KR101175737B1 - Defect pixel processing method and device - Google Patents

Abstract

Disclosed are a defective pixel processing method and apparatus for detecting and correcting a defective pixel resulting from a defect of an image sensor. An image pattern input unit for receiving an M x M Bayer pattern including a target pixel for determining whether a defective pixel is included in an input image, and extracting a magnetic component corresponding to the target pixel among the M x M Bayer patterns to obtain an N x N magnetic component A magnetic component extracting unit for generating a mask, a defective pixel determining unit determining whether the target pixel is a defective pixel by analyzing characteristics of the N × N magnetic component mask, and the N × N magnetic field when the target pixel is a defective pixel A correction unit for correcting by detecting an edge binary pattern of a component mask and when the target pixel is a defective pixel, outputting a correction value performed by the correction unit as a pixel value of the target pixel, and the target pixel is a normal pixel According to a defective pixel processing apparatus including an output unit for outputting the pixel value of the target pixel as it is, the Bayer image unit It is possible to improve the accuracy of detection and to remove the cluster defect pixels by performing feature component separation, characterization and strict detection method.

Description

Defect pixel processing method and device

The present invention relates to a defective pixel processing method and apparatus for detecting and correcting a defective pixel resulting from a defect of an image sensor.

An image sensor is a semiconductor device that converts an optical image into an electrical signal. Among these, a charge coupled device (CCD) is a device in which charge carriers are stored and transported in a capacitor while individual metal-oxide-silicon (MOS) capacitors are located in close proximity to each other. On the other hand, CMOS (Complementary MOS) image sensor, by using CMOS technology using a control circuit and a signal processing circuit as a peripheral circuit to make as many MOS transistors as the number of pixels, and sequentially output the pixels using the same It is a device that employs a switching method for detecting.

Portable devices (such as digital cameras, mobile communication terminals, etc.) equipped with such image sensors have been developed and sold. The image sensor is configured as an array of small photosensitive diodes called pixels or photosites. The pixels themselves usually do not extract color from light, but only convert photons from a broad spectral band into electrons. In order to record color images with a single sensor, the sensor is filtered so that different pixels receive different color illumination. This type of sensor is known as a color filter array (CFA). Different color filters are arranged in a predefined pattern across the sensor.

The color filter array of the color image generally follows a Bayer pattern. That is, half of the total number of pixels is green (G), and each quarter of the total number is allocated to red (R) and blue (B). In order to obtain the color information, the color image pixels are formed in a repeating pattern with a red, green or blue filter, for example, a 2 x 2 array in the case of the Bayer pattern. This Bayer pattern is based on the premise that the human eye derives most of the luminance information from the green content of the image. Thus, by allowing more pixels of all the pixels to be green, a higher resolution image can be generated compared to the alternate RGB color filter in the same number of red, green and blue pixels.

Such an image sensor may generate a defective pixel such as a dead pixel or a hot pixel due to an error in a manufacturing process. A dead pixel is a case in which the pixel itself does not develop color due to a defect, etc., and a hot pixel is a case in which the pixel itself generates color but does not develop a proper color. Such defective pixels are an important factor in determining the grade and price of an image sensor.

The pixel data generated by the defective pixel has a feature that is too large or vice versa smaller than the pixel data generated by the adjacent pixel. The conventional defect pixel processing method mainly detects a corresponding pixel as a defective pixel and corrects it when the difference exceeds a predetermined range compared with a reference pixel value. However, since the detection accuracy is low, the normal pixel is often detected as a defective pixel, and the corrected image has a problem of generating more distortion than the actual image.

The above-described background technology is technical information that the inventor holds for the derivation of the present invention or acquired in the process of deriving the present invention, and can not necessarily be a known technology disclosed to the general public prior to the filing of the present invention.

The present invention provides a defective pixel processing method and apparatus capable of identifying features after performing component separation from a Bayer image, applying a strict detection method to increase detection accuracy, and removing cluster defective pixels (two or more defective pixels). It is to provide.

Further, the present invention is to provide a defective pixel processing method and apparatus capable of correcting to an optimal pixel value that classifies an edge pattern with respect to a defective pixel so that false color does not occur.

In addition, the present invention is to provide a defect pixel processing method and apparatus that can detect and correct any number of defective pixels by simply adjusting the threshold value within an input mask, and is easy to tune due to the small number of adjustable parameters. .

Other objects of the present invention will be readily understood through the following description.

According to an aspect of the present invention, an apparatus for processing a defective pixel of an input image, the apparatus comprising: an image pattern input unit configured to receive an M x M Bayer pattern including a target pixel for determining whether a defective pixel is included in the input image; A magnetic component extractor for extracting a magnetic component corresponding to the target pixel of the M × M Bayer pattern to generate an N × N magnetic component mask, wherein M and N are natural numbers and M> N; A defective pixel determination unit determining whether the target pixel is a defective pixel by analyzing characteristics of the N × N magnetic component mask; A correction unit configured to identify and correct an edge binary pattern of the N × N magnetic component mask when the target pixel is a defective pixel; And an output unit configured to output a correction value performed by the correction unit as the pixel value of the target pixel when the target pixel is a defective pixel, and output the pixel value of the target pixel as it is when the target pixel is a normal pixel. A pixel processing apparatus is provided.

The M x M Bayer pattern may include the target pixel as a center pixel.

The magnetic component extractor may select pixels having the same color component as the color component of the target pixel among the pixels in the M × M Bayer pattern to generate the N × N magnetic component mask having a corresponding relative position.

The defective pixel determination unit may include: a reference value calculation module configured to calculate a reference value for determining whether the defective pixel is determined by calculating a difference with a neighboring pixel value based on a center pixel in the N × N magnetic component mask; A frequency difference calculation module for calculating a frequency difference value for inferring an amount of high frequency components included in the N × N magnetic component mask; A detection threshold calculation module for calculating a detection threshold according to the frequency difference value using a detection threshold graph; And a determination module configured to determine whether the target pixel is a defective pixel by comparing the reference value and the detection threshold value.

The reference value calculation module may calculate an absolute value of the difference between pixel values of each of the neighboring pixels of the N × N magnetic component mask and the center pixel, arrange the values in ascending order, and calculate an average of the first predetermined number as the reference value.

The frequency difference calculation module may calculate a difference between a maximum value and a minimum value among pixel values in the N × N magnetic component mask as the frequency difference value.

The detection threshold calculation module may calculate the detection threshold using a detection threshold graph according to the following equation.

Here, Freq_diff_value is a frequency difference value calculated by the frequency difference calculation module, dpc_freq_thr is a frequency threshold for distinguishing low frequency and high frequency, dpc_freq_width is a width interpolating the threshold of low frequency and high frequency, and dpc_lf_thr is a defect for low frequency. Is a pixel detection threshold, and dpc_hf_thr is a defective pixel detection threshold for high frequency.

The detection threshold graph may be adjusted according to the luminance area to which the target pixel belongs.

The correction unit matches a case in which the greater than or equal to the average of the maximum value and the minimum value is '1' for the remaining pixels except for the center pixel in the N × N magnetic component mask, and the case where the small value is '0'. The edge binary pattern of the N × N magnetic component mask may be classified by corresponding consecutive matching values of the peripheral pixels in a predetermined direction from the peripheral pixels to the binary values.

In addition, the correction unit may average the pixel values of the neighboring pixels matched with '0' according to the edge binary pattern classified with respect to the N × N magnetic component mask, the pixel values average of the neighboring pixels matched with '1', and the N x One of the middle values of pixel values of all pixels in the N magnetic component mask may be determined as a correction value of the target pixel.

Meanwhile, according to another aspect of the present invention, a method of processing a defective pixel of an input image and a recording medium on which a program for performing the same are recorded are provided.

According to an exemplary embodiment, a method of processing a defective pixel includes: receiving an M × M Bayer pattern including a target pixel for determining whether a defective pixel is included in the input image; Extracting a magnetic component corresponding to the target pixel of the M × M Bayer pattern to generate an N × N magnetic component mask, where M and N are natural numbers and M> N; Determining whether the target pixel is a defective pixel by analyzing characteristics of the N × N magnetic component mask; Determining an edge binary pattern of the N × N magnetic component mask and performing correction when the target pixel is a defective pixel; And outputting a correction value performed by the correction unit as a pixel value of the target pixel when the target pixel is a defective pixel, and outputting a pixel value of the target pixel as it is when the target pixel is a normal pixel. have.

The M x M Bayer pattern may include the target pixel as a center pixel.

In the magnetic component extraction step, the pixels having the same color component as that of the target pixel among the pixels in the M × M Bayer pattern may be selected to generate the N × N magnetic component mask having a corresponding relative position. .

The determining of the defective pixel may include calculating a reference value for determining whether the defective pixel is determined by calculating a difference from a neighboring pixel value based on a center pixel in the N × N magnetic component mask; Calculating a frequency difference value for inferring an amount of high frequency components included in the N × N magnetic component mask; Calculating a detection threshold according to the frequency difference using a detection threshold graph; The method may include determining whether the target pixel is a defective pixel by comparing the reference value and the detection threshold value.

The calculating of the reference value may include calculating an absolute value of a pixel value difference between each of the peripheral pixels of the N × N magnetic component mask and the center pixel; And sorting them in ascending order and calculating the average of the first predetermined number as the reference value.

In the calculating of the frequency difference value, a difference between a maximum value and a minimum value among pixel values in the N × N magnetic component mask may be calculated as the frequency difference value.

In the detecting threshold calculation step, the detection threshold may be calculated using a detection threshold graph according to the following equation.

Here, Freq_diff_value is a frequency difference value calculated in the frequency difference value calculating step, dpc_freq_thr is a frequency threshold for distinguishing low frequency and high frequency, dpc_freq_width is a width interpolating the threshold of low frequency and high frequency, and dpc_lf_thr is a defect for low frequency. Is a pixel detection threshold, and dpc_hf_thr is a defective pixel detection threshold for high frequency.

The detection threshold graph may be adjusted according to the luminance area to which the target pixel belongs.

The correcting step may include matching a case where a value greater than or equal to a maximum value and a minimum value of '1' and a case where the small value is '0' with respect to the remaining pixels except for the center pixel in the N × N magnetic component mask. ; And classifying the edge binary pattern of the N × N magnetic component mask by mapping the successive matching values of the respective peripheral pixels from the arbitrary peripheral pixels to the binary values in a predetermined direction.

In addition, the correcting may include: an average of pixel values of neighboring pixels matched with '0' according to the edge binary pattern classified for the N × N magnetic component mask, an average of pixel values of peripheral pixels matched with '1', and the N The method may further include determining one of an intermediate value of pixel values of all pixels in the x N magnetic component mask as a correction value of the target pixel.

Other aspects, features, and advantages will become apparent from the following drawings, claims, and detailed description of the invention.

According to an embodiment of the present invention, after performing component separation from a Bayer image, it is possible to identify features and to apply a strict detection method to increase detection accuracy and to remove cluster defective pixels (two or more defective pixels).

In addition, it is possible to classify the edge pattern with respect to the defective pixel and correct it to an optimal pixel value such that false color does not occur.

In addition, it is possible to detect and correct any number of defective pixels by adjusting only a threshold in the input mask, and there is an advantage in that tuning is easy due to the small number of adjustable parameters.

1 is a view showing a schematic configuration of a defective pixel processing apparatus according to an embodiment of the present invention;
2 is a view showing an example of a 5x5 Bayer pattern according to an embodiment of the present invention,
3 illustrates an example of a 3x3 magnetic component mask according to an embodiment of the present invention;
4 is a view showing a schematic configuration of a defective pixel determination unit according to an embodiment of the present invention;
5 is an exemplary diagram illustrating a detection threshold graph according to an embodiment of the present invention;
6 is an exemplary view illustrating a luminance expression color bar according to an embodiment of the present invention;
7 is an exemplary diagram of a matching process and a generated edge binary pattern for obtaining an edge binary pattern according to an embodiment of the present invention;
8 is a flowchart of a defective pixel processing method according to an embodiment of the present invention;
9 is an exemplary view comparing an image before processing and an image after processing by the defective pixel processing apparatus according to the exemplary embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, the terms "comprises" or "having" and the like refer to the presence of stated features, integers, steps, operations, elements, components, or combinations thereof, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Also, the terms " part, "" module," and the like, which are described in the specification, refer to a unit for processing at least one function or operation, and may be implemented by hardware or software or a combination of hardware and software.

In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a view showing a schematic configuration of a defective pixel processing apparatus according to an embodiment of the present invention, Figure 2 is a view showing an example of a 5x5 Bayer pattern according to an embodiment of the present invention, Figure 3 is a present invention 4 is a view showing an example of a 3x3 magnetic component mask according to an embodiment of the present invention, FIG. 4 is a view showing a schematic configuration of a defective pixel determination unit according to an embodiment of the present invention, and FIG. 5 is a view showing an embodiment of the present invention. 6 is an exemplary diagram illustrating a detection threshold graph according to an embodiment of the present invention, FIG. 6 is an exemplary diagram illustrating a luminance expression color bar according to an embodiment of the present invention, and FIG. 7 is a matching process for obtaining an edge binary pattern according to an embodiment of the present invention; It is an illustration of the generated edge binary pattern.

The defective pixel processing apparatus 100 according to the present exemplary embodiment receives an image pattern including a target pixel that is a target for determining whether a defect is present, and determines whether the defective pixel is determined by using only a magnetic component corresponding to the target pixel. The detected defective pixel is corrected by finding an optimal correction value using a pre-classified edge binary pattern and outputting the optimal correction value.

To this end, the defective pixel processing apparatus 100 according to the present exemplary embodiment receives an M × M Bayer pattern, and generates an N × N magnetic component mask using only the magnetic component corresponding to the target pixel therefrom. Here, M and N are natural numbers, and M> N. Hereinafter, a case in which M = 5 and N = 3 will be described by way of example, but the scope of the present invention is not limited thereto.

Referring to FIG. 1, the defective pixel processing apparatus 100 according to an exemplary embodiment may include an image pattern input unit 110, a magnetic component extracting unit 120, a defective pixel determining unit 130, and a correcting unit 140. And an output unit 150.

The image pattern input unit 110 receives a 5 × 5 Bayer pattern in which a target pixel, which is a determination target of a defective pixel, is located in an input image frame having a Bayer pattern.

Referring to FIG. 2, an example of a 5 × 5 Bayer pattern input to the image pattern input unit 110 is illustrated. For example, a 5x5 Bayer pattern 10 having the target pixel Rc of the R component as a center pixel of one image frame may be selected and input.

Although the target pixel Rc is illustrated as having an R component in FIG. 2, the 5x5 Bayer pattern around the pixel may be selected in the same manner with respect to the pixels of the remaining color components B, Gr, and Gb. Self-explanatory

The magnetic component extractor 120 extracts only a pixel having the same color component (ie, magnetic component) as the center pixel among the 5x5 Bayer patterns 10 input through the image pattern input unit 110 to generate a 3x3 magnetic component mask 20. do. Referring to FIG. 3, a 3x3 magnetic component mask 20 generated by extracting only the pixels of the hatched R component among the 5x5 Bayer patterns illustrated in FIG. 2 is illustrated. Each component Pi of the 3x3 magnetic component mask 20 corresponds to an R component at Ri (where i = 1 to 8, c) in the 5x5 Bayer pattern 10 (ie, Pi = Ri, i = 1-8, c).

The defective pixel determiner 130 determines whether the target pixel is a defective pixel by using a 3x3 magnetic component mask generated by the magnetic component extractor 120.

Referring to FIG. 4, the defective pixel determination unit 130 includes a reference value calculation module 132, a frequency difference calculation module 134, a detection threshold calculation module 136, and a determination module 138.

The reference value calculation module 132 calculates a reference value for determining whether a defective pixel is determined by calculating a difference between neighboring pixel values of the pixels in the 3 × 3 magnetic component mask based on the center pixel.

For example, according to the outlier remover filtering method, the absolute value S'i of the pixel value difference between each of the eight neighboring pixels and the center pixel is calculated as shown in Equation 1 below, and the values are sorted in ascending order. After that, the average of the first predetermined number (for example, four), that is, the average of the predetermined number in order of decreasing S'i can be used as the reference value.

[Equation 1]

S'i = | Pi-Pc |, i = 1 to 8

The reference value calculated as described above is used to find out how much the image is popping up when the center pixel is compared with the surrounding pixels, and is used to find the popping noise in the image frame.

The frequency difference calculation module 134 calculates frequency difference values that can be inferred from how high frequency components, such as, for example, edges, are included in the 3x3 magnetic component mask. For example, the frequency difference value may be calculated by Equation 2 below.

&Quot; (2) "

Freq_diff_value = max (P [1: 8, c])-min (P [1: 8, c])

Where Freq_diff_value is the frequency difference value, max (P [1: 8, c]) is the maximum value of the pixel values in the 3x3 magnetic component mask, and min (P [1: 8, c]) is the pixel in the 3x3 magnetic component mask. It represents the minimum value among the values.

The detection threshold calculation module 136 calculates a detection threshold for detecting the defective pixel according to the frequency difference value. The calculation of the detection threshold may be performed using the detection threshold graph 30 according to the frequency difference value illustrated in FIG. 5, and the detection threshold may be applied regardless of the type of color component.

In the detection threshold graph 30, dpc_thr is a threshold for detecting a defective pixel, and frequency represents a frequency difference value calculated by the frequency difference calculation module 134, and may be expressed as Equation 3 below.

&Quot; (3) "

Here, Freq_diff_value is a frequency difference value calculated by the frequency difference value calculation module 134, dpc_freq_thr is a frequency threshold for distinguishing low frequency and high frequency, dpc_freq_width is a width interpolating the threshold of low frequency and high frequency, and dpc_lf_thr is low frequency. Is a defective pixel detection threshold for dpc_hf_thr and a defective pixel detection threshold for high frequencies.

According to Equation 3, when the frequency difference value belongs to the low frequency region (Freq_diff_value <F1), the detection threshold value is the low frequency detection threshold value (dpc_lf_thr), and when the frequency difference value belongs to the high frequency region (Freq_diff_value> F2), the detection threshold value is high frequency detection. Threshold (dpc_hf_thr), where the frequency difference lies between the low frequency region and the high frequency region (otherwise), the detection threshold may be a value according to a linear function expressed as a frequency threshold, a frequency interpolation width, a low frequency detection threshold, and a high frequency detection threshold. have.

Here, the frequency threshold (dpc_freq_thr), the frequency interpolation width (dpc_freq_width), the low frequency detection threshold (dpc_lf_thr), and the high frequency detection threshold (dpc_hf_thr) are parameters that can be adjusted by the user, depending on the luminance region to which the target pixel, that is, the center pixel belongs. A tuning point at which values may be determined experimentally.

Referring to FIG. 6, a luminance expression color bar is illustrated. For example, when the luminance value of the center pixel is smaller than the first luminance threshold dpc_Luma_thr1, it is determined to belong to the region 1 and has a low luminance. If it is larger than the luminance threshold value dpc_Luma_thr2, it may be determined to belong to region 3 and have high luminance. If it is between the first and second luminance threshold values, it may be determined to belong to region 2 and have medium luminance. According to the determination result, parameters such as a frequency threshold, a frequency interpolation width, a low frequency detection threshold, a high frequency detection threshold, and the like are determined, and according to the determination, the detection threshold graph 30 as illustrated in FIG. 5 can be obtained.

The detection threshold calculation module 136 may find a detection threshold according to the frequency difference value calculated by the current frequency difference calculation module 134 using the detection threshold graph 30 obtained through the above process.

The determination module 138 checks whether the center pixel in the 3x3 magnetic component mask is a defective pixel using the reference value calculated in the reference value calculation module 132 and the detection threshold calculated in the detection threshold calculation module 136. For example, when the reference value is larger than the detection threshold value, it may be determined as a defective pixel, and when the reference value is less than or equal to the detection threshold value, it may be determined as a normal pixel.

Referring back to FIG. 1, the correction unit 140 performs pixel value correction on the target pixel determined as the defective pixel as a result of the determination by the defective pixel determination unit 130. The pixel value correction may be performed by classifying an edge binary pattern of the 3x3 magnetic component mask 20 having the target pixel as a center pixel, and finding an optimal correction value for the identified pattern.

One example of an edge binary pattern is illustrated in FIG. 7. In the 3x3 magnetic component mask, the remaining pixels except for the center pixel C are matched to '1' when the value is larger than or equal to the average of the maximum value and the minimum value, and to '0' when the value is smaller (see FIG. 7A). ). 8 bits corresponding to [b0: b7] by corresponding to b0 to b7 by matching the matching values of the respective peripheral pixels from a predetermined peripheral pixel (for example, the upper left pixel) to a predetermined direction (for example, a clockwise direction). An edge binary pattern may be generated (see FIG. 7B). For example, when b0 to b2 match '0' and b3 to b7 match '1', the generated edge binary pattern is '00011111'.

The edge binary pattern may have a total of 256 patterns. When the edge binary pattern of the center pixel currently determined as a defective pixel is identified, the edge binary pattern may be applied to a classifier calculated in advance to find an optimal correction value. Can be. For example, the pixel value average group of neighboring pixels matched with '0' is M1, and the pixel value plain group of neighboring pixels matched with '1' is M2, and the median value of all pixel values of the 3x3 magnetic component mask is determined. In the case of M3, one of M1, M2, and M3 may be determined experimentally according to the type of edge binary pattern.

Referring back to FIG. 1, when the determination result of the defective pixel determination unit 130 determines that the target pixel is a normal pixel, the output unit 150 determines an uncorrected original pixel value and determines that the target pixel is a defective pixel. In this case, the correction value by the correction unit 140 is output as the pixel value of the target pixel.

8 is a flowchart of a defective pixel processing method according to an exemplary embodiment of the present invention. Each step described below may be performed by each internal component of the defective pixel processing apparatus 100.

In operation S210, the image pattern input unit 110 receives a 5 × 5 Bayer pattern including a target pixel as a center pixel for determining whether a defective pixel is included in an input image frame. An example of a 5x5 Bayer pattern is shown by way of example in FIG.

In operation S220, pixels having the same color component as that of the center pixel of the 5 × 5 Bayer pattern are extracted as magnetic components to generate a 3 × 3 magnetic component mask. An example of a 3x3 magnetic component mask is shown by way of example in FIG. 3.

In operation S230, the characteristic of the 3x3 magnetic component mask is analyzed to determine whether the target pixel is a defective pixel.

To determine whether there is a defective pixel, first, in step S232, the reference value calculation module 132 calculates a difference between neighboring pixel values based on the center pixel to calculate a reference value for determining whether the defective pixel is present. The reference value may be an average of the first predetermined number selected after calculating the absolute value of the pixel value difference between each of the eight peripheral pixels of the 3x3 magnetic component mask and the center pixel as shown in Equation 1, sorting them in ascending order.

Independently of this, in step S234, the frequency difference calculation module 134 calculates a frequency difference that can be inferred from how much high frequency components are included in the 3x3 magnetic component mask. The calculation method of the frequency difference is shown in Equation 2 above.

Thereafter, the detection threshold calculation module 136 calculates a detection threshold for detecting the defective pixel according to the frequency difference value calculated in step S234 using the detection threshold graph. Here, the detection threshold graph for calculating the detection threshold has been described in detail with reference to Equation 3 and FIG. 5, which may be efficiently detected and corrected by the user according to the luminance region to which the target pixel belongs. Could be adjusted.

In step S238, the determination module 138 compares the reference value calculated in step S232 with the detection threshold value calculated in step S236 to determine whether the corresponding pixel is a defective pixel. For example, when the reference value is larger than the detection threshold, it may be determined as a defective pixel, and when it is the same or smaller, it may be determined as a normal pixel.

If it is determined that the defect pixel is determined, the process proceeds to step S240 and the correction unit 140 classifies and corrects the edge binary pattern of the defect pixel.

The classification of the edge binary pattern may be performed by matching pixel values of neighboring pixels to '0' or '1' with respect to a 3x3 magnetic component mask having a target pixel as a center pixel, and corresponding to an 8-bit binary value. This has been described in detail with reference to FIG. 7. The correction unit 140 finds the best correction value experimentally determined according to the classified edge binary pattern.

Thereafter, in step S250, the output unit 150 outputs the correction value corrected by the correction unit 140 to complete the processing for the defective pixel.

If it is determined in step S238 that the pixel is normal, the process proceeds to step S260 and the output unit 150 outputs the uncorrected center pixel value as it is.

It is apparent that the above-described defective pixel processing method may be performed by an automated procedure according to a time series sequence by a software program or the like embedded in the defective pixel processing apparatus 100. The codes and code segments that make up the program can be easily deduced by a computer programmer in the field. In addition, the program is stored in a computer-readable information storage medium, and the program is read and executed by a computer to implement the method. The information storage medium includes a magnetic recording medium, an optical recording medium, and a carrier wave medium.

9 is an exemplary diagram comparing an image before processing and an image after processing by the defective pixel processing apparatus according to the exemplary embodiment of the present disclosure.

FIG. 9A illustrates an original image input to a defective pixel processing apparatus according to an exemplary embodiment, and FIG. 9B illustrates a result image of detecting and correcting a defective pixel. Illustrated by way of example.

In FIG. 9A, a plurality of defective pixels displayed as dots are displayed on a part of the image. In FIG. 9B, all of the defective pixels are detected and corrected.

According to the present invention, only the magnetic component of the M x M Bayer pattern is used, and defect pixels are detected by identifying characteristics of peripheral pixels in the N x N magnetic component mask as well as the center pixel. Since the detected defective pixels classify the edge binary patterns of the masks to find an optimal correction value, the defective pixels can be removed while minimizing the occurrence of false color. In addition, there is an advantage in that the defect pixels can be detected and corrected only by adjusting the detection threshold without limiting the number, and the tuning is easy due to the small number of adjustable parameters.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention as defined in the following claims And changes may be made without departing from the spirit and scope of the invention.

100: defective pixel processing unit 110: image pattern input unit
120: magnetic component extraction unit 130: defective pixel determination unit
140: correction unit 150: output unit
10: 5x5 Bayer Pattern 20: 3x3 Magnetic Component Mask
132: reference value calculation module 134: frequency difference calculation module
136: detection threshold calculation module 138: decision module
30: detection threshold graph

Claims (21)

An apparatus for processing a defective pixel of an input image,
An image pattern input unit configured to receive an M × M Bayer pattern including a target pixel for determining whether a defective pixel is included in the input image;
A magnetic component extractor for extracting a magnetic component corresponding to the target pixel of the M × M Bayer pattern to generate an N × N magnetic component mask, wherein M and N are natural numbers and M>N;
A defective pixel determination unit determining whether the target pixel is a defective pixel by analyzing characteristics of the N × N magnetic component mask;
A correction unit configured to determine and correct an edge binary pattern of the N × N magnetic component mask when the target pixel is a defective pixel; And
A defective pixel including an output unit configured to output a correction value performed by the correction unit as a pixel value of the target pixel when the target pixel is a defective pixel, and output the pixel value of the target pixel as it is when the target pixel is a normal pixel Processing unit.
The method of claim 1,
The M x M Bayer pattern includes the target pixel as a center pixel.
The method of claim 1,
The magnetic component extractor selects pixels having the same color component as the color component of the target pixel among the pixels in the M × M Bayer pattern to generate the N × N magnetic component mask having a corresponding relative position. Defective pixel processing device.
The method of claim 1,
The defective pixel determination unit,
A reference value calculation module configured to calculate a reference value for determining whether there is a defective pixel by calculating a difference with a neighboring pixel value based on a center pixel in the N × N magnetic component mask;
A frequency difference calculation module for calculating a frequency difference value for inferring an amount of high frequency components included in the N × N magnetic component mask;
A detection threshold calculation module for calculating a detection threshold according to the frequency difference value using a detection threshold graph;
And a determination module which determines whether the target pixel is a defective pixel by comparing the reference value and the detection threshold value.
The method of claim 4, wherein
The reference value calculation module calculates an absolute value of a difference between pixel values of each of the neighboring pixels of the N × N magnetic component mask and the center pixel, sorts the values in ascending order, and calculates an average of the first predetermined number as the reference value. Defective pixel processing device.
The method of claim 4, wherein
And the frequency difference calculation module calculates a difference between a maximum value and a minimum value among pixel values in the N × N magnetic component mask as the frequency difference value.
The method of claim 4, wherein
The detection threshold calculation module calculates the detection threshold value using a detection threshold graph according to the following equation.
[Mathematical Expression]

Here, dpc_thr is the detection threshold value, Freq_diff_value is the frequency difference value calculated by the frequency difference calculation module, dpc_freq_thr is a frequency threshold for distinguishing low and high frequencies, dpc_freq_width is the width interpolating the threshold of low and high frequencies , dpc_lf_thr is a defective pixel detection threshold for low frequency, dpc_hf_thr is a defective pixel detection threshold for high frequency, F1 is a first reference value for calculating the detection threshold calculated as a difference of 1/2 of dpc_freq_thr and dpc_freq_width, F2 is a second reference value for calculating the detection threshold calculated as a sum of 1/2 of dpc_freq_thr and dpc_freq_width.
The method of claim 4, wherein
And the detection threshold graph is adjustable according to a luminance region to which the target pixel belongs.
The method of claim 1,
The correction unit matches a case in which the greater than or equal to the average of the maximum value and the minimum value is '1' for the remaining pixels except for the center pixel in the N × N magnetic component mask, and the case where the small value is '0'. And the edge binary pattern of the N × N magnetic component mask is classified by corresponding consecutive matching values of the peripheral pixels in a predetermined direction from the peripheral pixels to binary values.
10. The method of claim 9,
The corrector is a pixel value average of neighboring pixels matched with '0' according to the edge binary pattern classified for the N × N magnetic component mask, a pixel value average of neighboring pixels matched with '1', and the N × N magnetic field And determining one of intermediate values of pixel values of all pixels in the component mask as a correction value of the target pixel.
A method of processing a defective pixel of an input image,
Receiving an M x M Bayer pattern including a target pixel for determining whether a defective pixel is included in the input image;
Extracting a magnetic component corresponding to the target pixel of the M × M Bayer pattern to generate an N × N magnetic component mask, where M and N are natural numbers and M>N;
Determining whether the target pixel is a defective pixel by analyzing characteristics of the N × N magnetic component mask;
Determining an edge binary pattern of the N × N magnetic component mask and performing correction when the target pixel is a defective pixel; And
Outputting the correction value corrected by the performing of the correction as the pixel value of the target pixel when the target pixel is a defective pixel, and outputting the pixel value of the target pixel as it is when the target pixel is a normal pixel. Defective pixel processing method comprising.
The method of claim 11,
The M x M Bayer pattern includes the target pixel as a center pixel.
The method of claim 11,
In the magnetic component extraction step, the pixels having the same color component as that of the target pixel among the pixels in the M × M Bayer pattern are selected to generate the N × N magnetic component mask having a corresponding relative position. A defective pixel processing method.
The method of claim 11,
The defective pixel determination step,
Calculating a reference value for determining whether there is a defective pixel by calculating a difference from a neighboring pixel value based on the center pixel in the N × N magnetic component mask;
Calculating a frequency difference value for inferring an amount of high frequency components included in the N × N magnetic component mask;
Calculating a detection threshold according to the frequency difference using a detection threshold graph;
And determining whether the target pixel is a defective pixel by comparing the reference value and the detection threshold value.
15. The method of claim 14,
The reference value calculation step,
Calculating an absolute value of a pixel value difference between each of the peripheral pixels of the N × N magnetic component mask and the center pixel;
And arranging them in ascending order and calculating an average of the first predetermined number as the reference values.
15. The method of claim 14,
And the frequency difference calculation step calculates a difference between a maximum value and a minimum value among pixel values in the N × N magnetic component mask as the frequency difference value.
15. The method of claim 14,
In the detecting threshold calculation step, the detection threshold value is calculated using a detection threshold graph according to the following equation.
[Mathematical Expression]

Here, dpc_thr is the detection threshold value, Freq_diff_value is the frequency difference value calculated by the frequency difference calculation module, dpc_freq_thr is a frequency threshold for distinguishing low and high frequencies, dpc_freq_width is the width interpolating the threshold of low and high frequencies , dpc_lf_thr is a defective pixel detection threshold for low frequency, dpc_hf_thr is a defective pixel detection threshold for high frequency, F1 is a first reference value for calculating the detection threshold calculated as a difference of 1/2 of dpc_freq_thr and dpc_freq_width, F2 is a second reference value for calculating the detection threshold calculated as a sum of 1/2 of dpc_freq_thr and dpc_freq_width.
15. The method of claim 14,
And the detection threshold graph is adjustable according to the luminance region to which the target pixel belongs.
The method of claim 11,
The correction step,
Matching a case in which the greater than or equal to the average of the maximum value and the minimum value is '1' for the remaining pixels except for the center pixel in the N × N magnetic component mask as '0';
And classifying an edge binary pattern of the N × N magnetic component mask by mapping a successive matching value of each peripheral pixel from a random peripheral pixel to a binary value in a predetermined direction.
20. The method of claim 19,
The correcting step may include: an average of pixel values of neighboring pixels matched with '0' according to the edge binary pattern classified for the N × N magnetic component mask, an average of pixel values of peripheral pixels matched with '1', and the N x N And determining one of intermediate values of pixel values of all pixels in a magnetic component mask as a correction value of the target pixel.
21. A recording medium on which a program of instructions executable by a digital processing apparatus for performing the defective pixel processing method according to any one of claims 11 to 20 is tangibly embodied and can be read by the digital processing apparatus.

Images