UTN- FRC -Producción Académica de Investigación y Desarrollo - Artículos

Permanent URI for this collectionhttp://48.217.138.120/handle/20.500.12272/2453

Browse

Author: search.filters.author.Nemer Pelliza, Karim Alejandra

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Analysis of the efficiency of the adaptive canny method for the detection of icebergs at open sea
    (IEEE, 2020) Nemer Pelliza, Karim Alejandra; Pucheta, Martín Alejo
    The detection of icebergs in the open sea, as well as its evolution in displacement and shrinking, is vital for navigation, the study of the evolution of Polar regions, and the Earth climate change, among others. In order to carry out these studies, it is necessary to delimit accurately the icebergs in satellite images, mainly of the Synthetic Aperture Radar (SAR) type. The Adaptive Canny method has shown to be efficient for the detection of edges of objects in SAR images, according to recent publications and confer- ences. These studies were only carried out for images that had approximately half of each backscatter, without considering that the dimension of the objects can affect the edge detection process. Here, we present the results of the efficiency of the Adaptive Canny method as the size of the object, from which it is intended to extract the contour, decreases. A systematic analysis of the behavior of the method has been performed with objects of variated dimensions, through a Monte Carlo type experiment with synthetic images, where the contours of the figures were extracted with the Adaptive Canny method and compared with the Ground Truth (GT). Then, the method was tested on real images of the Antarctic Ocean, with blocks of ice of different sizes to contrast the results with those obtained with synthetic images
  • Thumbnail Image
    Item
    Comparación de métodos para la detección de bordes en imágenes satélitales SAR
    (Asociación Argentina de Mecánica Computacional, 2019) Ré, Miguel A.; Nemer Pelliza, Karim Alejandra
    Resumen. El reconocimiento y delimitación de regiones homogéneas en imágenes satelitates de ra- dar de apertura sintética (SAR) es un problema de interés por las diversas aplicaciones posibles. Sin embargo la presencia del ruido de speckle en la retrodispersión dificulta el procesamiento de las imá- genes con métodos tradicionales. Surge así el interés en el análisis y evaluación de distintas alternativas para superar esta dificultad. Presentamos en esta comunicación un análisis comparativo del desempe- ño de dos métodos desarrollados por uno de los autores con una nueva alternativa en desarrollo ba- sada en la divergencia de Jensen Shannon, una medida de semejanza entre distribuciones de probabi- lidad. Las densidades a comparar se estiman por el método del kernel de densidad. Se generan imá- genes sintéticas y se evalúa el desempeño de cada método por la cifra de mérito de Pratt (PFoM).
  • Thumbnail Image
    Item
    Optimal canny’s parameters regressions for coastal line detection in satellite-based SAR images
    (2020) Nemer Pelliza, Karim Alejandra; Pucheta, Martín Alejo; Flesia, Ana Georgina
    Canny’s algorithm is a very well-known and widely implemented multistage edge detector. The extraction of coastal lines in space-borne-based synthetic aperture radar (SAR) images using this algorithm is particularly complicated because of the multiplicative speckle noise present in them and can only be used if Canny’s parameters (CaPP) are chosen appropriately. This letter introduces a methodology for computing functional forms for the CaPP, using functions of the image characteristics through a system that combines artificial neural networks (ANN) with statistical regression. A set of CaPP functional forms is obtained by applying this method on synthetic SAR images. Pratt’s fig- ure of merit (PFoM) is used to measure the performance of them, obtaining more than 0.75, on average, in the 14 400 synthetic SAR images analyzed. Finally, this set of formulas has been tested for extracting coastal edges from real polynyas SAR images, acquired from Sentinel-1.