a are formed. 2 is divided into peripheral pixels 74a. Here, the reference value is a second peripheral pixel 74a having a gray level not similar to the first peripheral pixels 72a and 73a having a gray level similar to the gray level of the center pixel 71a among the pixels in the window 70a. Provides criteria separated by.

For example, when the reference value is set to a gray level of 15, as shown in FIG. 4, the peripheral pixels 72a and 73a having a difference value of 0 and 10 smaller than the reference value are gray levels of the center pixel 71a. Recognized as having a gray level similar to the first peripheral pixels (72a, 73a) and can be given any value. In the present invention, an arbitrary value assigned to the first peripheral pixels 72a and 73a is '1' as an example.

In addition, when the reference value is set to a gray level of 15, the peripheral pixel 74a having the difference value 140 larger than the reference value is recognized as having a gray level that is not similar to the gray level of the center pixel 71a. It can be divided into two peripheral pixels 74a and given an arbitrary value. In the present invention, an arbitrary value given to the second peripheral pixel 74a is '0' as an example.

The adder 40 counts the number of the first peripheral pixels 72a and 73a in the window.

Here, a method of counting the number of the first peripheral pixels 72a and 73a will be described with reference to FIG. 5. As shown in FIG. 5, pixels arranged in rows of the window 70a are divided into a plurality of first clusters 80a, 80b, and 80c. The number of pixels included in the first clusters 80a, 80b, and 80c may be arbitrarily set, and in this description, two first clusters 80a, 80b and three pixels including four pixels from the left in the drawing are illustrated. One example is divided into one first cluster (80c) that includes, and divided into a total of three first cluster (80a, 80b, 80c). In addition, the number of the first peripheral pixels 72a and 73a included in each of the first clusters 80a, 80b, and 80c is counted, and the first peripheral pixels (counted between the first clusters 80a, 80b and 80c) are counted. When the number of 72a and 73a is counted, the number of first peripheral pixels 72a and 73a included in the corresponding row in the window 70 is counted.

The number of first peripheral pixels 72a and 73a included in the corresponding row in the window 70a is divided into a plurality of second clusters 81a, 81b and 81c arranged in columns. The number of the first peripheral pixels 72a and 73a included in the corresponding row in the window 70 included in the second cluster may be arbitrarily set. In the present description, two pixels including four pixels from the upper side in the drawing are included. For example, the second clusters 81a and 81b and one second cluster 81c including three pixels are divided into three second clusters 81a, 81b and 81c. The number of first peripheral pixels 72a and 73a included in each of the second clusters 81a, 81b and 81c is counted, and the first peripheral pixels (counted between the second clusters 81a, 81b and 81c) are counted. When the number of 72a and 73a is counted, the number 82 of the corresponding total first peripheral pixels 72a and 73a in the window 70a is counted. Here, the number of second peripheral pixels can be counted in the same manner, and can be compared with the reference pixel number later.

The output determination unit 50 compares the number of reference pixels with the number of the first peripheral pixels summed by the adder 40, outputs the center pixel invariably according to the comparison result, and minimizes the gray level of the center pixel. One of the outputs is determined.

The reference pixel number means the number of pixels in which the first peripheral pixel should be disposed at least in the window, and may be arbitrarily set. For example, assuming that the number of reference pixels is 60, at least 60 first peripheral pixels must be arranged in the window so that the center pixel is invariably outputted. If the number is less than 60, the center pixel is regarded as a matching error and 0 is the lowest gray level. Is output.

That is, since the number of the first peripheral pixels 72a and 73a is 74, the window 70a shown in FIGS. 3 and 4 is larger than 60, which is the reference pixel number. In this case, the center pixel 71a shown in Figs. 2 and 3 is regarded as not a matching error and the gray level of each pixel is output unchanged.

However, in the window 70b shown in FIGS. 6 and 7, the number of first peripheral pixels 71b and 72b having a gray level difference less than or equal to the center pixel 71b is 47, and the center pixel 71b is shown. The number of second peripheral pixels 74b having a gray level difference greater than or equal to 73 is 73. In this case, since the number 47 of the first peripheral pixels 71b and 72b is less than the reference pixel number 60, the center pixel 71b is regarded as a matching error and is output as 0 having the lowest gray level. Here, the gray level is output as 0, which is the lowest, may be a background (black) of a general input image.

Here, in the present description, the number of first peripheral pixels included in the window is added and compared with the reference pixel number to determine whether the corresponding center pixel is a matching error. However, the number of second peripheral pixels in the window is added together. It can be provided to compare with the reference pixel number. In this case, the center pixel of the window where the number of the second peripheral pixels is less than the reference pixel number is not considered a matching error and is output unchanged. The center pixel whose number of the second peripheral pixels is greater than or equal to the reference pixel number is regarded as a matching error so that the gray level is reduced. Output is at the lowest 0 (black).

Here, each center pixel which is invariably output by the output determination unit 50 is stored in the storage unit 60 and is disposed at the coordinates of the center pixel included in the pixel data based on the pixel data provided, and the pixel data The gray level of the center pixel included in the output is output.

After the window is generated by the image processor 10, the storage unit 60 takes time to determine whether the window is output through the difference value calculator 20, the comparator 30, and the adder 40. Pixel data corresponding to the coordinates of the center pixel and the gray level of the center pixel are stored. Here, since the coordinates of the center pixel are stored in the window is sequentially generated by the image processing unit 10, the difference value calculation unit 20, the comparison unit 30, the adding unit 40 and the output determination unit 50 This is to prevent the center pixel from being output at the coordinates shifted from the window coordinates.

The storage unit 60 may include a first storage unit 61, a second storage unit 62, and a third storage unit 63.

The first storage unit 61 stores the pixel data during the time when the window generated by the image processor 10 calculates the difference value of the gray level between the center pixel and the peripheral pixel by the difference value calculator 20.

The second storage unit 62 stores pixel data stored in the first storage unit 61 during the time when the peripheral pixel of the window is divided into the first peripheral pixel and the second peripheral pixel by the comparator 30. .

The third storage unit 63 stores the pixel data that was stored in the second storage unit 62 during the time when the number of the first peripheral pixels is added by the adder 40.

Each storage unit 60 may be provided with a transmission unit (not shown) capable of transmitting pixel data after a time at which the corresponding window is calculated by the difference value operation unit 20, the comparison unit 30, and the summer unit 40. Can be.

In addition, as shown in FIGS. 8 and 9, the output determination unit 50 has a gray level between the center pixel 71c of the window 70c and the four corner pixels 72c positioned at the corners of the window 70c. If both are the same, the center pixel 71c can be output.

This means that if the gray levels are the same between the center pixel 71c and the four corner pixels 72c positioned at the corners of the window 70c, it is highly likely that all pixels included in the window 70c have mutually similar gray levels. Therefore, it is very likely that the center pixel 71c is not a matching error.

Through the above configuration, the stereo image A including the matching error C shown in FIG. 10 is the stereo image A ′ in which the matching error C is removed and the matching error C shown in FIG. 11 is removed. Is output.

Hereinafter, a stereo image matching error elimination method using the stereo image matching error elimination apparatus 1 according to the present invention will be described in detail with reference to FIG. 2.

When the stereo image A including the matching error is input (S1), a window is generated by the image processor 10 (S2).

In the difference value calculator 20, the difference value between the gray level of the center pixel of the generated window and the gray level of the peripheral pixel is calculated (S3). The pixel data is stored in the first storage unit 61 during this time t1 (S12).

The difference value of each peripheral pixel calculated by the difference value calculation unit 20 is compared with the comparison reference value (S4), and the peripheral pixel is divided into a first peripheral pixel and a second peripheral pixel according to the comparison result, and is set (S5). . During this time t2, the pixel data stored in the first storage unit 61 is stored in the second storage unit 62 (S13).

When the peripheral pixels are divided into the first peripheral pixel and the second peripheral pixel, the number of the first peripheral pixels is counted (summed) by the adder 40 (S6). During this time t3, the pixel data stored in the second storage unit 62 is stored in the third storage unit 63 (S14).

Then, the output determination unit 50 compares whether the gray level of the center pixel of the window and the gray level of the corner pixel among the peripheral pixels are the same (S7). For example, when the gray level of the center pixel of the window 70 and the gray level of the corner pixel among the peripheral pixels are the same, pixel data stored in the third storage unit 63 is reflected to output the center pixel without change (S9). In this case, the pixel data is reflected and the center pixel is constantly output and is disposed in the coordinates of the center pixel included in the pixel data based on the pixel data stored in the storage unit 60 and provided. Output is to have a gray level of the center pixel.

If the gray level of the center pixel of the window and the gray level of the corner pixel among the peripheral pixels are not the same, the output determination unit 50 compares the number of the first peripheral pixels in the window with the number of reference pixels (S8) to determine the first level. When the number of neighboring pixels is larger than the number of reference pixels, the output determination unit 50 reflects the pixel data stored in the third storage unit 63 so that the center pixel is continuously output (S9).

For example, when the number of the first peripheral pixels is smaller than the number of reference pixels, the output determination unit 50 applies data corresponding to the coordinates of the center pixel stored in the third storage unit 63 to the center pixel, but minimizes the gray level. To be output as 0 (S10).

Through the above process, the stereo image A is converted into a stereo image A 'from which the matching error C has been removed and output (S11).

In addition, the matching error (C) can be eliminated sequentially through the window sequentially generated. For example, it is assumed that a window in the stereo image A is provided with a total of three windows: a first window, a second window, and a third window, and each window is sequentially generated by the image processor 10.

In this case, as shown in FIG. 2, while the first window performs steps S3, S4, and S13 corresponding to time t2, the second window performs steps S3 and S12 corresponding to time t1.

In addition, while the first window performs steps S6 and S14 corresponding to time t3, the second window performs steps S4, S5 and S13 corresponding to time t2, and the third window corresponds to S3, corresponding to time t1. Follow step S12.

As described above, successively generating windows sequentially perform each step according to time, thereby reducing matching error calculation time. That is, as shown in FIG. 2, after the first window takes a time t1 to t3 and is output, a second window may be output after t3 time. Only t3 hours are output.

In addition, the coordinates and gray levels of the center pixel of the window are stored separately so that the center coordinates may be output from the original position when the center coordinates are output.

Although embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that the present embodiments may be modified without departing from the spirit or spirit of the invention. It is intended that the scope of the invention be defined by the claims appended hereto and their equivalents.

1 is a block diagram of a stereo image matching error cancellation apparatus according to the present invention,

2 is a flowchart of a stereo image matching error elimination method according to the present invention;

3 to 9 are views for explaining the division of the first peripheral pixel and the second peripheral pixel and counting the number of the first peripheral pixel in the window through the stereo image matching error elimination method according to the present invention,

10 is a diagram illustrating an example of a stereo image including a matching error.

11 is a diagram illustrating an example of a stereo image from which a matching error is eliminated through the stereo image matching error removing device according to the present invention;

12 is a diagram illustrating a situation in which a window is generated while moving in an image in the stereo image matching error elimination method according to the present invention.

Claims (14)

An image processor configured to use any one of pixels included in a stereo image as a center pixel and to generate a window including the center pixel and a plurality of peripheral pixels adjacent to the center pixel; A difference value calculator for calculating a difference value between the pixel value of the center pixel and the pixel value of the peripheral pixel according to a color in the window; A comparison unit for dividing the peripheral pixel into a first peripheral pixel having a difference value less than or equal to a comparison reference value and a second peripheral pixel having a difference value exceeding the comparison reference value; A summation unit which counts any one of the first and second peripheral pixels; An output determination unit for comparing the counted number of one of the first peripheral pixel and the second peripheral pixel with a preset reference pixel number, and determining whether the pixel value of the center pixel is variable according to the comparison result Stereo image matching error elimination device, characterized in that. The method of claim 1, The summation unit counts the number of the first peripheral pixels, The pixel value is provided in gray level, The output determining unit determines whether the center pixel is variable according to a comparison result. Compare the counted number of the first peripheral pixels with the reference pixel number, and output the center pixel unchanged when the counted number of the first peripheral pixels is equal to or greater than the reference pixel number, and the number of the first peripheral pixels And a gray level of the center pixel is minimized when the center pixel is regarded as a matching error when the number of reference pixels is less than the reference pixel. The method of claim 1, The summation unit counts the number of the second peripheral pixels, The pixel value is provided in gray level, The output determining unit determines whether the center pixel is variable according to a comparison result by comparing the counted number of the second peripheral pixels with the reference pixel number, and the counted number of the second peripheral pixels being less than the reference pixel number. And outputting the center pixel unchanged, and if the number of the second peripheral pixels is equal to or greater than the reference pixel number, the center pixel is regarded as a matching error and the gray level of the center pixel is outputted as the minimum. Matching Error Elimination Device. The method according to claim 2 or 3, After the window is generated by the image processor, the window corresponds to the gray level and coordinates of the center pixel and the peripheral pixel during the time when the window is determined whether the window is output through the difference calculator, the comparator, or the adder. Further comprising a storage unit for storing the pixel data, The output determination unit And a center pixel of the window determined to output the center pixel based on the pixel data stored in the storage unit. The method of claim 4, wherein The storage unit A first storage unit storing the pixel data for a time period during which the window generated by the image processor is calculated by the difference value calculator; A second storage unit storing the pixel data stored in the first storage unit for a time when the peripheral pixel of the window is divided into the first peripheral pixel and the second peripheral pixel by the comparing unit; And a third storage unit storing the pixel data stored in the second storage unit during a time when the number of the first peripheral pixels is counted by the summation unit. The output determination unit Based on the pixel data stored in the third storage unit, the center pixel determined to be output is arranged in the coordinates of the center pixel and the center pixel is output with the gray level. Removal device. The method of claim 2, The summator counts the number of the first peripheral pixels. The pixels of the window are divided into a plurality of first clusters arranged in a row, each number of the first peripheral pixels in each of the first clusters is counted, and the counts of the first peripheral pixels counted between the first clusters. The number of pixels included in each row is counted by counting the number, and the number of the first peripheral pixels included in the row is divided into a plurality of second clusters arranged in columns, and the first cluster in the second cluster is counted. And counting the number of neighboring pixels and counting the sum of the sum of the first neighboring pixels between the second clusters. The method of claim 1, The output determination unit And when the pixel values of the center pixel and the pixel values of the pixels positioned at the vertex of the window pixel are the same, determining to output the center pixel unchanged. In the stereo image matching error removal method, Generating a window including any one of pixels included in a stereo image as a central pixel and a plurality of peripheral pixels adjacent to the central pixel; Calculating a difference value between the pixel value of the center pixel and the pixel value of the peripheral pixel according to a color in the window; Setting a comparison reference value and comparing the comparison reference value with the difference value; Setting the peripheral pixel having the difference value equal to or less than the reference value to a first peripheral pixel and setting the peripheral pixel having the difference value exceeding the reference value to a second peripheral pixel; Counting any one of the first peripheral pixel and the second peripheral pixel; And comparing a counted number of one of the first peripheral pixel and the second peripheral pixel with a preset reference pixel number, and determining whether the pixel value of the center pixel is variable according to the result. How to eliminate image matching error. The method of claim 8, Counting the number of any one of the first peripheral pixel and the second peripheral pixel is a step of counting the number of the first peripheral pixel, The pixel value is provided in gray level, Comparing the number of the reference pixel number and any one of the counted first peripheral pixel and the second peripheral pixel is a step of comparing the number of the first peripheral pixel and the reference pixel number, Determining whether the pixel value of the center pixel is variable includes outputting the center pixel unchanged when the number of the first peripheral pixels is greater than or equal to the reference pixel, and when the number of the first peripheral pixels is less than the reference pixel number. And the pixel is regarded as a matching error, and the gray level of the center pixel is lowest and outputted. The method of claim 8, Counting the number of any one of the first peripheral pixel and the second peripheral pixel is a step of counting the number of the second peripheral pixel, The pixel value is provided in gray level, Comparing the number of the reference pixel and the number of the counted first peripheral pixel and the second peripheral pixel is a step of comparing the number of the second peripheral pixel and the reference pixel number, Determining whether the pixel value of the center pixel is variable may include outputting the center pixel unchanged when the number of the second peripheral pixels is less than the number of reference pixels, and outputting the center pixel without change. And the pixel is regarded as a matching error, and the gray level of the center pixel is lowest and outputted. The method of claim 9, Pixel data corresponding to pixel values and coordinates of the center pixel and the peripheral pixel is generated for the time required from the step of generating the window to counting any one of the first peripheral pixel and the second peripheral pixel. Further comprising the step of being stored, If the center pixel is constantly outputted by determining one of the output of the center pixel unchanged and the gray level of the center pixel being lowest and outputted, based on the stored pixel data When the center pixel is output and the center pixel is regarded as a matching error, the stereo image matching error is output based on the data corresponding to the coordinates of the center pixel stored and the gray level becomes the lowest. How to remove. The method of claim 11, The step of storing the pixel data Providing a storage unit including a first storage unit, a second storage unit, and a third storage unit in which the pixel data is stored; Storing the pixel data in the first storage unit for a time of calculating a difference value of the gray level based on the generated window and comparing the difference value with the comparison reference value;  Storing the pixel data stored in the first storage unit in the second storage unit during a time of setting the first peripheral pixel and the second peripheral pixel; Counting the number of any one of the first peripheral pixel and the second peripheral pixel, and comparing the number of the reference pixel number and the number of the first peripheral pixel counted to the second storage unit during the time And storing the stored pixel data in the third storage unit. The method of claim 9, Counting the number of the first peripheral pixel is Dividing the pixels of the window into a plurality of first clusters arranged in rows; Counting the number of the first peripheral pixels in each of the first clusters; Counting the counted number of the first peripheral pixels between the first clusters and extracting the number of the first peripheral pixels included in each row; Dividing the number of pixels included in the row into a plurality of second clusters arranged in columns; Counting the number of the first peripheral pixels in the second cluster; And counting the number of the counted first neighboring pixels between the second clusters. The method of claim 8, Before comparing the counted number of any one of the first peripheral pixel and the second peripheral pixel with a preset reference pixel number, And if the pixel values of the center pixel and the pixel values of pixels positioned at the vertex of the window pixel are the same, determining to output the center pixel invariably.

Images