BASIC OPERATIONS IN IMAGE PROCESSING USING MATLAB

Transcription

1 BASIC OPERATIONS IN IMAGE PROCESSING USING MATLAB Er.Amritpal Kaur 1,Nirajpal Kaur 2 1,2 Assistant Professor,Guru Nanak Dev University, Regional Campus, Gurdaspur Abstract: - This paper aims at basic image processing operations using Matlab. Matlab is versatile software used as proposal in laboratory training and the problems classes in image dispensation. Fundamental operations like image cropping, brightness, rotation, edge detection, blurring and deblurring of images are discussed in this paper. Keywords: - Sobel, Canny, Prewitt, Roberts, Log filters, RGB I. INTRODUCTION Today computer vision has an important role in our life and has many different application areas [1,2]. To use computers in image processing has different objectives. One of the major objectives is to create more suitable images for people to recognize. The areas of image processing is quite heterogeneous, to gain basic knowledge of these areas, it is essential to classify images according to their sources like electromagnetic energy spectrum, acoustic, ultrasonic and electronic and synthetic images generated by computer[3]. An image is rectangular array of pixels. Each pixel represents the dimension of some property of a picture deliberate over a fixed area. These properties should be many things like brightness, contrast of image, alternation of a image, recognition of edges of image and many more. All these properties depend on certain application and requirement. II. BASIC OPERATIONS ON IMAGE First step is to load image into Matlab. Matlab supports certain types of images like GIF,JPEG,TIF,BMP.It uses command imread to load image. After that various operations are applied on loaded image which are discussed. Image processing using Matlab has wide range of applications such as space exploration, image transmission and storage for business applications, medical processing, radar, sonar and acoustic image processing, robotics and automated industrial inspection [4]. a) CHANGING INTENSITY OF IMAGE Images are usually captured with pixels in each channel being represented by eight bit integers. An intensity image values represents its brightness. A binary image pixel has two values. A RGB image has three channels representing intensities in terms of wavelengths corresponding to red, green and blue illumination. In digital image processing perception, the intensity of a image is referred to global measure of that image such as mean pixel intensity. A relative measure of image intensity could be how bright the image appears compared to other image. Rgb2gray converts RGB image to gray image. Im2bw converts the gray image to binary image. The command bw=im2bw(i,.4) converts gray image to binary with intensity 0.4.Similarly im2bw(i,.6), im2bw(i,.2), im2bw(i,.9) converts gray image to binary with intensities 0.6,0.2,0.9 respectively. Graythresh converts gray image to binary image using threshold method. Figure 1 shows image with different intensities All Rights Reserved 21

2 Original Image Gray Image with 0.4 intensity with 0.6 intensity with 0.2 intensity with 0.9 intensity binary threshold Figure 1:- Image with different intensity levels b) INVERSION & CROPPING OF IMAGE Cropping refers to deduction of outer parts of image in order to cross out an unwanted subject or irrelevant details from that image. There are certain reasons for cropping. For instance to improve photo composition, to give your photo a focus, zoom in for a subject, to change the aspect ratio for printing[5]. Inversion is the process which takes binary or greyscale image as an input and produces its photographic negative [6]. The resulting value for each pixel is the input value subtracted from 255. In figure 2 inversion algorithm is applied to RGB components of a image. The R,G,B components of image are represented as unit8 values, ranging from 0 to 255. The command imcrop creates an interactive Crop Image tool associated with the image displayed in the current figure, called the target image. The Crop Image tool is a moveable, resizable rectangle that you can position interactively using the mouse. When the Crop Image tool is active, the pointer changes to cross hairs when you move it over the target image. Using the mouse, you specify the crop rectangle by clicking and dragging the mouse. You can move or resize the crop rectangle using the mouse. The command imcrop (I,[ ]) crops the original image I having xmin=35, ymin=75, width=112, height=132] that specifies the size and position of the crop image. The command imcomplement will complement the image i.e converts image in its negative All Rights Reserved 22

3 original image Crop image Invert image Image with changed intensity Figure 2:- Crop and inverted Image c) ADJUST BRIGHTNESS & CONTRAST OF IMAGE Brightness is an aspect of visual observation in which a source seems to be radiating or reflecting a specific amount of light [7]. Brightness levels should be proper in order to see the objects in an image. Contrast is ocular property that separates it from other objects. Brightness and contrast terms are often confusing. A very simple difference between the two is that the later is overall lightness or darkness of an image whereas the former is difference in brightness of objects or regions in an image. Figure 3 illustrates an image with different brightness and contrast levels. The command B=I+125 increases brightness of image by factor of 125. D=I-125 decreases brightness of image by factor of 125. C=I*5 decreases contrast of image by factor of 5.E=I/5 increases contrast of image by factor of All Rights Reserved 23

4 Original image Increase brightness Decrease brightness Decrease contrast Increase contrast Figure 3:- An image with different brightness and contrast levels. d) EDGE DETECTION Edge detection is one of the image segmentation techniques. Edge is the most crucial feature of an image. Edges contain lot of information and meaningful features of an image. The edges in an image refer to rapid changes in physical operations such as geometry, reflectivity and illumination [8]. An edge pixel is defined by two important parameters, primarily the edge strength, which is equal to the magnitude of gradient and secondly the edge direction which is equal to the direction of the gradient [9].Five types of edge detection techniques are used in this paper- Sobel, Canny, Prewitt,Roberts and Log edge detectors. Sobel operator performs 2D spatial gradient measurement and has smoothing effect to random noise. Canny detector is used for step edges corrupted by white noise [10].Prewitt is frequently used for detection of horizontal and vertical edges. Roberts operator works well under noisy conditions whereas Log consists of differentiation of Laplace of Gaussian which consists of adjustable features. EDGE(J, 'CANNY') does edge detection of image J by Canny method. M= EDGE(J, 'PREWITT') does edge detection of image J by Prewitt method. N= EDGE(J, 'ROBERTS') does edge detection of image J by Roberts method. Y= EDGE(J, 'LOG') does edge detection of image J by Log method which is shown in figure All Rights Reserved 24

5 GRAY IMAGE SOBEL FILTER CANNY FILTER PREWITT FILTER ROBERTS FILTER LOG FILTER Figure 4:- An image with different edge detection methods. e) ROTATION OF IMAGE Rotation is very fundamental image processing operation. Many applications such as radiology and photographic analysis require very high quality image rotation [6].IMROTATE command rotates the image. Imrotate(I,45 degree), Imrotate(I,90 degree), Imrotate(I,180degree) rotates image by 45,90,180 degrees respectively. Original image 45 degree rotation 90 degree rotation 180 degree rotation Figure 5:- An image with rotation. f) BLURRING & DE BLURRING OF IMAGE Blurring of image is used to reduce image noise and reduce detail. Blur is unsharp image area caused by camera or subject movement, inaccurate focussing or by the use of an aperture that gives shallow depth of field [11]. Deblurring or restoration is the process of recovering original image from blurred image. There are number of algorithms for this. A point-spread function PSF is used here for blurring and deblurring (restoration) of image.psf corresponding to the linear motion across 31 All Rights Reserved 25

6 (LEN=31), at an angle of 11 degrees (THETA=11). To simulate the blur, convolve the filter with the image using imfilter. The simplest syntax for deconvwnr is deconvwnr(a, PSF, NSR), where A is the blurred image, PSF is the point-spread function, and NSR is the noise-power-to-signal-power ratio. Original Image Blurred Image Restored Image Figure 6:- A blurred and restored image. III. CONCLUSION AND FUTURE SCOPE In this paper, various image processing techniques like image cropping, rotation, change of intensity and brightness, blurring and restoration of images are discussed which has numerous applications in a variety of fields. There are many more complex modifications which can be made to images for better results. REFERENCES 1. Er Kiranpreet Kaur, Er Vikram mutenja, Er Indrajeet Singh Gill, Fuzzy Logic based Image Edge detection Algorithm in MATLAB,International Journal of Computer Applications ( ), Vol S.Nagendram, G. Divya, L.Avinash Bhardwaj, P. Dharanijam, A Novel method of exploring Human Reasoning Power in Image Analysis,International Journal of Science and Advanced Techonolgy,Vol.1,Issue 8, Oct RC Gonzalez, RE Woods, Digital Image Processing, 3 rd Edition, Pearson Prentice Hall, A.K.Jain, Fundamentals of Digital Image Processing,, Pearson Prentice Hall, Image Processing: Brightness, Contrast, Gamma, and Exponential/Logarithmic Settings in Pro Analyst 8. Z.J Hou,G.W.Wei, A new approach to edge detection, Pattern Recognition, Vol.35 (2002),Page G.T. Shrivakshan, Dr.C.Chandrashekhar, A comparison of various edge detection techniques used in image processing, IJCSI International Journal of Computer Science Issues, Vol.9, Issue 5 No 1, September 2012, page J. Canny, A computational approach to edge detection, IEEE Trans. Patt. Anal. Machine Intell, Vol. PAMI- 8.pp , Dr.Salem Saleh Al-amri, Dr.Ali Salem Ali, Restoration and Deblured Motion Blurred Images, IJCSI International Journal of Computer Science Issues, Vol. 11, Issue 1, No 1, January 2014 ISSN (Print): ISSN (Online): All Rights Reserved 26