Estado de conservación de la Puya raimondii Harms mediante técnicas de teledetección y modelos Deep Learning en el área de conservación regional bosque de Puya Raimondi - Titankayocc, Ayacucho
No hay miniatura disponible
Fecha
2023-07-26
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Pontificia Universidad Católica del Perú
DOI
Resumen
Los estudios de la Puya raimondii Harms en el Perú son escasos, pese a su valor
ecológico y económico para los ecosistemas altoandinos. Actualmente, su
situación es grave debido a las amenazas climáticas y antropogénicas que afectan
en el crecimiento poblacional de la especie. Consecuencia de ello, la P. raimondii
se encuentra declarada en peligro de extinción, ya que presenta poca variabilidad
genética para soportar dichos cambios; además, produce una sola inflorescencia
al final de su periodo vegetativo. De manera que, el objetivo general de esta tesis
es estudiar y evaluar el estado de conservación de la P. raimondii a través de la
teledetección y el uso de nuevas técnicas de detección de objetos como son los
algoritmos de Deep Learning aplicado en un área representativa de puyas como
es el Área de Conservación Regional Bosque de Puya Raimondi - Titankayocc,
departamento de Ayacucho. La metodología implica el uso de herramientas de
Sistemas de Información Geográfica y análisis espacial basado en la
geoestadística para estimar el número de individuos a través de imágenes
satelitales de Google Earth; posteriormente, calcular los valores de las variables
ambientales como el Índice de Vegetación de Diferencia Normalizada (NDVI) y
el Índice de Rugosidad del Terreno (TRI) provenientes de satélites de alta
resolución, CBERS-4A y SRTM respectivamente; finalmente, discretizar la
información hallada para caracterizar el hábitat de la P. raimondii dentro del área
de conservación. En ese sentido, los resultados alcanzados concluyeron en la
detección de 58 607 individuos usando imágenes Google Earth. Asimismo, la
actividad fotosintética registrada tenía como valor promedio un 0.23 según el
NDVI; de igual manera, para el caso del TRI se identificaron los hábitats más
propicios para la especie los cuales fueron suelos rugosos ligeros a elevados
ubicados principalmente en los ejes Este y Sur. Dicho esto, la propuesta de
nuevas estrategias para el estudio de conservación implicó abordar los conceptos
relacionados a la ecología vegetal, análisis espacial e inteligencia artificial.
Studies on Puya raimondii Harms in Peru are scarce, despite its ecological and economic value for high Andean ecosystems. Currently, its situation is serious due to climate and anthropogenic threats that affect the population growth of the species. As a result, P. raimondii has been declared in danger of extinction since it has little genetic variability to withstand such changes; in addition, it produces only one inflorescence at the end of its vegetative period. Therefore, the general objective of this thesis is to study and evaluate the conservation status of P. raimondii through remote sensing and the use of new object detection techniques such as Deep Learning algorithms applied in a representative area of puyas, namely the Regional Conservation Area of Puya Raimondi Forest - Titankayocc, department of Ayacucho. The methodology involves the use of Geographic Information Systems tools and spatial analysis based on geostatistics to estimate the number of individuals through Google Earth satellite images; subsequently, calculate the values of environmental variables such as the Normalized Difference Vegetation Index (NDVI) and the Terrain Roughness Index (TRI) from high-resolution satellites, CBERS-4A and SRTM respectively; finally, discretize the information found to characterize the habitat of P. raimondii within the conservation area. In this sense, the results achieved concluded in the detection of 58,607 individuals using Google Earth images. Likewise, the registered photosynthetic activity had an average value of 0.23 according to the NDVI; similarly, in the case of the TRI, the most favorable habitats for the species were identified, which were light rugged soils to elevated ones located mainly in the eastern and southern axes. That said, the proposal of new strategies for conservation study implied addressing concepts related to plant ecology, spatial analysis and artificial intelligence.
Studies on Puya raimondii Harms in Peru are scarce, despite its ecological and economic value for high Andean ecosystems. Currently, its situation is serious due to climate and anthropogenic threats that affect the population growth of the species. As a result, P. raimondii has been declared in danger of extinction since it has little genetic variability to withstand such changes; in addition, it produces only one inflorescence at the end of its vegetative period. Therefore, the general objective of this thesis is to study and evaluate the conservation status of P. raimondii through remote sensing and the use of new object detection techniques such as Deep Learning algorithms applied in a representative area of puyas, namely the Regional Conservation Area of Puya Raimondi Forest - Titankayocc, department of Ayacucho. The methodology involves the use of Geographic Information Systems tools and spatial analysis based on geostatistics to estimate the number of individuals through Google Earth satellite images; subsequently, calculate the values of environmental variables such as the Normalized Difference Vegetation Index (NDVI) and the Terrain Roughness Index (TRI) from high-resolution satellites, CBERS-4A and SRTM respectively; finally, discretize the information found to characterize the habitat of P. raimondii within the conservation area. In this sense, the results achieved concluded in the detection of 58,607 individuals using Google Earth images. Likewise, the registered photosynthetic activity had an average value of 0.23 according to the NDVI; similarly, in the case of the TRI, the most favorable habitats for the species were identified, which were light rugged soils to elevated ones located mainly in the eastern and southern axes. That said, the proposal of new strategies for conservation study implied addressing concepts related to plant ecology, spatial analysis and artificial intelligence.
Descripción
Palabras clave
Inteligencia artificial, Sistemas de información geográfica, Plantas--Perú--Ayacucho, Flora--Perú--Ayacucho, Teledetección
Citación
Colecciones
item.page.endorsement
item.page.review
item.page.supplemented
item.page.referenced
Licencia Creative Commons
Excepto se indique lo contrario, la licencia de este artículo se describe como info:eu-repo/semantics/openAccess