1 PROCESSING OF DIGITAL IMAGES OBTAINED BY DIFFERENT TECHNIQUES OF MICROSCOPY OF NATURAL RUBBER MEMBRANES USING WAVELET AND FOURIER TRANSFORMS Alexandre Fioravante de Siqueira 1 , Messias Meneguette Junior 2 , Aldo Eloizo Job 3 1 Faculdade de Ciências e Tecnologia - Unesp, Presidente Prudente, Brazil, siqueiraaf@gmail.com 2 Faculdade de Ciências e Tecnologia - Unesp, Presidente Prudente, Brazil, messias@fct.unesp.br 3 Faculdade de Ciências e Tecnologia - Unesp, Presidente Prudente, Brazil, job@fct.unesp.br Abstract: This work is about a study on image processing of natural rubber membranes. A software is being created to use wavelet and Fourier transforms in digital images of natural rubber; with these results, is expected the use of pattern recognition to those images. Keywords: Natural Rubber, Fourier Transform, Wavelets. 1. INTRODUCTION The latex, which is obtained from “Hevea brasiliensis” tree using the bleeding process, is a colloidal dispersion of a polymeric substance in an aqueous environment, mainly composed by the cis-1,4-isoprene monomer [1] which has visco-elastic behavior. It comprises approximately 30-45% by weight of hydrocarbons and non-rubber constituents totaling 3-5%, depending on climatic factors, frequency of bleeding and soil type. There is much speculation about the function of latex in the plant, and a part of this points to the transport and storage of food nutrients. However, the vessels that carry it (laticifers petioles) can not lead the photosynthesized leaves, or they would become blocked. It can hardly be regarded as a reserve of food to be used when the plant is in a position not very favorable, and when exposed to air has protective action against mechanical injury. Natural rubber obtained from latex is a potential source in several areas of research, although in the past three years, few articles were found in the literature that would contribute to studies of incorporation of colloidal metal nanoparticles in natural rubber latex. Recent studies have shown that electrical properties of the latex (observed for a period of forty-eight months) has negative charges that influence the stability. Brazilian researchers have contributed to the development of new materials with elastomeric properties using natural and/or synthetic rubber, seeking various application areas. The most outstanding of these materials have conductive properties. In this work mathematical tools were used to study latex images. These tools are the wavelet, Fourier and Gabor transforms. With these transforms it can be assigned a vector of coefficients to each processed image, and using this vector, these images can be compared by algorithms of pattern recognition. Among the techniques for pattern recognition is the Euclidean distance, which can be applied depending on the kind of vector obtained by analysis of given transforms [2]. Concerning the database, can be used a method called Content Based Image Retrieval (CBIR) [3], which allows the creation of a database with the coefficient vectors of the images. As the program is developed, is expected to also implement a technique called neural networks for pattern recognition; using this technique is expected more reliable results. The software that has been created, named "Image Segmentation", makes it possible for the user to choose among types of wavelets (ie, Haar and Daubechies) and transforms (Fourier and Gabor) proposed in the work, to compare obtained results. Wavelet and pattern recognition techniques have been explored together in recent years [2,3,4]. Because of the relatively low computational cost of some wavelet, and also by the simplicity of this tool, they are preferred instead of more complicated and expensive techniques. Wavelets came to fill gaps left in Fourier analysis, bringing numerous benefits to mathematicians, physicists, engineers, and others, by the versatility of the tool. 2. PURPOSE This study aims at analyzing images obtained by atomic force, scanning and optical microscopies, using different types of wavelets, namely Haar and Daubechies as well as Fourier and Gabor transforms. Also, algorithms related to pattern recognition will be implemented in research and set up classes for selection and discrimination of images. 3. METHODS 3.1. The Fourier Transform The Fourier transform is an integral transform that expresses a function in terms of sinusoidal basis functions, ie as a sum or integral of sinusoidal functions multiplied by coefficients. Being a versatile tool, it is used in various fields of science, like physics, chemistry, number theory, combinatory analysis, signal and image processing, probability theory, and other areas. The extensive use of such transform is given by the possibility of making a difficult problem to solve in a more simple to be solved. In Proceedings of the 9th Brazilian Conference on Dynamics Control and their Applications Serra Negra, SP - ISSN 2178-3667 824