Diseño de un esquema de muestreo de datos para la carga y almacenamiento progresivo de nubes de puntos
Cargando...
Archivos
Autores
Arboleda Duque, Sebastian
Tipo de contenido
Document language:
Español
Fecha
Título de la revista
ISSN de la revista
Título del volumen
Documentos PDF
Resumen
En este documento se presenta una técnica para simplificar nubes de puntos, la cual se basa en la información espacial contenida en una nube de puntos no estructurada. Para lograr esto, se pretende reorganizar el conjunto de datos de manera tal que al inicio del archivo se ubiquen los puntos más significativos para el modelo 3D, esto permite que el conjunto de datos pueda ser cargado, visualizado o almacenado de forma progresiva. Para clasificar los puntos de acuerdo a su importancia dentro del conjunto de datos, el modelo 3D es dividido en segmentos relativamente planos usando un algoritmo de crecimiento de regiones, luego cada región es analizada para clasificar los datos dentro de ella como puntos pertenecientes a algún borde o puntos pertenecientes a zonas intermedias de la superficie. Finalmente se realiza un muestreo de los datos clasificados y se ordenan en el archivo. De esta forma el usuario puede tener el control de la cantidad de puntos que desea cargar del archivo y así se obtiene un modelo de diferentes resoluciones
Abstract : In this document a technique to simplify point clouds is presented, which is based in the spatial information contained in a unstructured point cloud. To achieve this, the intention is to reorganize the data set in such a way that the most significant dots for the 3D model are located at the beginning of the file, which allows that the data set can be uploaded, viewed or stored in a progressive way. To classify the dots according to their importance within the data set, and in this way re arrange the file, the 3D model is split into segments relativity plane using a region growing algorithm, next each region is analyzed to classify the data inside it as dots belonging to any border or dots belonging to intermediate zones of the surface. Finally, a sampling of the classified data is made and they are arranged in the file. In this way the user can control the amount of dots that he wants to upload from the file and so he gets a model of different resolutions
Abstract : In this document a technique to simplify point clouds is presented, which is based in the spatial information contained in a unstructured point cloud. To achieve this, the intention is to reorganize the data set in such a way that the most significant dots for the 3D model are located at the beginning of the file, which allows that the data set can be uploaded, viewed or stored in a progressive way. To classify the dots according to their importance within the data set, and in this way re arrange the file, the 3D model is split into segments relativity plane using a region growing algorithm, next each region is analyzed to classify the data inside it as dots belonging to any border or dots belonging to intermediate zones of the surface. Finally, a sampling of the classified data is made and they are arranged in the file. In this way the user can control the amount of dots that he wants to upload from the file and so he gets a model of different resolutions
Abstract
Palabras clave propuestas
Simplificación de nubes de puntos; Estructuras de datos espaciales; Sistemas multirresolución; Algoritmo de niveles de detalle; Carga progresiva; Visión por computador; Point cloud simplification; Spatial data structures; Multiresolution systems,; Level of detail algorithms; Progressive loading; Computer vision