International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING & ISSN

0976 - 6375(Online), Volume 4, Issue 6, November - December (2013), IAEME

TECHNOLOGY (IJCET)

ISSN 0976 6367(Print) ISSN 0976 6375(Online) Volume 4, Issue 6, November - December (2013), pp. 121-126 IAEME: www.iaeme.com/ijcet.asp Journal Impact Factor (2013): 6.1302 (Calculated by GISI) www.jifactor.com

IJCET
IAEME

APPLICATION OF MEDIAN FILTER WITH THE THRESHOLD TECHNIQUE TO REDUCE AND REMOVE GAUSSIAN NOISE ON THE IMAGE EDGES PRODUCED BY SOBEL OPERATOR
H.A. Alshamarti1*,
1

Ali K. Hussein2,

B.A. Almayahi3

Department of Physics, College of Science, University of Kufa, Iraq 2 College of Dentistry, University of Kufa, Iraq 3 Department of Environment, College of Science, University of Kufa, Iraq

ABSTRACT In this paper, a new method to remove Gaussian noise on the image edges produced by Sobel operator is designed. The mean filter was used in literatures to removes or reduces Gaussian noise, but this filter is not enough. Therefore, in this work median filter is added with the function of threshold on the image edges, which it filtered by mean filter for clear the image using MATLAB software. The comparison between the treatment image edges is conducted using Root Mean Square Error (RMSE). Keywords: Gaussian Noise, Sobel Operator, Edge Detection, Threshold Function. 1. INTRODUCTION The process of image may generate images without quality due to mechanical problems, out of focus blur, motion, illumination unsuitable, and noises. The different procedures related to the types of noise are introduced to the image. There are many noises: Gaussian or White, Rayleigh, Shot or Impulse, periodic, sinusoidal or coherent, uncorrelated, and granular (Gonzalez & Woods 2004). Image processing algorithms tend to perform worse when operating on images with noise. Therefore, it is necessary to employ processing noise to reduction filters, which it product much of the original image details (Azzam et al. 2008). This paper aims to removal of the Gaussian noise presented on the image edges. 1.1. Noise Models The principal source of noise in digital images arises during the image acquisition (digitization) and transmission. The performance of image sensors is affected by a variety of factors
121

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 4, Issue 6, November - December (2013), IAEME

such as environmental conditions during image acquisition and by the quality of the sensing elements. In acquiring images with a CCD camera, light levels, and the sensor temperature are major factors affect the noise in the image (Senthilkumarn & Rajesh 2009). Removal of noise is important for most of the digital imaging applications (Gonzalez 2004). Some of the noise removals algorithms are require prior knowledge about the noise in the image. The standard noise types (Gaussian, speckle, salt, and paper) can be expressed in terms of noise variance or standard noise deviation (Kutty & Ojha 2012). 1.2. Smooth Image All in the case of Additive White Gaussian noise (AWGN) and all the image pixels deviate from their original values following the Gaussian curve. The probability density function (PDF) for a zero mean Gaussian distribution is (Gajanand 2011):

(1)

where z= represents gray level, = the mean of average value of z, and = standard deviation. The standard deviation ( 2)= the variance of z. For each image pixel with intensity value Iij (1 i m, 1 j n; for the image (m * n)), the corresponding pixel of the noisy image Nij is Nij = Iij + Gij (2)

where each noise value (G) is drawn from a zero-mean Gaussian distribution. The main aim of image smoothing is to remove noise in digital images. It is a classical matter in digital image processing to smooth image. It has been widely used in many fields, such as image display, image transmission and image analysis,.etc. Image smoothing is a method of improving the quality of images. Because image smoothing is a classical matter, many filters come into practice based on the practical requirement and the development of related technology (Keiji 2001). An averaging filter is useful for removing noise from an image. Because each pixel is set to the average of the pixels in its neighborhood and local variations caused by grain are reduced (Yong & Kassam 1985). Median filtering is similar the average filter, except that the value of an output pixel is determined by the median of the neighborhood pixels, rather than the mean (Chen et al. 1999). 1.3. Mean and Median Filters The Mean Filter is a linear filter, which it uses a mask over each pixel in the signal. Each of the components of the pixels, which fall under the mask are averaged together with form a single. The Mean filter is defined (Padmavathi et al. 2009): Mean ilter

(3)

where is the image pixel range. The neighboring pixels are ranked according to brightness (intensity) and the middle value (median value) becomes the new value for the central pixel. Its can do an good job of rejecting certain types of noise, in particular, shot or impulse noise in which some individual pixels have extreme values (Ko & Lee 1991).

122

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 4, Issue 6, November - December (2013), IAEME

1.4. Edge Detectors Types Edge detectors can be classified into two broad categories (Senthilkumarn & Rajesh 2009; Ravi & Khan 2012; Alshamarti 2013): 1. First Derivative Operators a. Roberts operator b. Prewitts operator c. Sobels operator 2. Second Derivative Operators a. Laplacian of Gaussian Operator b. Canny Edge Detector 1.5. Sobel Operator The kernels can be applied to the input image to produce separately measurements of the gradient component in each orientation (Sx and Sy). These can be combined to find the absolute magnitude of the gradient at each point and the orientation of gradient. The masks used to convolute Sobel operator are: - - 1 2 1 0 0 0 1 2 1 Row Mask (SX) 0 1 1 0 2 2 0 1 1 Column Mask (Sy)

The Sobel operator is the magnitude of gradient and can be calculated (Maarten 2001):
M

(4)

1.6. Threshold Technique Threshold processing aims to remove fine fragments mixed with objects. Large lump ores in original images are often mixed with fine sands and rocks, which have similar illumination reflection and texture. It makes getting continue boundary of objects with very difficult ordinary edge detection algorithms. Remove most of fine fragments as background and make edge detection algorithms focus on objects (Maarten 2001). Threshold is one of the widely methods used for enhancement the image edge. It is useful in discriminating foreground from the background. Select suitable threshold value (T) and the image gray-level can be converted to binary image. The first binary image reduces the complex of data and simplifies the process of recognition and classification. The common way to convert an image gray-level to a binary image is to select a single threshold value (T). Then all the gray-level values below T will be black (0) and above T will be white (1) (Salem et al. 2010). In this paper the all gray level values below T will be classified as black (0) and above T will be white (z). 1.7. Algorithm The fundamental steps in algorithm application are: a. The smooth of Image: The application mean filter for primarily reduction of noise Gaussian. b. Detection of edges: Local operations that select all the possible edges in the image and select the true edges from the list of the possible edges using Sobel operator. c. Final enhancement: It uses threshold technique and median filter, where there is a very well removal of noise.
123

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 4, Issue 6, November - December (2013), IAEME

2. RESULTS The well-known Lenna image was used as shown in Fig. 1, the test image corrupted by different amount of Gaussian noise and convolution by Sobel operator as shown in Fig. 2. The mean filter and edge detection are used as shown in Fig. 3. The results of the enhancement image edge after applying the threshold technique and median filter are shown in Fig. 4.

(a) (b) (c) Figure 1. (a) original test image (b) with sobal operator (c) threshold technique

=0.004

=0,008

=0.012

=0.016

=0.02

RMSE =44.37 RMSE =58.40 RMSE =68.50 RMSE =77.04 RMSE = 82.36 Figure 2. Lenna images corrupted by different Gaussian noises with edge detection (Sobal operator 3*3)

RMSE =27.88 RMSE =32.22 RMSE =36.14 RMSE =39.33 RMSE =41.86 Figure 3. Mean filter (Fig. 2) and down images represent edge detection using sobal operator 3*3
124

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 4, Issue 6, November - December (2013), IAEME

RMSE =10.83 RMSE =11.75 RMSE =13.15 RMSE =14.82

RMSE =16.26

RMSE =12.94 RMSE =13.306 RMSE =13.81 RMSE =14.55 RMSE =14.89 Figure 4. Removal of Gaussian noise after applying median filter + threshold technique with T= 128

The RMSE for Gaussian noise at =0.02:0.004:0.04 as shown in Table 1. Table 1. Mean square error with Gaussian for varies methods. Sobel Sobel and mean Sobel+mean+threshold Different noise operator filter (SMT) RMSE 74.523 35.603 28.563 (=0.02) RMSE 80.269 38.334 30.240 (=0.024) RMSE 84.614 41.035 31.790 (=0.028) RMSE 88.978 42.685 33.021 (=0.032) RMSE 92.390 44.889 34.737 (=0.036) RMSE 94.112 46.355 35.961 (=0.04) 3. CONCLUSION Edge detector method using Sobal operator with mean filter failed to remove noise that has different Gaussian noise amount for image edge. In this study concluded that the median filter with threshold technique (threshold value =128) are very well for removal of Gaussian noise from image edges. 4. ACKNOWLEDGMENT Financial support was provided by the College of Science, University of Kufa.
125

Median (SMT) 28.580 29.240 30.053 30.239 31.282 31.995

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 4, Issue 6, November - December (2013), IAEME

REFERENCES
[1] [2] [3] [4] Alshamarti H. A., (2013), Removal of Gaussian noise on the image edges using the Prewitt operator and threshold function technical, Journal of Computer Engineering, 15, 81-85. Azzam S., Wesam A., Mohamad Q., Shatha A., Oraib Al-M,( 2008), Recognizing Objects by Detecting Multiple Moving Parts, The Journal of American Science, 4:4, 38-49. Chen T. C., K. K. Ma, L. H. Chen, (1999), Tri-state Median Filter for Image Denoising, IEEE Transactions on Image Processing, 8:12, 1834-1838. Gajanand Gupta, (2011), Algorithm for Image Processing Using Improved Median Filter and Comparison of Mean, Median and Improved Median Filter. International Journal of Soft Computing and Engineering, ISSN: 2231-2307, 1. Gonzalez & Woods, (2004), Digital Image Processing, Prentice Hall, 3rd edition. Keiji Taniguchi, (2002), DIGITAL IMAGE PROCESSING, (Basical)[M], Beijing : Science Press and Kyoritsu Shuppan Co., Ltd. Ko S. J., Y. H. Lee., (1991), Center Weighted Median Filters and their Applications to Image Enhancement, Transactions on Circuits and Systems, 38: 9, 984-993. Kutty K., S. Ojha. , (2012), A Generic Transfer Function based Technique for Estimating Noise from Images, International Journal of Computer Applications (0975 8887), 51. Maarten Jansen, (2001), Noise Reduction by Wavelet Thresholding, 161. Springer Verlag, United States of America, 1st edition. Padmavathi G., P. Subashini, M. Muthu Kumar, Suresh Kumar Thakur, (2009), Performance analysis of Non Linear Filtering Algorithms for underwater images, (IJCSIS) International Journal of Computer Science and Information Security, 6. Ravi S., A. M. Khan., (2012), Operators Used In Edge Detection Computation: A Case Study, International Journal of Applied Engineering Research, ISSN 0973-4562, 7. Salem Al-amri, N.V. Kalyankar, Khamitkar S.D., (2010), Image Segmentation by Using Thershod Techniques, Journal of Computing, 2, 2151-9617. Senthilkumarn N., R.Rajesh, (2009), Edge Detection Techniques for Image Segmentation- A Survey of Soft Computing Approaches, IJRTE, 1, 250-254. Yong Lee, S. Kassam, (1985), Generalized Median Filtering and Related Nonlinear Filtering Techniques, IEEE Transactions on Acoustics, Speech and Signal Processing, 33:3, 672683. Shruti V Kamath, Mayank Darbari and Dr. Rajashree Shettar, Content Based Indexing and Retrival from Vehicle Surveillance Videos using Gaussian Mixture Model, International Journal of Computer Engineering & Technology (IJCET), Volume 4, Issue 1, 2013, pp. 420 - 429, ISSN Print: 0976 6367, ISSN Online: 0976 6375. Lalit Saxena, Effective Thresholding of Ancient Degraded Manuscript Folio Images, International Journal of Computer Engineering & Technology (IJCET), Volume 4, Issue 5, 2013, pp. 285 - 291, ISSN Print: 0976 6367, ISSN Online: 0976 6375. Shameem Akthar, Dr. D Rajaylakshmi and Dr. Syed Abdul Sattar, A Modified PSO Based Graph Cut Algorithm for the Selection of Optimal Regularizing Parameter in Image Segmentation, International Journal of Advanced Research in Engineering & Technology (IJARET), Volume 4, Issue 3, 2013, pp. 273 - 279, ISSN Print: 0976-6480, ISSN Online: 0976-6499. J.Rajarajan and Dr.G.Kalivarathan, Influence of Local Segmentation in the Context of Digital Image Processing A Feasibility Study, International Journal of Computer Engineering & Technology (IJCET), Volume 3, Issue 3, 2012, pp. 340 - 347, ISSN Print: 0976 6367, ISSN Online: 0976 6375. Mane Sameer S. and Dr. Gawade S.S., Review on Vibration Analysis with Digital Image Processing, International Journal of Advanced Research in Engineering & Technology (IJARET), Volume 4, Issue 3, 2013, pp. 62 - 67, ISSN Print: 0976-6480, ISSN Online: 0976-6499.
126

[5] [6] [7] [8] [9] [10]

[11] [12] [13] [14] [15]

[16]

[17]

[18]

[19]