APPARATUS AND METHOD FOR IMPROVING IMAGE QUALITY OF IMAGE SENSOR

• US 20090262223A1
• Filed: 06/30/2009
• Published: 10/22/2009
• Est. Priority Date: 11/01/2005
• Status: Active Grant

• Current Assignee: Intellectual Ventures II LLC (Intellectual Ventures LLC)
• Original Assignee: Crosstek Capital LLC (Intellectual Ventures LLC)
• Inventors: Hwang, Sung-Hyun
• Granted Patent: US 8,199,216 B2
• Time in Patent Office: Days
• US Class Current: 348/241
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/20012 Locally adaptive
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 3/4015 Image demosaicing, e.g. colour filter arrays [CFA] or Bayer patterns
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: H04N 1/58 Edge or detail enhancement; Noise or error suppression, e.g. colour misregistration correction H04N1/62 takes precedence
• CPC Class Codes: H04N 9/68 for controlling the amplitude of colour signals, e.g. automatic chroma control circuits H04N9/71, H04N9/73 take precedence; camera processing pipelines for controlling the colour saturation of colour signals H04N23/86

Abstract

Provided are an apparatus and method for improving an image quality of an image sensor, capable of adaptively removing noise occurring in a de-mosaicking that is performed for generating three-channel data of R, G and B from a single channel pixel structure based on a bayer pattern. The apparatus includes a first converting unit for converting RGB color data into a YCbCr color data, the RGB color data being obtained from bayer data through a de-mosaicking, a noise removal unit for removing noise from a Cb and a Cr color data outputted from the first converting unit, and a second converting unit for converting the Cb and the Cr color data from the noise removal unit and a Y data from the first converting unit into the RGB color data.

Claims

1. An apparatus for improving an image quality of an image sensor, the apparatus comprising:

a first converting unit for converting an RGB color data obtained from bayer data through a de-mosaicking into a YCbCr color data;

a noise removal unit for removing a noise of a Cb and a Cr color data outputted from the first converting unit, and for determining a threshold value to be used during a filtering process by using data located at a center from the Cb and the Cr color data outputted from the first converting unit; and

a second converting unit for converting Cb and Cr color data outputted from the noise removal unit, and a Y color data from the first converting unit, into the RGB color data,

wherein, when a statistical characteristic of noise added to adjacent pixels follows a Gaussian distribution, the adjacent pixels being disposed within a predetermined range that is determined by a threshold value for centering on a center pixel, are smoothed, the threshold value being a function of a standard deviation of the noise adjacent to the adjacent pixels.

2. The apparatus as recited in claim 1, wherein the first converting unit converts the RGB color data into the YCbCr color data using a matrix-vector product.

3. The apparatus as recited in claim 1, wherein the Cb and the Cr color data outputted from the first converting unit include color information.

4. The apparatus as recited in claim 1, wherein the filtering is performed by a one-dimensional filter that includes:

a register for sequentially storing the Cb and the Cr color data outputted from the first converting unit;

a threshold generator for determining the threshold value to be used during the filtering process by using the data located at the center among the data stored in the register;

a data selector for selecting the output data of the register using the threshold value outputted from the threshold generator; and

a linear filter for removing a noise from the Cb and the Cr color data selected by the data selector.

5. The apparatus as recited in claim 4, wherein the register sequentially receives the Cb and the Cr color data from the first converting unit to output the Cb and the Cr color data in a pipeline form.

6. The apparatus as recited in claim 5, wherein the threshold generator operates in response to an enable signal and determines the threshold value to be used during the filtering process by using an input value of the data positioned at the center among the data stored in the register.

7. The apparatus as recited in claim 4, wherein the data selector selects the data sequentially inputted from the register by using the threshold value determined by the threshold generator.

8. The apparatus as recited in claim 4, wherein the linear filter includes:

a multiplexer for selecting a weight for smoothing with respect to the data selected by the data selector in response to a select signal;

a multiplier for multiplying the selected weight by the data selected through the data selector;

a first accumulator for sequentially adding data outputted from the multiplier;

a second accumulator for sequentially adding the weights selected by the multiplexer; and

a divider for dividing the output data of the first accumulator by output data of the second accumulator for normalization.

9. The apparatus as recited in claim 8, wherein the multiplexer selects a Gaussian weight or a unit weight according to the select signal.

10. The apparatus as recited in claim 9, wherein the unit weight is a weight for an arithmetic average.

11. The apparatus as recited in claim 8, wherein the select signal is inputted from a source external to the image sensor.

12. The apparatus as recited in claim 8, wherein the multiplier multiplies the weight selected by the multiplexer by the data selected by the data selector to be sequentially inputted.

13. The apparatus as recited in claim 1, wherein the apparatus is disposed within at least one of a digital camera or an apparatus disposed to perform video tracking.

14. The apparatus as recited in claim 1, wherein the apparatus is disposed within a mobile terminal including a built-in camera.

15. The apparatus as recited in claim 1, wherein the image sensor is a complementary metal oxide semiconductor image sensor.

16. A method for improving an image quality of an image sensor, the method comprising:

converting an RGB color data into a YCbCr color data, the RGB color data being obtained from a bayer pattern through a de-mosaicking;

removing a noise from a Cb and a Cr color data among the data converted into the YCbCr color data and determining a threshold value to be used during a filtering process by using data located at a center from the Cb and the Cr color data; and

converting, the Cb and the Cr color data outputted after removing the noise from the Cb and the Cr color data, and Y data, into the RGB color data,

wherein, when a statistical characteristic of noise added to adjacent pixels follows a Gaussian distribution, the adjacent pixels being disposed within a predetermined range that is determined by a threshold value for centering on a center pixel, are smoothed, the threshold value being a function of a standard deviation of the noise adjacent to the adjacent pixels.

17. The method as recited in claim 16, wherein the RGB color data is converted into the YCbCr color data using a matrix-vector product.

18. The method as recited in claim 17, wherein the predetermined range is provided from a first threshold value to a second threshold value, wherein a magnitude of the first threshold value and the second threshold value are the same and the first threshold value is a negative of a second threshold value.

19. The method as recited in claim 17, wherein the removing of the noise further comprises:

performing a smoothing using the determined threshold value, the smoothing being performed on a boundary region by smoothing similar values to a value of the center pixel.

20. The method as recited in claim 16, wherein the removing of the noise is performed using a one-dimensional filter or a two-dimensional filter.

21. The method as recited in claim 16, wherein the removing of the noise from the Cb and the Cr color data includes:

storing the Cb and the Cr color data sequentially in a register;

determining the threshold value to be used during the filtering process by using the data located at the center among the data stored in the register;

selecting the data stored in the register using the threshold value; and

removing noise from the selected data.

22. The method as recited in claim 21, wherein the removing of the noise from the selected data includes:

selecting a weight for smoothing with respect to the selected data in response to a select signal;

multiplying the selected weight by the selected data;

summing sequentially the multiplied data;

summing sequentially the selected weights; and

dividing the sequentially summed multiplied data by the sequentially summed weight, and quantifying the sequentially summed multiplied data.

23. The method as recited in claim 16, wherein the steps of the method are performed in a video tracking apparatus or a camera.

24. A computer program product executing computer-executable instructions to perform the following acts:

removing a noise from a Cb and a Cr color data among data converted into YCbCr color data and determining a threshold value to be used during a filtering process by using data located at a center from the Cb and the Cr color data,

wherein, when a statistical characteristic of noise added to pixels adjacent to Cb and Cr color data follows a Gaussian distribution, the adjacent pixels being disposed within a predetermined range that is determined by a threshold value for centering on a center pixel, are smoothed, the threshold value being a function of a standard deviation of the noise adjacent to the adjacent pixels.

25. An apparatus for improving an image quality of an image sensor, the apparatus comprising:

a means for removing a noise from a Cb and a Cr color data among data converted into YCbCr color data and determining a threshold value to be used during a filtering process by using data located at a center from the Cb and the Cr color data,

wherein, when a statistical characteristic of noise added to pixels adjacent to Cb and Cr color data follows a Gaussian distribution, the adjacent pixels being disposed within a predetermined range that is determined by a threshold value for centering on a center pixel, are smoothed, the threshold value being a function of a standard deviation of the noise adjacent to the adjacent pixels.