Uso de imágenes de satélite de diferente resolución espacial en la estimación de producción en parronales
Loading...
Date
2006
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
La aplicación de la teledetección, a través del uso de imágenes multiespectrales registradas tanto en plataformas aéreas como espaciales, ha tenido un gran desarrollo en los últimos años en la agricultura de precisión. En el presente estudio, se pretende considerar las diferencias espaciales existentes en los cultivos, en particular en parronales, en la estimación de la producción de uva de mesa (cv. Red Globe) mediante el uso de una imagen capturada por el sensor multiespectral del satélite QuickBird y otra de mejor resolución espacial, que se obtuvo al realizar un proceso de fusión de la imagen multiespectral, antes mencionada, y su correspondiente imagen capturada por el sensor pancromático, del mismo satélite. Se identificó sectores cuyo rendimiento es conocido (regiones de interés, ROI) para luego ubicarlos en la imagen. Se analizó la respuesta espectral y se calculó el NDVI (Normalized Difference Vegetation Index) de cada ROI para estudiar su relación con el rendimiento. Se aplicó la Transformada de Tasseled Cap como otra forma de detectar diferencias entre las distintas regiones de interés. Luego de realizado este estudio, no se encontraron diferencias significativas en la respuesta espectral, ni en las componentes de Tasseled Cap de las hileras con distinto rendimiento. Por otro lado, si bien es posible apreciar una variabilidad espacial en los mapas de NDVI obtenidos para la zona bajo estudio, no se pudo expresar ésta en función de los rendimientos.
The application of remote sensing, through the use of multispectral images registered as much in aerial as in space platforms, has been a great development in the last years in precision agriculture. In the present study it is tried to consider the existing space differences in crops, particularly in grape, in the estimation of production in table grape (cv. Red Globe) using an image captured by the multispectral sensor of the QuickBird satellite and another one of better space resolution, obtained through a process of fusion of the multispectral image, before mentioned, and its corresponding image captured by the pancromatic sensor, of the same satellite. Once identified sectors whose yield is known (region of interest, ROI) they were located in the image. For each ROI, the spectral response was analyzed and the NDVI (Normalized Difference Vegetation Index) was calculated in order to study its relationship with the crop yield. The Tasseled Cap Transform was applied as another way to detect differences between the different regions of interest. After this study, non significant differences were found in the spectral response, nor in the Tasseled Cap components of rows with different yield. On the other hand, although it is possible to appreciate a spatial variability in the NDVI maps obtained for the area under study, it was not possible it to be expressed based on the crop yield.
The application of remote sensing, through the use of multispectral images registered as much in aerial as in space platforms, has been a great development in the last years in precision agriculture. In the present study it is tried to consider the existing space differences in crops, particularly in grape, in the estimation of production in table grape (cv. Red Globe) using an image captured by the multispectral sensor of the QuickBird satellite and another one of better space resolution, obtained through a process of fusion of the multispectral image, before mentioned, and its corresponding image captured by the pancromatic sensor, of the same satellite. Once identified sectors whose yield is known (region of interest, ROI) they were located in the image. For each ROI, the spectral response was analyzed and the NDVI (Normalized Difference Vegetation Index) was calculated in order to study its relationship with the crop yield. The Tasseled Cap Transform was applied as another way to detect differences between the different regions of interest. After this study, non significant differences were found in the spectral response, nor in the Tasseled Cap components of rows with different yield. On the other hand, although it is possible to appreciate a spatial variability in the NDVI maps obtained for the area under study, it was not possible it to be expressed based on the crop yield.
Description
Tesis presentada para optar al título de Ingeniero Civil Agrícola
Keywords
Parronales, Imagen satelital