Digital Image Noise Processing Based On MATLAB Software
Digital Image Noise Processing Based On MATLAB Software
Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Digital Image Noise Processing Based on
MATLAB Software
Wang Mengyi
Department of Communication engineering
Shandong University of Technology
Shandong Province, China
Abstract:- On the basis of the analysis of digital image On the beauty of the whole photo. In order to make the
effect improvement basic theory research, a clearer picture more clear, it is necessary to eliminate the noise in the
processing algorithm of university image effect is picture and deal with the deformation of the picture. The
proposed to make the image more distinct and more operation of eliminating interference information and
quickly presented.By MATLAB software to write retaining relevant information is usually called image
relevant simulation code, respectively using the mean preprocessing, and the step of image noise reduction is also
filtering method and median filtering method to denoise called picture smoothing processing[3].
the image, so as to improve the image clarity, brightness,
finally the two methods are compared, in order to better There is a lot of noise in real life (such as grating
understand the theoretical knowledge. The software scanning, film particles, machine parts, signal propagation,
running results show that this method can achieve image etc.), which can be divided into many kinds, including Gauss,
denoising succinctly and accurately, and has high pepper salt, Rayleigh, etc. Generally, it can be removed by
learning and practical value. filtering method. General filtering methods include mean
method, median filtering method, adaptive, adjacent average,
Keywords:- Image Denoising; Image Definition; Filtering; selective average, low-pass filtering, etc[4].
MATLAB.
This paper is based on MATLAB simulation, by adding
I. INTRODCUTION Gaussian noise and salt and pepper noise to the original image
respectively, and then using the mean filter method and
Digital image technology is a new technology proposed median filter method respectively, to compare the effect of
under the support of a number of technical achievements and noise processing and draw a conclusion.
theoretical achievements. In the practical application process,
it relies on the functional support of the computer to transform II. NOISE ADDITION
the image information into two-dimensional digital signals, so
as to achieve various operations. Therefore, the performance The gray image polluted by Gaussian noise points has
of this technology cannot be separated from the support of the any noise position, but the gray value of this kind of noise
computer, which is also known as computer image points is randomly distributed. The gray image polluted by
processing.This technology includes the functions of pepper and salt noise points has random scattered noise
denoising, enhancing image sharpness, recognition, coding, positions and is shown in black and white dots on the screen,
restoration, segmentation, etc. With the continuous deepening but they have no connection with the gray value of the
of research in related fields, the technical performance is original image, usually the gray value of this kind of noise
continuously enhanced, and the restoration degree of image points and the gray value of nearby pixels have a big
reproduction is greatly improved. Based on the performance difference, in general is the extreme value[5].
advantages of this technology and the analysis of technical
advantages, the digital image processing technology has a For the processing of MATLAB image, the imnoise
more extensive application prospect in the future. However, function can be added to the noise, and its imperative
the indispensable part of life is the impact of noise on the parameter is g=imnoise (f, type, parameters), where g is the
image[1]. image after adding noise, f is the original image,type indicates
the type of the noise to be added, and parameters indicate
Noise has a great impact on image quality and makes it some parameters of the noise.
difficult to analyze the image. Image noise is a random
variation of brightness or color information in an image. It is III. NOISE REMOVAL
generally produced by the sensors and circuits of scanners or
digital cameras. It may also be caused by the inevitable loose A. Mean filtering method
grain noise in film particles or ideal photodetectors[2]. The Let the signal polluted by noise be g(t), the original
most immediate phenomenon is the blurring that occurs when signal be f(t), and the noise be n(t).g(t)=f(t)+n(t).The common
you zoom in.There is no doubt that these noises will have a expression of mean filtering is
certain impact
IJISRT23APR631 www.ijisrt.com 590
Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
(1) B.Code design of mean filter
Where: , is the length of the mean filter 5. The image with Gaussian noise and salt and pepper
noise is filtered by mean value. According to the filtering
principle of mean filter, the code is as follows:
B.Median filtering method
The core operation of median filtering is to sort the data in [height width]=size(G);
the template. In this way, if the noise of a bright spot is sorted, X1=G;
it will be placed on the most right or left side of the data X2=K;
sequence. Therefore, the value in the middle of the final for i=2:height-1
selected data sequence is generally not the value of the noise for j=2:width-1
point, so as to achieve the purpose of noise suppression[6]. sum1=uint16(0);
sum2=uint16(0);
IV. DESIGN OF CODE for i1=i-1:i+1
for j1=j-1:j+1
A.Code design with added noise sum1=sum1+uint16(G(i1,j1));
sum2=sum2+uint16(K(i1,j1));
Using the imnoise function, the code of adding Gaussian end
noise and salt and pepper noise on the basis of the original end
image is as follows: X1(i,j)=sum1/9;
X2(i,j)=sum2/9;
F=imread('E:\man.bmp');% Image path end
G=imnoise(F,'gaussian');% Gaussian noise end
K=imnoise(F,'salt & pepper',0.05);% salt and pepper noise
The code runs as follows:
The code running effect is shown in Figure 1:
(a) The original image (a) mean filtering of Gaussian noise
(b) After adding Gaussian noise
(b) mean filtering of salt and pepper noise
Fig 2 Image processed by mean filter
C.Code design of median filter
The image with Gaussian noise and salt and pepper
noise is filtered by mean value. According to the filtering
principle of mean filter, the code is as follows:
y1 = double(G);
(c)After adding salt and pepper noise y2 = double(K);
Y1=y1;
Fig 1 Images before and after adding noise Y2=y2;
for i = 2: height-1
IJISRT23APR631 www.ijisrt.com 591
Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
for j = 2:width-1 V. CONCLUSION
mb = y1( i-1:i+1,j-1:j+1 );
mb = mb(:); For Gaussian noise, the mean filtering effect is better
mm = median(mb); than the median filtering effect. According to its principle, a
Y1( i,j ) = mm; simple analysis can be made: Gaussian noise amplitude is
mb = y2( i-1:i+1, j-1:j+1 ); approximately normal distribution, but the distribution is on
mb = mb(:); every pixel. Because every point in the image is a polluted
mm = median(mb); point, the median filtering method cannot select a suitable
Y2( i,j ) = mm; clean point, because the mean value of positive distribution is
end 0, so the mean filtering can reduce the noise. For salt and
end pepper noise, the median filtering effect is better than the
mean filtering effect. The amplitude of salt and pepper noise
The code runs as follows: is approximately equal but randomly distributed in different
positions. There are clean points and polluted points in the
image. Median filtering is to select appropriate points to
replace the value of polluted points, so the processing effect is
good. Because the mean value of noise is not 0, mean filtering
cannot remove noise well.
VI. REFERENCES
[1]. Shang Y H. Application of Big data and digital image
processing technology [J]. Integrated Circuit
Applications,2021,38(10):90-91. (in Chinese)
(a) median filtering of Gaussian noise [2]. Yan Zhengguo, Lei Yu, Wang Haoran, et al.Research
on digital image watermarking [J]. Industrial Control
Computer, 20, 33 (11) : 100-102.( in Chinese)
[3]. Liu Jiamin, Peng Ling, Yuan Jiacheng, Liu
Junwei.Image denoising method based on two-
dimensional variational mode decomposition and
adaptive Median filtering [J]. Application Research of
Computers,2017.34(10): 349-3152(in Chinese)
[4]. Xiang Rui, Wang Li.Image denoising method based on
wavelet Transform [J]. Electronic Science and
Technology,2016.29(7):82-84,87. (in Chinese)
[5]. Yan Hongwen, Deng Xuefeng.Application of Median
Filter in digital mage denoising [J]. Computer Times,
(b) median filtering of salt and pepper noise 2020, 47-48. (in Chinese)
[6]. Zhu Hong. Foundation and Application of Digital
Fig 3 Image processed by median filter Image Processing [M]. Tsinghua University Press,
2013:30-33. (in Chinese)
