KR100696167B1 - Bad pixel processing unit - Google Patents

Abstract

A dead pixel processing device is provided to output a pixel value of a 6Î6 structure corrected for a dead pixel. A component dividing unit(110) divides an inputted bayer pattern image into component-classified data. A dead pixel detecting unit(120) obtains a pattern of pixel values based on a median value of respective data from the component-classified data, compares the obtained pattern with a pattern of the other components, obtains a comparison value on the basis of respective components, compares the comparison value with a dead-pixel threshold value, judges whether a current pixel is a dead pixel, and outputs corrected component-classified data.

Description

Bad pixel processing unit {Method for processing dead pixel}

1 is a structural diagram of an embodiment of a bad pixel processing apparatus according to the present invention;

FIG. 2A illustrates an example of a 6 × 6 Bayer pattern image input to FIG. 1;

FIG. 2B is an exemplary view illustrating the separation of the Bayer pattern of FIG. 2A into components having a 3 × 3 structure; FIG.

3 is a detailed structural diagram of an embodiment of a bad pixel detection unit of FIG. 1;

4 is an exemplary diagram for describing an operation of a median detector, an absolute detector, and a comparison detector of FIG. 3;

5 is an exemplary diagram illustrating an image output by the bad pixel detection unit of FIG. 1;

6 is a detailed structural diagram of an embodiment of a hot pixel detection unit of FIG. 1;

Figure 7a is an exemplary view showing an original image input to the device of the present invention,

FIG. 7B illustrates an example of correcting by detecting a bad pixel and a hot pixel with respect to FIG. 7A; FIG.

<Explanation of symbols for the main parts of the drawings>

110: component separation unit 120: bad pixel detection unit

130: hot pixel detection unit

The present invention relates to a bad pixel processing apparatus, and more particularly to a bad pixel processing apparatus for use in an image sensor or the like.

Image sensor is composed of several pixels arranged in a two-dimensional structure, and each pixel converts them into electrical signals according to the brightness of incoming light. The amount of light can be known, and an image can be constructed using the pixel unit.

Such an image sensor is composed of a pixel array consisting of hundreds of thousands to millions of pixels, a device for converting analog data sensed by pixels into digital data, and hundreds to thousands of storage devices. Due to such a large number of devices, there is a possibility that an image sensor may have a process error. For this reason, a dead pixel or a hot pixel may be generated among the pixels. This is an important factor in determining the grade and price of a. Pixel data generated by bad pixels is characterized by being too large or vice versa smaller than pixel data generated by adjacent pixels.

In the conventional bad pixel processing method, when the difference exceeds a predetermined range compared with a reference pixel value, the bad pixel is defined as a bad pixel and corrected. However, such a conventional method uses a lot of hardware resources, and in the case of a dense image or a video having a large difference in brightness, a difference occurs between adjacent pixels. Therefore, even if it is not a bad pixel, the correction is treated as a bad pixel. As a result, the reconstructed image has a problem in that a lot of distortion occurs than the actual image.

The present invention has been proposed to solve the above problems, and an object thereof is to provide a bad pixel processing apparatus for determining whether a bad pixel or a hot pixel is compared by comparing a pattern of each component within a 6x6 mark. .

Another object of the present invention is to provide a bad pixel processing apparatus for outputting a pixel value of a 6x6 structure corrected for a bad pixel.

In order to achieve the above object, according to a preferred embodiment of the present invention, a separation unit for separating the input Bayer pattern image into the component-specific data; And calculating the distribution (pattern) of pixel values based on the median value of each data in the component-specific data, comparing patterns of other components, obtaining a comparison value based on each component, and comparing the comparison value with a bad pixel threshold. A bad pixel processing apparatus including a bad pixel detection unit for determining whether a bad pixel is compared with a value and outputting corrected component-specific data is provided.

In this case, the bad pixel detection unit may include a first center value detection unit for detecting a median value of each component data; A first absolute value detector for detecting an absolute value of a value obtained by subtracting the median value from each element of the component-specific data; A first comparison value detector configured to obtain an absolute value of a value obtained by subtracting the other output from one output with respect to the output of the first absolute value detector; A threshold setting unit for setting a threshold value using a threshold value for illuminance classification and the median value detected by the first median detector; And a first output determination unit for comparing the comparison value detected by the first comparison value detector with a bad pixel threshold value to determine output data according to the presence or absence of a bad pixel.

Here, the first output determination unit compares the comparison value detected by the first comparison value detection unit with the bad pixel threshold value, and when there is no comparison value larger than the bad pixel threshold value, the component-specific data. Is determined as an output, and the comparison value detected by the first comparison value detection unit and the bad pixel threshold value are respectively compared to each other. If the absolute value is all greater than the threshold value, the absolute value of the value obtained by subtracting the neighboring member elements from the corresponding member elements for each member is compared with the threshold determined by the threshold determination unit. The corresponding element is determined as a median value among the surrounding elements, and the comparison value detected by the first comparison value detector is compared with the bad pixel threshold value, respectively. When there is only one comparison value larger than the bad pixel threshold value, each component of the classification data is determined, but the absolute value of the value obtained by subtracting neighboring members from the corresponding member for any member is determined. Compared with the threshold determined by the value determining unit, if the absolute value is not greater than the threshold, the corresponding element is determined as an output.

On the other hand, the processing apparatus of the present invention obtains a pattern from component data received from the bad pixel detection unit, compares patterns of other components, obtains a comparison value based on each component, and compares the hot pixel comparison value with a threshold value. In addition, it is preferable to further include a hot pixel detection unit for outputting the corrected component-specific data by determining whether or not the hot pixel.

The hot pixel detection unit may include a second center value detection unit for detecting a median value of each component data received from the bad pixel detection unit; A second absolute value detector for detecting an absolute value of a value obtained by subtracting a median value detected by the second median detector from each component of the component-specific data received from the bad pixel detector; A second comparison value detector configured to obtain an absolute value of a value obtained by subtracting the other output from one output with respect to the output of the second absolute value detector, and add the sum to detect the comparison value for each component data; And a second output determination unit for comparing the comparison value detected by the second comparison value detection unit with the hot pixel threshold value to determine output data according to the presence or absence of a hot pixel.

In this case, the second output determiner compares the comparison value detected by the second comparison value detector with the hot pixel threshold value, and outputs the component-specific data when there is no comparison value larger than the hot pixel threshold value. And comparing the comparison value detected by the second comparison value detector with the hot pixel threshold value, and when there is only one comparison value that is larger than the hot pixel threshold value, determine each component of each component data. The absolute value of the value obtained by subtracting neighboring elements from the corresponding element for any member is compared with a threshold determined by the threshold determination unit, and if the absolute value is larger than the threshold value, the corresponding element is selected. The hot value is determined by a median value among the surrounding members, and the comparison value detected by the second comparison value detector is compared with the hot pixel threshold value, respectively. If there is only one comparison value that is larger than the threshold value, each component of the component-specific data is determined, but the absolute value of the value obtained by subtracting the surrounding components from the corresponding component for any member is determined. Compared with a threshold value, when the absolute value is not greater than the threshold value, the corresponding element is determined as an output.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. First of all, in adding reference numerals to the components of each drawing, it should be noted that the same components have the same number as much as possible even if displayed on different drawings. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a structural diagram of an embodiment of a bad pixel processing apparatus according to the present invention.

As shown in the figure, the processing apparatus according to the present invention includes a component separation unit 110, a bad pixel detection unit 120 and a hot pixel detection unit 130.

The component separator 110 separates the input 6 × 6 Bayer pattern into four components having a 3 × 3 structure as shown in FIG. 2. FIG. 2A illustrates an example of a 6 × 6 Bayer pattern image input to FIG. 1, and FIG. 2B illustrates an example of separating the Bayer pattern of FIG. 2A into components having a 3 × 3 structure.

As shown in the figure, the component separation unit 110 of the present invention is a 6 × 6 Bayer pattern of the same components, that is, R (red), Gr (G of GR line; green), Gb (G of GB line) and The data is separated into a 3 × 3 data structure composed only of B (blue) components. As shown in Fig. 2B, Gr, R, B, and Gb component data having a 3x3 structure will be referred to as C1, C2, C3, and C4 for convenience.

The bad pixel detection unit 120 obtains a pattern of each component data (the distribution of pixel values based on a median. The same applies below), and then compares the patterns of other components to measure the values based on each component. Then, the measured values are compared to determine whether or not the pixels are defective. Hereinafter, with reference to the drawings will be described in more detail.

3 is a detailed structural diagram of an exemplary bad pixel detection unit of FIG. 1.

As shown in the drawing, the bad pixel detection unit 120 of the present invention includes a median detection unit 121, an absolute value detection unit 122, a comparison value detection unit 123, a threshold value setting unit 124 and an output determination unit ( 125). Except for the input signals (C1, C2, C3, C4) in the drawings, the signal flow is processed as one line, but it is obvious to those skilled in the art that various data are included therein. .

The median value detector 121 detects median values of C1, C2, C3, and C4 data having an input 3x3 structure, respectively. Let the medians corresponding to the above data be M1, M2, M3 and M4, respectively.

For example, the absolute value detection unit 122 detects the absolute value Abs_1i of the value obtained by subtracting the median value M1 from each of the constituent elements (Gri, i = 1 to 9). Similarly for the C2, C3 and C4 data, the absolute values Abs_2i, Abs_3i and Abs_4i of the values obtained by subtracting the median values M2, M3 and M4 from the respective constituent elements (Ri, Bi and Gbi, i = 1 to 9) are detected. Abs_1i, Abs_2i, Abs_3i and Abs_4i are represented by Equation 1 below.

Where i is a natural number from 1 to 9.

The comparison value detector 123 obtains a comparison value by obtaining an absolute value of a value obtained by subtracting the other output from one output with respect to the output of the absolute value detector 122 obtained above. That is, the absolute values of the values obtained by subtracting the remaining Abs_2i, Abs_3i and Abs_4i from Abs_1i are calculated, and the sums thereof are calculated to obtain COMP_1i. In this manner, COMP_2i, COMP_3i and COMP_4i are obtained. This is expressed as the following equation.

Where i is a natural number from 1 to 9. FIG. 4 is an exemplary diagram for describing an operation of a median detector, an absolute detector, and a comparison detector of FIG. 3.

)으로 정의했을 때, 임계값 설정부(124)는 Y_m이 Y_threshold보다 작은 경우에는 thr을 0으로 설정하고, 그렇지 않은 경우에는 thr을 50으로 한다. 다만, 이에 한정되는 것은 아니며, 시스템에 따라 그 값이 달라질 수 있을 것이다.The threshold setting unit 124 sets the threshold value thr for the internal flag by using the threshold Y_threshold for illuminance classification applied and M1 to M4 which are medians of 3 × 3 data. Y_threshold is system adjustable. The arithmetic mean of the median of each component data is Y_m (i.e.

), The threshold value setting unit 124 sets thr to 0 when Y_m is less than Y_threshold, and sets thr to 50 otherwise. However, the present invention is not limited thereto, and the value may vary depending on the system.

The output determination unit 125 compares the four comparison values detected by the comparison value detection unit 123 and the applied bad pixel threshold value (dead_threshold), respectively, and if there is no comparison value larger than the bad pixel threshold value, the original image. Outputs That is, a 3 × 3 Cj (j = 1 to 4) image is output as it is. 5 is a diagram illustrating an image output by the bad pixel detection unit of FIG. 1. As shown in the drawing, the bad pixel detection unit 120 outputs a 3 × 3 Cj (j = 1 to 4) image. Therefore, the data output by the output determination unit 125 becomes an original image of 6 × 6 size. In the drawing, 'Size3 × 3_Cj1' refers to the first component of Cj data having a size of 3 × 3. It will be described below with the data of this structure.

If there is only one comparison value larger than the bad pixel threshold, the output determiner 125 reconstructs the image, and takes a method of constructing pixels for each element.

In order to construct Size3 × 3_Cj1 (the first element of Size3 × 3_Cj, j = 1 to 4), the output determining unit 125 includes the surrounding elements (that is, Size3 × 3_Cj2, Size3 × 3_Cj5 and The absolute value of the value minus Size3 × 3_Cj4) is compared with the threshold value thr, and if any one is smaller than the threshold value, the original input value, Size3 × 3_Cj1, is output; The median of 3_Cj2, Size3x3_Cj4, and Size3x3_Cj5 is output.

In addition, in order to configure Size3 × 3_Cj2 (the second element of Size3 × 3_Cj, which is j = 1 to 4), the output determination unit 125 includes the surrounding elements (that is, Size3 × 3_Cj1 and Size3 × in FIG. 5). Compare the absolute values of 3_Cj3, Size3 × 3_Cj4, Size3 × 3_Cj5, and Size3 × 3_Cj6) with the threshold thr, and if any one is smaller than the threshold, output the original input, Size3 × 3_Cj2. Otherwise, the median values of the surrounding elements Size3x3_Cj1, Size3x3_Cj3, Size3x3_Cj4, Size3x3_Cj5, and Size3x3_Cj6 are output. In this way, an output value is determined for all the elements of Size3 × 3_Cj (j = 1 to 4), and the Size3 × 3_Cj (j = 1 to 4) data is transmitted to the hot pixel determination unit 130 of FIG.

The hot pixel determination unit 130 obtains a pattern of input component data, compares patterns of other components, obtains measurement values based on each component, and compares the measured values to determine whether the pixel is a hot pixel. Hereinafter, with reference to the drawings will be described in more detail.

6 is a detailed structural diagram of an embodiment of the hot pixel detection unit of FIG. 1.

As shown in the figure, the hot pixel detector 130 of the present invention includes a median detector 131, an absolute value detector 132, a comparison value detector 133, and an output determiner 134. In the drawing, except for the input signals (Size3 × 3C1, Size3 × 3C2, Size3 × 3C3, Size3 × 3C4), the signal flow is treated as one line, but various data are included in the drawings. Self-explanatory to those who have knowledge.

The median value detector 131 detects median values of C1, C2, C3, and C4 data having an input 3x3 structure, respectively. Let the median values corresponding to the above data be M5, M6, M7 and M8, respectively. However, the median values M5, M6, M7, and M8 detected by the median detector 131 may use the median detected by the median detector 121 of the bad pixel detector 120 of FIG. 3 as it is.

For example, the absolute value detection unit 132 detects the absolute value Abs_5i of the value obtained by subtracting the median value M5 from each component (Size3 × 3C1i, i = 1 to 9). Similarly for Size3 × 3C2, Size3 × 3C3, and Size3 × 3C4 data, the absolute value of each component (Size3 × 3C2i, Size3 × 3C3i and Size3 × 3C4i, i = 1 to 9) minus the median values M6, M7, and M8, respectively The values Abs_6i, Abs_7i and Abs_8i are detected. Abs_5i, Abs_6i, Abs_7i, and Abs_8i are represented by Equation (3).

Where i is a natural number from 1 to 9.

The comparison value detector 133 obtains a comparison value by obtaining an absolute value of the output obtained by subtracting the other output from one output, with respect to the output of the absolute value detector 132 obtained above. That is, the absolute values of the values obtained by subtracting the remaining Abs_6i, Abs_7i, and Abs_8i from Abs_5i are calculated, and the sums are calculated to obtain COMP_5i. In this manner, COMP_6i, COMP_7i and COMP_8i are obtained. This is expressed as the following equation.

Where i is a natural number from 1 to 9.

The output determiner 134 compares the four comparison values detected by the comparison value detector 133 and the applied hot pixel threshold value hot_threshold, respectively, and outputs the original image when there is no comparison value larger than the hot pixel threshold value. do. That is, a 3 × 3 Cj (j = 1 to 4) image is output as it is. Therefore, the data output by the output determination unit 134 becomes a 6 × 6 original image.

If there is only one comparison value larger than the hot pixel threshold value, the output determination unit 134 reconfigures the image, and takes a method of constructing pixels for each element.

In order to construct Size3 × 3_Cj1 (the first element of Size3 × 3_Cj, j = 1 to 4), the output determining unit 134 includes the surrounding elements (that is, Size3 × 3_Cj2, Size3 × 3_Cj5 and The absolute value of the value obtained by subtracting Size3 × 3_Cj4) is compared with the threshold value thr determined and transmitted by the threshold value determination unit 124 of the bad pixel detection unit 120, and if any one is smaller than the threshold value, the original input is performed. The value Size3x3_Cj1 is output. Otherwise, the median values of the surrounding elements Size3x3_Cj2, Size3x3_Cj4 and Size3x3_Cj5 are output.

In addition, in order to configure Size3 × 3_Cj2 (the second element of Size3 × 3_Cj, j = 1 to 4), the output determining unit 134 includes the surrounding elements (that is, Size3 × 3_Cj1 and Size3 × in FIG. 5). Compare the absolute values of 3_Cj3, Size3 × 3_Cj4, Size3 × 3_Cj5, and Size3 × 3_Cj6) with the threshold thr, and if any one is smaller than the threshold, output the original input, Size3 × 3_Cj2. Otherwise, the median values of the surrounding elements Size3x3_Cj1, Size3x3_Cj3, Size3x3_Cj4, Size3x3_Cj5, and Size3x3_Cj6 are output. In this way, the output values are determined for all the members of Size3x3_Cj (j = 1-4), and the Size3x3_Cj (j = 1-4) data can be output. This outputs 6x6 data.

FIG. 7A is an exemplary diagram illustrating an original image input to an apparatus of the present invention, and FIG. 7B is an exemplary diagram illustrating correcting by detecting a bad pixel and a hot pixel with respect to FIG. 7A.

As shown in the figure, it can be seen that a bad pixel and a hot pixel are detected and corrected.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes can be made in the art without departing from the technical spirit of the present invention. It will be clear to those of ordinary knowledge.

As described above, the present invention compares and analyzes the pattern of the component including the central pixel and the patterns of other components in the 6 × 6 mask to accurately determine whether the pixel is a bad pixel or a hot pixel, thereby correcting the 6 × 6 structure. This has the effect of letting you print an image.

Claims (10)

A separation unit for separating the input Bayer pattern image into its component-specific data; And From the component data, the distribution (pattern) of pixel values based on the median value of each data is obtained, the patterns of other components are compared to obtain a comparison value based on each component, and the comparison value is referred to as a bad pixel threshold value. And a bad pixel detection unit for determining whether or not the pixel is defective and outputting corrected component data. The method of claim 1, wherein the bad pixel detection unit, A first median detector for detecting a median of the component-specific data; A first absolute value detector for detecting an absolute value of a value obtained by subtracting the median value from each element of the component-specific data; A first comparison value detector configured to obtain an absolute value of a value obtained by subtracting the other output from one output with respect to the output of the first absolute value detector; A threshold setting unit for setting a threshold value using a threshold value for illuminance classification and the median value detected by the first median detector; And And a first output determiner configured to compare the comparison value detected by the first comparison value detector and the bad pixel threshold value to determine output data based on the presence of a bad pixel. 3. The method of claim 2, wherein the first output determiner compares the comparison value detected by the first comparison value detection unit with the bad pixel threshold value, and there is no comparison value larger than the bad pixel threshold value. The bad pixel processing apparatus for determining the component-specific data as an output. 3. The method of claim 2, wherein the first output determiner compares the comparison value detected by the first comparison value detector with the bad pixel threshold value, and there is only one comparison value larger than the bad pixel threshold value. And, respectively, determine the members of the data for each component, and compare the absolute value of the value obtained by subtracting the surrounding components from the corresponding member elements with respect to the threshold value determined by the threshold determination unit. And all of the neighboring elements are determined as the median of the neighboring elements when all of them are larger than the threshold. 3. The method of claim 2, wherein the first output determiner compares the comparison value detected by the first comparison value detector with the bad pixel threshold value, and there is only one comparison value larger than the bad pixel threshold value. And, respectively, determine the members of the data for each component, and compare the absolute value of the value obtained by subtracting the surrounding components from the corresponding member elements with respect to the threshold value determined by the threshold determination unit. The defective pixel processing device that determines the component as an output when it is not larger than the threshold. The method according to any one of claims 1 to 3, wherein a pattern is obtained from component-specific data received from the bad pixel detection unit, a pattern of other components is compared, a comparison value is calculated based on each component, and the hot pixel comparison is performed. And a hot pixel detector configured to determine whether the pixel is a hot pixel by comparing the value with a threshold value and output corrected component-specific data. The method of claim 6, wherein the hot pixel detection unit, A second center value detector for detecting a median value of the component-specific data received from the bad pixel detector; A second absolute value detector for detecting an absolute value of a value obtained by subtracting a median value detected by the second median detector from each component of the component-specific data received from the bad pixel detector; A second comparison value detector configured to obtain an absolute value of a value obtained by subtracting the other output from one output with respect to the output of the second absolute value detector, and add the sum to detect the comparison value for each component data; And And a second output determination unit for comparing the comparison value detected by the second comparison value detection unit with a hot pixel threshold value to determine output data according to the presence or absence of a hot pixel. 8. The method of claim 7, wherein the second output determination unit compares the comparison value detected by the second comparison value detection unit with the hot pixel threshold value, respectively, and when there is no comparison value larger than the hot pixel threshold value, the component. Defective pixel processing unit that determines the star data as output. 8. The method of claim 7, wherein the second output determination unit compares the comparison value detected by the second comparison value detection unit with the hot pixel threshold value, respectively, and when there is only one comparison value larger than the hot pixel threshold value, the component. Each member of the star data is determined, but the absolute value of the value obtained by subtracting the surrounding member from the corresponding member for any member is compared with the threshold determined by the threshold determination unit, and the absolute value is the threshold. If larger than all, the bad pixel processing apparatus for determining the corresponding element as the median value of the surrounding elements. 8. The method of claim 7, wherein the second output determination unit compares the comparison value detected by the second comparison value detection unit with the hot pixel threshold value, respectively, and when there is only one comparison value larger than the hot pixel threshold value, the component. Each member of the star data is determined, but the absolute value of the value obtained by subtracting the surrounding member from the corresponding member for any member is compared with the threshold determined by the threshold determination unit, and the absolute value is the threshold. The defective pixel processing device that determines the component as an output when it is not larger.

Images