Alejandro Restrepo-Martínez

Evolución de la Intensidad de Pixeles para Describir la Deformación de Películas Plásticas

Resumen es: Este trabajo propone una metodologia para la descripcion de compor-tamientos mecanicos en la deformacion de una pelicula plastica termode-formable utiliz...

Computational analysis of stress map variations by industrial light sources and load additions in digital photoelasticity

In digital photoelasticity, evaluating the stress map is often affected in regions with critical values. This phenomenon is associated to color degradation effect and high fringe densities. It is a consequence of different experimental conditions, such as: type of birefringent material, relative spectral content of light source, relative spectral response of camera sensor, polarization optical elements, load application, etc. In this study field, the main goal accounts for evaluating the stress values, as better as possible, from photoelasticity images. Which turns the view towards the process that allow to acquire photoelasticity images with more complete information. This makes necessary to analyze the possible effects that each element could introduce into the photoelasticity image generation. This paper presents a computational analysis on the effect that different industrial light sources introduces for recovering the stress maps. Hence, four common industrial light sources are considered for generating the photoelasticity images. In this case, results reveal that there are light sources which represent stronger limitations for evaluating the stress, and that Such effect varies with the load increments. This approach is useful for predicting the possible effect that a light source selection could introduce into the stress evaluation process.

Modified correlation method to autofocus particles in digital in-line holography

Abstract. A method is proposed to automatically evaluate the focal planes of spherical particles. This method compares the correlation coefficients of multiple reconstructed planes relative to a reference plane. The particles are located where a minimum correlation is found, and reconstructions are made using an angular spectrum propagator. The Hough transform is employed to segment the hologram, thereby enabling the detection of circular shapes, such as Airy patterns, and the edges of the particles themselves. The autofocus is improved by creating a correlation matrix using an iterative process, which reduces the computational cost of the particle display processes in their respective focal planes. A theoretical model was studied to estimate the longitudinal and transverse magnifications of the focused particles caused by the influence of aberrations in the reconstruction of digital holograms due to the spherical reference wave used. Experimentally, laser light was used to illuminate 5-μm latex particles, which was recorded by a CCD camera with a 9.9-μm pixel size. The reconstructions measured an average particle radius of 71.3±16.3  μm in the average focal plane, which was estimated to be 60.65±0.22  mm from the hologram where the magnifications were considered.

Morphological Analysis for Temporary Behaviour Description of Photoelasticity Images on Plastic Films

Thermo-deformable plastic films experience an optical birefringence phenomenon when they are subjected to mechanical strain, this phenomenon can be described by color fringes behavior observed in sequences of images acquired through photoelasticity assemblies. This paper proposes a temporary morphological analysis that color fringes experience when the plastic film is deformed. For the image analysis, stereology and erosion surfaces techniques are applied. It achieves the relationship between morphological description and mechanical behavior of the plastic films deformation process. The stereology and erosion surfaces allowed the identification of the appearance of color fringes that gradually increase in size into the image sequence. The morphological results showed that temporary behavior of color fringes could be used for modeling the mechanical behavior of the plastic films deformation process.

Compensador de Iluminación Utilizando un Diodo LED como Sensor y Actuador de Luz

This work presents an electronic system capable of controlling the direct and inverse polarization applied to the terminals of a LED diode; this allows configuring it as an actuator and light sensor interchangeably. The final goal is to design a system that compensates levels, set by the user, of luminesce potency. The design is described by an Analog block and a low-tech processor Arduino UNO. The Arduino UNO allowed for the functioning of the LED to operate dually by the implemented algorithm. An equation for the developed set was defined experimentally which determines the luminesce potency that the compensating LED should provide; the duty cycle for the PWM applied to the LED terminals was calculated too. Through the developed set it was possible to measure the photocurrent given by the LED configured as light sensor and then turn it into a voltage through the analog block. FInally an algorithm was obtained to compensate luminesce potency.

Clasificación de Mariposas por Modelos de Color HSI y RGB Usando Redes Neuronales

This study aims the classification of Butterfly species through the implementation of Neural Networks and Image Processing. A total of 9 species of Morpho genre which has blue as a characteristic color are processed. For Butterfly segmentation we used image processing tools such as: Binarization, edge processing and mathematical morphology. For data processing RGB values are obtained for every image which are converted to HSI color model to identify blue pixels and obtain the data to the proposed Neural Networks: Back-Propagation and Perceptron. For analysis and verification of results confusion matrix are built and analyzed with the results of neural networks with the lowest error levels. We obtain error levels close to 1% in classification of some Butterfly species.

Caracterización de Bordes y Contrastes en una Imagen Digital ante la Variación de Parámetros de los Filtros Pasa-Altos usados en la Estimación de Visibilidad Atmosférica

espanolEn este documento se consideran los bordes y los contrastes obtenidos con los filtros pasa-altos que se utilizan en la estimacion de la visibilidad atmosferica diurna desde imagenes digitales, cuyo comportamiento es caracterizado con la variacion de los parametros correspondientes a filtros pasa-altos tales como: el Ideal, Gaussiano, y Homomorfico-Gaussiano. Una imagen sintetica que contiene regiones con distintos contrastes se utiliza para aplicar los diferentes filtros; luego, se define un indice para medir la calidad de los bordes obtenidos en las imagenes filtradas y se emplea para analizar los resultados. Los resultados muestran que tanto la seleccion del filtro como la seleccion de sus parametros, afectan las carac-teristicas y la calidad de los bordes detectados en la imagen filtrada, tambien determinan la cantidad de ruido que el filtro agrega a la imagen (artefactos que no estaban presentes en la imagen original) y ademas establecen si se logra o no, la deteccion de los bordes; los resultados obtenidos tambien evidencian que el indice de calidad de los bordes alcanza valores maximos con ciertas combinaciones de los parametros de los filtros. Esto significa que algunas combinaciones de parametros disminuyen las situaciones que distorsionan los bordes y que distorsionan las medidas de visibilidad atmosferica basadas en la transformada de Fourier. Asi que, estos parametros, que proporcionan maximos de calidad para los bordes, se esta-blecen como convenientes para ser utilizados en las medidas de visibilidad. EnglishThis paper considers the edges and contrasts obtained with high-pass filters used in the estimation of daytime atmospheric visibility from digital images, and the behavior of these edges and contrasts is characterized by varying the parameters of high-pass filters such as the Ideal, Gaussian, and Homomorphic-Gaussian. A synthetic image of regions with different contrasts is used to apply different filters, then, we define an index to measure the quality of the edges obtained in the filtered image and it is used to analyze the results. The results show that both, the filter selection and the selection of its parameters: affects the characteristics and quality of the detected edges in the filtered image, also determine the amount of noise that the filter added to the image (artifacts that were not present in the original image), and also establish if achieved, or not, the edge detection. The results also show that the edge quality index reaches maximum values at certain combinations of the filters parameters, which means that some combinations of parameters reduce situations distorting the edges and distorting atmospheric visibility measures based on the Fourier transform. So these parameters which provide maximum quality edges are established as suitable for use in visibility measurement.This paper considers the edges and contrasts obtained with high-pass filters used in the estimation of daytime atmospheric visibility from digital images, and the behavior of these edges and contrasts is characterized by varying the parameters of high-pass filters such as the Ideal, Gaussian, and Homomorphic-Gaussian. A synthetic image of regions with different contrasts is used to apply different filters, then, we define an index to measure the quality of the edges obtained in the filtered image and it is used to analyze the results. The results show that both, the filter selection and the selection of its parameters: affects the characteristics and quality of the detected edges in the filtered image, also determine the amount of noise that the filter added to the image (artifacts that were not present in the original image), and also establish if achieved, or not, the edge detection. The results also show that the edge quality index reaches maximum values at certain combinations of the filters parameters, which means that some combinations of parameters reduce situations distorting the edges and distorting atmospheric visibility measures based on the Fourier transform. So these parameters which provide maximum quality edges are established as suitable for use in visibility measurement.

Incidencia de la Segmentación en la Obtención de Región de Interés en Imágenes de Palma de la Mano

Una de las mayores dificultades en el reconocimiento de patrones y en particular en sistemas biometricos basados en la palma de la mano, es obtener en el preprocesamiento un adecuado calculo de la region de interes (Region Of Interest, ROI), debido a que esta influye directamente en los resultados finales del sistema de identificacion. Este articulo presenta una comparacion entre imagenes que contienen la region de interes de la palma de la mano, obtenidas mediante el uso de tecnicas de umbralizacion, Otsu y Huang para la base de datos PolyU y Otsu e Isodata para la base de datos CASIA. Tambien se realizaron pruebas con dos metodos para la ubicacion del cuadrado correspondiente a la ROI, utilizando como medidas de similitud: la correlacion, la informacion mutua, el error cuadratico medio, la relacion de uniformidad de la imagen y el area comun entre las imagenes, Todo esto, para determinar que tan similares o diferentes son las imagenes generadas por diferentes metodos de ubicacion de ROI y por distintos metodos de umbralizacion. Esta investigacion demostro que efectivamente el cambio de los metodos de umbralizacion generan cambios en la ROI de la palma de la mano ya que la localizacion de los puntos valle cambia segun el metodo utilizado. Se concluye finalmente que al variar el metodo de umbralizacion, las imagenes generadas por cada metodo son diferentes y al variar el metodo de obtencion de la ROI la variacion es mayor.

Optimización de la Segmentación Local de Sauvola Aplicada a la Detección de Defectos Superficiales en Escenas con Iluminación No Homogénea

The presence of non-homogeneous illumination in real scenes images is an actual problem that difficult the correct segmentation of these. This paper presents a methodology for optimizing Sauvola local segmentation for the detection of superficial defects in non-homogeneous illuminated images by adjusting its parameters through genetic algorithms. The methodology consists of these stages: First, the problem is proposed from the perspective of genetic algorithms where each individual in the population represents the values for Sauvolas parameters. Then several fitness functions are proposed using comparison metrics between a Sauvolas segmentation and one performed manually. Each function is evaluated by running the genetic algorithm with it in a subset of images. The best fitness function, according to the results of optimization, is used again in a larger sample. Finally, the last optimization results are analyzed by a clustering analysis. The results show that it is possible to adjust Sauvolas parameters to successfully segment each image but these do not exhibit a tendency to a specific point that allow to suggest unique parameters to segment all images with a high performance.