CN1301621C - Method and device for dynamically detecting pixel values - Google Patents

Abstract

The method and the device for dynamically detecting the pixel value receive a pixel value array, and the pixel value array has the pixel value to be detected. And subtracting the values of the reference pixels and the values of the pixels to be detected respectively to obtain a plurality of difference values. The sampling pixel of the reference pixel value is a sampling pixel which is positioned near the pixel value to be detected, and the reference pixel value and the pixel value to be detected represent the same color. Comparing the difference value with the reference value, and judging that the pixel value to be detected is wrong when the absolute value of the difference value is greater than the reference value.

Description

Method and device for dynamically detecting pixel value

technical field

The present invention relates to a digital color image sensing device, and more particularly to a method and device for dynamically judging the correctness of pixel values and correcting them in real time.

Background technique

Most of the current digital color photography systems use sensors, such as Charge-Coupled Devices (CCD) or Complementary Metal-Oxide Semiconductor (CMOS) sensors, for image sampling. Work. In general, sensing devices utilize the principle that photons will excite electrons (and holes) to hop between different energy levels, and convert the sensed incident light into electrical signals.

In order to detect light of different colors, different sampling pixels must be prepared for red light, blue light and green light respectively. Take the most commonly used color filter array (Bayer ColorFilter Array, CFA) that can effectively reduce the size of the sensing device as an example, each sampling pixel is only responsible for one color, such as full color (ie red, green, one-third of the blue tricolor). In order to complement the complete color structure of each sampling pixel, the sensing device must perform color interpolation (Color Interpolation) to obtain the other two colors that each sampling pixel lacks through interpolation to improve resolution.

In practice, some sampled pixels, such as pixels with defects, may obtain wrong pixel values. If these erroneous pixel values are not further corrected, the image quality will be affected. The simplest traditional correction processing method is to test the sensing device before leaving the factory, record the position of each defective pixel, and add additional correction to the corresponding pixel value in the subsequent numerical processing.

However, for example, new defective pixels that have been generated for a long time, or wrong pixel values obtained due to dust covering the lens, or defective pixel positions that are different from the original ones due to replacement of sensing devices, etc. The existing correction processing method cannot correct the factors that exist when the test components leave the factory.

US patent application 2004/0032516 A1 discloses a "Digital Image System and Method Combining Demosaicing and Bad Pixel Correction", which uses dynamic detection to correct the above-mentioned erroneous pixel values, and can be combined with color interpolation. method to process all pixel values together. However, the prior art disclosed in this US patent application has a complex circuit structure and requires huge calculation processing, which not only requires high-speed system computing capability to support, but also increases the cost burden of design and manufacture. In addition, the images obtained by this prior art tend to be too blurred and distorted visually due to too many numerical operations.

Contents of the invention

An object of the present invention is to provide a method for dynamically detecting pixel values, which is used to dynamically detect and correct wrong pixel values in real time, and avoid the problem of distortion caused by complex numerical calculations.

Another object of the present invention is to propose a system for dynamically detecting pixel values, which has a simple circuit structure and does not require huge calculation processing, and can be combined with the original color interpolation calculation device to reduce design and manufacturing costs.

According to the present invention, a method and device for dynamically detecting pixel values are provided. The pixel value array is stored in the register, and the pixel value array has the pixel values to be detected. A subtracter is used to subtract the multiple reference pixel values from the pixel values to be detected to obtain multiple difference values. The sampling pixels of the value of the reference pixel are the sampling pixels surrounding the value of the pixel to be detected, and the value of the reference pixel is the same as the color represented by the value of the pixel to be detected. A comparator is used to compare the difference with a reference value, and when the absolute values of the differences are greater than the reference value, it is determined that the value of the pixel to be detected is wrong.

According to a preferred embodiment of the present invention, the device for dynamically detecting pixel values further includes an adder and a divider. When the value of the pixel to be detected is wrong, the average value of the reference pixel value is used to replace the value of the reference pixel. The value of the pixel to be detected, and when one of the absolute values of the difference is not greater than the reference value, the value of the pixel to be detected is retained.

According to a preferred embodiment of the present invention, the distances between the sampling pixels of the reference pixel value and the sampling pixels of the pixel value to be detected are equal, and the sampling pixels of the reference pixel value surround the sampling pixels of the pixel value to be detected. Furthermore, the dynamic detection pixel value device is combined with the color interpolation operation device.

The method and device for dynamically detecting pixel values of the present invention compare the difference obtained by subtraction with a reference value to determine whether the pixel value is wrong. Furthermore, the present invention can be combined into the original color interpolation operation device of the sensing device, without complicated additional circuit improvement, which can effectively reduce the difficulty and cost of design and manufacture. In addition, compared with the image obtained in the prior art, the image processed by the present invention does not undergo much numerical calculation, which solves the problem of existing image distortion.

Description of drawings

Fig. 1A is a method flowchart of a preferred embodiment of the present invention;

Figure 1B is a schematic diagram of a device according to a preferred embodiment of the present invention; and

FIG. 2 is a schematic diagram of a pixel value array in a preferred embodiment of the present invention.

Detailed ways

FIG. 1A is a flow chart of a method of a preferred embodiment of the present invention, and FIG. 1B is a schematic diagram of an apparatus of a preferred embodiment of the present invention.

First, with a sensing device, such as a charge-coupled device (CCD), a complementary metal-oxide-semiconductor (CMOS) sensing device, or a color filter array sampling sensing device, the sensed incident light is Color sampled pixels are converted to multiple raw pixel values. Of course, if the light is monochromatic or black-and-white images are to be made, only one color sampling pixel or sampling pixels that cannot distinguish colors can be used. Then, according to the spatial distribution of each sampled pixel, part or all of the pixel values are stored in a register 132 to form a pixel value array (step 102 ).

When checking whether any of the plurality of sampling pixels included in the pixel value array is a bad (or faulty) sampling pixel, each pixel can be detected one by one. For example, starting from the upper left of the entire pixel value array, each pixel is gradually detected row by row and column by row. When detecting a certain pixel, the entire array of pixel values can be regarded as a value of a pixel to be detected and a plurality of reference pixel values. The sampling pixels of the value of the reference pixel are the sampling pixels surrounding the value of the pixel to be detected, and the value of the reference pixel is the same as the color represented by the value of the pixel to be detected. For example, if the color represented by the value of the pixel to be detected is red, then the colors represented by the value of the reference pixel selected from the pixel value array must also be red. Next, subtract the value of the reference pixel from the value of the pixel to be detected by the subtractor 134 to obtain a plurality of difference values respectively corresponding to each reference pixel value (step 104 ).

The difference is then compared with a reference value using a comparator 136 (step 106). When the absolute values of the differences are greater than the reference value, it is determined that the value of the pixel to be detected is wrong (step 112 ). At this point, if the wrong value of the pixel to be detected is to be further repaired, the value of the pixel to be detected can be replaced by the average value of the reference pixel values by the adder 142 and the divider 144 (step 114 ). On the other hand, when one of the absolute values of the difference is not greater than the reference value, it is determined that the value of the pixel to be detected is correct (step 122 ). At this time, keep the value of the pixel to be detected, and continue to detect the value of the next pixel to be detected (step 124). At this time, it is possible to use the detected pixel values that are determined to be normal, or even use the repaired original error detection pixel values to process other pixel values to be detected that will be processed later. Moreover, the reference value is an adjustable parameter, which can be adjusted according to the required picture quality or the performance of the sensing device, etc., and even different reference values can be used in different parts of the same pixel value array.

FIG. 2 is a schematic diagram of a pixel value array of a preferred embodiment of the present invention. The detection process of this preferred embodiment will be described below using FIG. 2 to further explain the content of the present invention in detail and clearly.

As shown in FIG. 2 , the pixel value array 200 is an array of 5×5 pixel values, and the arrangement of the pixel values is spatially a Bayer Color Filter Array. In the pixel value array 200 , R is a pixel value representing red, G is a pixel value representing green, and B is a pixel value representing blue.

Firstly, _R5 in the pixel value array 200 is selected as the pixel value to be detected, and the other four _R2 , _R4 , _R6 and _R8 of the same color closest to and surrounding _R5 are selected as the reference pixel value. The reference pixel values R ₂ , R ₄ , R ₆ , and R ₈ are subtracted from the pixel value R ₅ to be detected to obtain a plurality of difference values, and then the difference values and the reference value are compared.

As mentioned above, when the absolute value of the difference is greater than the reference value, it is determined that the value of the pixel to be detected is wrong, as shown in the inequalities (1), (2), (3) and (4):

|R ₂ -R ₅ |＞Reference value (1)

|R ₄ -R ₅ |＞Reference value (2)

|R ₆ -R ₅ |＞Reference value (3)

|R ₈ -R ₅ |＞Reference value (4)

That is to say, when the above inequalities (1), (2), (3) and (4) are all true, it is determined that the value R5 of the pixel to be detected is wrong. At this time, the original value R5 of the pixel to be detected is replaced by the average value R5 ′ of the reference pixel values R2 , R4 , R6 and R8 . The average value R5' is shown in equation (5).

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 5</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 6</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 8</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

On the other hand, when one of the absolute values of the difference is not greater than the reference value, it is determined that the value of the pixel to be detected is correct. At this time, keep the value of the pixel to be detected, and continue to detect the value of the next pixel to be detected.

In this preferred embodiment, four other R ₂ , R ₄ , R ₆ and R ₈ of the same color closest to and surrounding the pixel value R ₅ to be detected are selected as the reference pixel value. However, other embodiments of the present invention can also use another four R ₁ , R ₃ , R ₇ and R ₉ of the same color but farther away from R ₅ as reference pixel values; or even use 7×7 pixels The outermost reference pixel value in the array of values. The focus of the present invention is to apply reference pixel values of the same color around the pixel value R ₅ to be detected, without limiting the reference pixel values used.

Further, to obtain more accurate judgment results and/or better average values, more pixel values of the same color can be selected within the range that the hardware computing power can load, for example, comprehensive use of pixel values R ₁ and R ₂ , R ₃ , R ₄ , R ₆ , R ₇ , R ₈ and R ₉ are used together as reference pixel values for judging and/or calculating the average value.

In addition, since this device uses other pixel values of the same color near the value of the pixel to be detected as the reference pixel value when judging and calculating the average value, the present invention compares the value of the pixel to be detected at the edge of the sensing device In a preferred embodiment, several rows and columns of dummy pixel values can be reserved at the edge of the sensing device without performing calculations, so as to solve the problem that the pixel values to be detected at the edge of the sensing device lack enough available pixels on the edge side. Refer to the problem of pixel value.

To sum up, the dynamic detection pixel value device of the present invention has a simple circuit and easy operation, and can easily share the same set of circuits with the original color interpolation operation device of the general sensing device. In particular, since the operation of the present invention does not need to record the positions of defective pixels in advance, even if the sensing device is replaced, new defective pixels appear in the sensing device, or dust is attached to the sensing device, the application of the present invention will not be affected. Influence.

Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person familiar with the art may make various equivalent changes or substitutions without departing from the spirit and scope of the present invention. , so the scope of protection of the present invention shall prevail as defined by the appended claims of the application.

Claims (11)

1. the method for a detection of dynamic pixel number comprises at least:

Receive a pixel number array;

Choose a pixel number to be detected from this pixel number array;

Respectively a plurality of reference pixel numerical value and this pixel number to be detected are subtracted each other to draw a plurality of differences, wherein the pixels sampled of this reference pixel numerical value is the pixels sampled on every side that is positioned at this pixel number to be detected, and this reference pixel numerical value is identical with the color of this pixel number representative to be detected; And

When the absolute value of this difference during, judge that this pixel number to be detected is for wrong all greater than a reference value.

2. the method for claim 1 is characterized in that when this pixel number to be detected is mistake, with this pixel number to be detected of mean value replacement of this reference pixel numerical value.

3. the method for claim 1 when it is characterized in that one of them is not more than this reference value when the absolute value of this difference, keeps this pixel number to be detected.

4. the method for claim 1 is characterized in that it being to allow the distance between the pixels sampled of pixels sampled and this pixel number to be detected of this reference pixel numerical value all equate.

5. the method for claim 1, the pixels sampled that it is characterized in that this reference pixel numerical value is that the pixels sampled on every side that is positioned at this pixel number to be detected further comprises: be positioned at top, below, left and at least four right-hand pixels sampled of this pixel number to be detected, or be positioned at least four pixels sampled of upper right side, lower right, upper left side and the lower left of this pixel number to be detected.

6. the method for claim 1, it is characterized in that can also be at the edge of this pixel number array the dummy pixel numerical value of pre-residual row and ordered series of numbers do not do the processing of pixel number to be detected.

7. the device of a detection of dynamic pixel number comprises at least:

One register stores a pixel number array, and wherein this pixel number array has a pixel number to be detected;

One subtracter, respectively a plurality of reference pixel numerical value and this pixel number to be detected are subtracted each other to draw a plurality of differences, wherein the pixels sampled of this reference pixel numerical value is the pixels sampled a plurality of pixels sampled on every side that are positioned at this pixel number to be detected, and this reference pixel numerical value is identical with the color of this pixel number representative to be detected; And

One comparator, relatively this difference and a reference value when the absolute value of this difference during all greater than this reference value, judge that this pixel number to be detected is for wrong.

8. device as claimed in claim 7, it is characterized in that this device also comprises an adder and a divider, when this pixel number to be detected when being wrong, replace this pixel number to be detected with the mean value of this reference pixel numerical value by this adder and this divider.

9. device as claimed in claim 7, when it is characterized in that one of them is not more than this reference value when the absolute value of this difference, this comparator keeps this pixel number to be detected.

10. device as claimed in claim 7, the pixels sampled that it is characterized in that this reference pixel numerical value all equates with distance between the pixels sampled of this pixel number to be detected.

11. device as claimed in claim 7, the pixels sampled that it is characterized in that this reference pixel numerical value is that the pixels sampled on every side that is positioned at this pixel number to be detected further comprises: be positioned at top, below, left and at least four right-hand pixels sampled of this pixel number to be detected, or be positioned at least four pixels sampled of upper right side, lower right, upper left side and the lower left of this pixel number to be detected.

Images