CN1633159A - A Method for Removing Image Noise - Google Patents

Abstract

This invention discloses a method for removing image noise including the following steps: A, acquiring the pixel data of the image B, using all gray values of said image to compute the mean gray value and its variance value, C, fetching the gray values of all pixels and judging if the gray value of each pixel is in the sphere of three times variance one by one, if so, the value of said pixel is not corrected, otherwise, the pixel is the noise removed by revising the gray value of said pixel.

Description

A Method for Removing Image Noise

technical field

The invention relates to image processing technology, in particular to a method for removing image noise.

Background technique

One of the most basic purposes of image processing is to improve image quality and provide a good premise environment for subsequent processing operations. Removing image noise is a relatively effective method to improve image quality. There are many reasons for the formation of noise, which may be generated during the imaging process or during the transmission process. The existence of noise will bring great inconvenience to subsequent image processing operations. Therefore, noise removal can be said to be necessary for all image processing. one step.

Currently, there are many methods for removing image noise. Mean filter smoothing image is a commonly used method to remove image noise. This method mainly uses the template operator to replace its own value with the mean value of the surrounding pixel values of a certain pixel, so as to achieve the purpose of removing noise and smoothing the image.

Referring to FIG. 1 , FIG. 1 is a schematic diagram of the principle of mean value filtering in the prior art. Wherein, the pixel value of pixel A is replaced by the average value of four pixels B, C, D, E around it, that is, the four pixels on the circle edge in FIG. 1 . Its specific processing process is referring to Fig. 2, and Fig. 2 is the flow chart that removes image noise with mean filtering mode in the prior art, and this flow process comprises the following steps:

Step 201, read the coordinate value and gray value data of each pixel of the image, and store them in the function f(x, y), wherein the horizontal and vertical coordinates of each pixel are stored, represented by x, y, and the value of each pixel is also stored A grayscale value (often also called a pixel value), represented by the value of f(x, y).

Step 202, traverse the entire image, calculate the new gray value of each pixel by formula (1), and store it.

${\mathrm{<span\; class="notranslate">\; f</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}{\stackrel{\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; no</span>}}{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}{\stackrel{\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, a, b are the horizontal and vertical coordinates of the pixel (x, y), and n is the step size.

Step 203, read the gray value of each pixel stored in the function f(x, y), and replace the gray value of each pixel with a new gray value according to the coordinates of each pixel, that is, use f( The value of x, y) is replaced by the value of f(x, y).

Since the essence of the mean filtering method in the prior art to remove image noise is to replace the pixel value of the noise with the mean value of the pixels around the noise, although the noise can be effectively removed, but after the image is processed in this way, the gray of adjacent pixels The brightness value may be relatively close, that is to say, the difference between the gray values of adjacent pixels is reduced, so the image may be blurred.

In addition, the above method for removing image noise uses a template operator to scan the image point by point, so the amount of calculation is large, and it needs to be calculated point by point, and a separate storage space is required to temporarily store intermediate data, which wastes system resources. At present, the usually selected step size n=1, that is, the average value is calculated in a 3*3 window. If the step size increases, the amount of calculation will increase linearly.

Contents of the invention

In view of this, the main purpose of the present invention is to provide a method for removing image noise, which can not only effectively remove image noise, but also reduce image blur caused by image noise removal processing.

In order to achieve the above object, the technical solution of the present invention is specifically realized in the following way:

A method for removing image noise, the method comprises the following steps:

A. Obtain each pixel data of the image;

B. Using the gray values of all pixels in the image, calculate the gray mean value and the gray variance value of the image;

C. Read the gray value of all pixels in the image, and judge whether the gray value of each pixel falls within 3 times the variance of the mean; if so, do not modify the gray value of the pixel; otherwise, the pixel is noise, passed Modify the gray value of the pixel to remove noise.

Wherein, the image may be the whole image or a region in the image.

The method for calculating the gray mean value of the image described in step B can be:

After summing the gray values of all pixels, find the average value.

The method for calculating the image grayscale variance value described in step B may include:

B1. For all pixels, find the gray difference between the gray value and the gray mean, and find the square of the gray difference;

B2. After summing the squares of the grayscale differences of all pixels, an average value is obtained, and the square root of the average value is obtained to obtain the grayscale variance value of the image.

The method for removing image noise by modifying the gray value described in step C can be:

For pixels whose grayscale value is greater than the grayscale mean plus 3 times the variance, the grayscale value is reduced;

For pixels whose gray value is less than the gray mean value minus 3 times the variance, the gray value is increased.

The method for removing image noise by modifying the gray value described in step C may specifically be:

Modify the grayscale value of the pixel whose grayscale value is greater than the grayscale mean plus 3 times the variance to the grayscale mean plus 3 times the variance;

Change the gray value of the pixel whose gray value is less than the gray mean minus 3 times the variance to the gray mean minus 3 times the variance.

The method for removing image noise by modifying the gray value described in step C can also be:

Predetermine an adjustment gray value;

Modify the gray value of the pixel whose gray value is greater than the gray mean value plus 3 times the variance to the original gray value minus the predetermined adjusted gray value;

Modify the gray value of the pixel whose gray value is less than the gray mean value minus 3 times the variance to the original gray value plus a predetermined adjusted gray value.

It can be seen from the above-mentioned technical solution that the method for removing image noise in the present invention utilizes the 3θ principle in the theory of probability and statistics, and removes pixels whose gray value falls outside the mean value by 3 times the variance as noise. Can effectively remove image noise.

Moreover, since the present invention only modifies the gray value of pixels whose gray value falls outside the range, instead of replacing all pixels with their respective calculated mean values as in the prior art, the present invention ensures that gray The gray value of the pixel whose intensity value falls within this range is not modified, thereby reducing the phenomenon of image blurring caused by image noise removal processing, and the calculation amount is small, which can save system resources.

Description of drawings

FIG. 1 is a schematic diagram of the prior art mean filter principle;

Fig. 2 is the flow chart of removing image noise with mean filtering mode in the prior art;

Fig. 3 is a processing flowchart of a preferred embodiment of the method for removing image noise of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and examples.

The method for removing image noise in the present invention utilizes the 3θ principle in the theory of probability and statistics, which characterizes that the overall majority information exists within 3 times the variance above and below the mean. Based on this theory, the image is regarded as a whole to calculate the gray mean and variance of the image, and by point-by-point scanning pixel comparison, the pixels whose gray value falls outside the mean value by 3 times the variance are regarded as noise, and their gray value is regarded as noise. The value is modified to the mean plus/minus 3 times the variance to eliminate noise.

Referring to FIG. 3 , FIG. 3 is a processing flowchart of a preferred embodiment of the method for removing image noise according to the present invention. The process includes the following steps:

Step 301, read and store data such as the gray value and coordinate value of each pixel of the image.

In this step, the storage method can be the same as that of the prior art, that is, store in the function f(x, y). Of course, it can be stored in other ways, as long as the gray value and coordinate value of each pixel can be stored correspondingly.

Step 302, read the grayscale values of all pixels in the image, and calculate the grayscale mean μ. The gray mean value μ can be calculated according to the formula (2):

$\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Step 303, calculating the variance θ with the gray mean value. It can be calculated first according to formula (3):

${\mathrm{<span\; class="notranslate">\; \&theta;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Among them, the average value of x is μ.

Then the value of the variance θ is calculated by the square root operation.

Step 304, read the gray value of a pixel.

Step 305, judging whether the gray value falls within the range of [μ-3θ, μ+3θ], if yes, execute

Step 307; otherwise, execute step 306.

Step 306, if the gray value is less than μ-3θ, replace the gray value with μ-3θ; if the gray value is greater than μ+3θ, replace the gray value with μ+3θ.

The processing process of steps 305-306 can be expressed by formula (4).

$\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}& \mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\\ \mathrm{<span\; class="notranslate">\; p</span>}& \mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\\ \mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}& \mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

where p represents the gray value of each pixel.

The essence of steps 305-306 is: decrease the gray value of the pixel whose gray value is greater than μ+3θ; increase the gray value of the pixel whose gray value is smaller than μ−3θ. Therefore, other ways are also possible.

For example: Predetermine an adjusted gray value; modify the gray value of a pixel whose gray value is greater than μ+3θ to the original gray value minus the predetermined adjusted gray value; change the gray value of a pixel whose gray value is smaller than μ-3θ The value is modified to the original gray value plus a predetermined adjusted gray value.

Step 307, judging whether there are unread pixels, if so, return to step 304, and read the next pixel; otherwise, end the processing flow.

The present invention can implement denoising processing on the whole set of images, and can also select a certain area in the image to implement denoising processing as required. If the denoising process is performed on the entire image, the image described in the above process is the entire image; if the denoising process is performed on a certain area, the image described in the above process is the selected one. image area. The process is exactly the same, only the size of the range being processed is different.

It can be seen from the above-mentioned embodiments that the method for removing image noise in the present invention utilizes the 3θ principle in the theory of probability and statistics, and removes pixels whose gray value falls outside the mean value by 3 times the variance as noise. Can effectively remove image noise.

Moreover, since the present invention only modifies the gray value of pixels whose gray value falls outside the range of [μ-3θ, μ+3θ], instead of using their respective calculated mean values for all pixels as in the prior art Instead, the present invention ensures that the gray values of pixels whose gray values fall within the range are not modified, thereby reducing image blurring caused by image noise removal.

In addition, the present invention only executes two operations of formula (2) and (3) on all pixels of the image, and executes formula (4) by way of comparison, which has a small amount of calculation, simple and convenient processing method, and can directly calculate the grayscale of the original image. Values can be modified without additional storage space, saving system resources.

Claims (7)

1. A method for removing image noise, characterized in that the method may further comprise the steps: A. Obtain each pixel data of the image; B. Using the gray values of all pixels in the image, calculate the gray mean value and the gray variance value of the image; C. Read the gray value of all pixels in the image, and judge whether the gray value of each pixel falls within 3 times the variance of the mean; if so, do not modify the gray value of the pixel; otherwise, the pixel is noise, by Modify the gray value of the pixel to remove noise. 2. The method according to claim 1, characterized in that: the image is the whole image or a region in the image. 3. The method according to claim 1, wherein the method for calculating the gray mean value of the image in step B is: After summing the gray values of all pixels, find the average value. 4. The method according to claim 1 or 3, characterized in that the method for calculating the image grayscale variance value described in step B comprises: B1. For all pixels, find the gray difference between the gray value and the gray mean, and find the square of the gray difference; B2. After summing the squares of the grayscale differences of all pixels, an average value is obtained, and the square root of the average value is obtained to obtain the grayscale variance value of the image. 5. The method according to claim 1, characterized in that the method for removing image noise by modifying the gray value in step C is: For pixels whose grayscale value is greater than the grayscale mean plus 3 times the variance, the grayscale value is reduced; For pixels whose gray value is less than the gray mean value minus 3 times the variance, the gray value is increased. 6. The method according to claim 5, characterized in that the method for removing image noise by modifying the gray value in step C is: Modify the grayscale value of the pixel whose grayscale value is greater than the grayscale mean plus 3 times the variance to the grayscale mean plus 3 times the variance; Change the gray value of the pixel whose gray value is less than the gray mean minus 3 times the variance to the gray mean minus 3 times the variance. 7. The method according to claim 5, wherein the method for removing image noise by modifying the gray value in step C is: Predetermine an adjustment gray value; Modify the gray value of the pixel whose gray value is greater than the gray mean value plus 3 times the variance to the original gray value minus the predetermined adjusted gray value; Modify the gray value of the pixel whose gray value is less than the gray mean value minus 3 times the variance to the original gray value plus a predetermined adjusted gray value.

Images