Please use this identifier to cite or link to this item:
Title: Procesamiento de imágenes multiespectrales para el análisis del estado de la vegetación
Author: Canalejo Ariza, Sergio
Director: Ventura Royo, Carles  
Tutor: Kanaan Izquierdo, Samir
Others: Universitat Oberta de Catalunya
Keywords: convolutional neural network
vegetation indices
multispectral images
Issue Date: Jun-2018
Publisher: Universitat Oberta de Catalunya
Abstract: The present work tries to develop a multispectral image analysis system with the values of the different spectral bands, so that it is capable of processing said images and quickly obtain the value of a specific vegetation index that gives us an information about the state of the vegetation at the time of the images. To do this, a convolutional neuronal network has been designed, acting as a regressor, capable of processing the data and offering the value of the selected vegetation index in a short period of time. This type of systems has been chosen for its great capacity and accuracy in the image treatments. This system aims to expedite the obtaining of these data by the Department of prevention of incendis since it currently depends on an external system that gives the data in ranges of two weeks. After the design and training of the convolutional neuronal network, it has been possible to obtain results very close to the actual results for a specific pixel of the image immediately, assuming a great improvement with the way of obtaining that data at present.
Language: Spanish
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
scanalejocTFM0618memoria.pdfMemoria del TFM1.06 MBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons