Procesamiento de imagenes

Computer Vision using MatLAB and the Toolbox of Image Processing Technical Report B-05-09
Erik Cuevas1,2, Daniel Zaldivar1,2, and Raul Rojas1 Freie Universität Berlin, Institut für Informatik Takusstr. 9, D-14195 Berlin, Germany 2 Universidad de Guadalajara, CUCEI Av. Revolucion No. 1500, C.P 44430, Guadalajara, Jal, Mexico. {cuevas, zaldivar, rojas}@inf.fu-berlin.de November 12, 2003.
1Abstract
During the implementation of computer vision algorithms the manipulation of pointers, memory administration and some other resources are expensive in time even for friendly programming language. All these problems can be resolved if the implementation test is carried out in MatLAB using its toolbox of image processing with it the time of implementation becomes the minimum with the trust ofusing algorithms scientifically proven and robust. In this work we show the form in which can be used matlab and its toolboxes to solve common problems of computer vision efficiently.

Introduction During the implementation of computer vision algorithms the manipulation of pointers, memory administration and some other resources are expensive in time even for friendly programming language. Toimplement it in C++ language (due to their characteristics of high and low level it is appropriate for the implementation of computer vision algorithms) it would suppose the lost of valuable time that which is very important in those cases in which we do not have the security of what we want to implement finally will work. Also, to use C++ for the test period demands a correction time due normalerrors in the implementation process of the algorithm, that is to say programmatic errors made to the moment to multiply two matrix or any other resemblance. All these problems can be resolved if the implementation test is carried out in MatLAB using its toolbox of image processing with it the time of implementation becomes the minimum with the trust of using algorithms scientifically proven androbust. The toolbox of image processing implements a group of well-known algorithms to work with binary images, geometric transformation, morphology and color manipulation that together with the functions already integrated with matlab allows to implement analysis and transformations of images in the domain of the frequency (Fourier and Wavlets transform). This work is divided in 3 parts, the firstone treats the basic concepts of the images how they are represented in matlab as well as an introduction to the basic operations of handling of files. The second part covers the common and representative image processing functions in the area of computer vision; explaining the use of these

functions through examples. Finally the third part explains the use of the tool vfm used to capture imagesof devices installed in the computer such as video cards and USB Webcams. Basic concepts of the images and manipulation functions In matlab an grayscale image is represented by a two-dimensional matrix of m x n elements where n represents the number of pixels of wide and m the number of pixels of long. The element v11 corresponds to the top left element. Where each element of the image matrix canhave the value from 0 (black) to 255 (white). The figure 1 shows in detail these concepts.
⎡ v11 v12 ⎢v v22 I ( x, y ) = ⎢ 21 ⎢ ⎢ ⎣ vm1 vm 2 v1n ⎤ v2 n ⎥ ⎥ ⎥ ⎥ vmn ⎦

m

n
Figure 1. Grayscale image representation in MatLAB.

On the other hand a RGB color image (several types exist however it is the broadly used for computer vision, besides representing for matlab the default option) it isrepresented by a three-dimensional matrix m x n x p, where m and n have the same significance for the case of the grayscale images while p represents the plane that can be 1 for the red, 2 for the green and 3 for the blue. The figure 2 shows details of these concepts. Reading and writing images from a file To read images in matlab contained in a file the function imread is used whose syntax is...