Cargando...
Fecha
2024-10-22
Derechos de acceso
info:eu-repo/semantics/openAccess
Título de la revista
ISSN de la revista
Título del volumen
Editorial
Universidad Nacional de Educación a Distancia (España), Universidad de Concepción - Chile. Departamento de Ingeniería Mecánica
Citas
Resumen
El maíz forrajero es parte fundamental en la alimentación del ganado, por lo que se requiere cuantificar la cantidad de materia verde a cosechar para establecer estrategias de ensilado, debido a lo anterior se propone una metodología para estimar biomasa verde de maíz mediante imágenes multiespectrales de alta resolución y una red neuronal artificial. Para ello, se tomaron 80 muestras georreferenciadas de alturas y pesos de las plantas, se obtuvieron los índices de vegetación NDVI, EVI, GNDVI, WDRVI, CIre, RVI, SAVI, VARI, RGBVI, NGRDVI, ExG y el índice de área foliar. Posteriormente, se realizó la matriz de correlación y se aplicó el algoritmo de componentes principales creando una base de datos con seis componentes principales, generándose el modelo predictivo basado en un perceptrón multicapa, que mostró un buen ajuste entre la biomasa real y la estimada, con un 𝑅2= 92.9% y un error absoluto medio de 0.3 kg m-1.
Forage corn is a fundamental part of livestock feed, so it is necessary to quantify the amount of green matter to be harvested to establish silage strategies. Therefore, a methodology is proposed to estimate green biomass of corn using high resolution multispectral images and an artificial neural network. For this purpose, 80 georeferenced samples of plant heights and weights were taken and the vegetation indices NDVI, EVI, GNDVI, WDRVI, CIre, RVI, SAVI, VARI, RGBVI, NGRDVI, ExG and the leaf area index were obtained. Subsequently, the correlation matrix was performed and the principal component algorithm was applied creating a database with six principal components, generating the predictive model based on a multilayer perceptron, which showed a good fit between the real and estimated biomass, with a 𝑅2= 92.9% and a mean absolute error of 0.3 kg m-1.
Forage corn is a fundamental part of livestock feed, so it is necessary to quantify the amount of green matter to be harvested to establish silage strategies. Therefore, a methodology is proposed to estimate green biomass of corn using high resolution multispectral images and an artificial neural network. For this purpose, 80 georeferenced samples of plant heights and weights were taken and the vegetation indices NDVI, EVI, GNDVI, WDRVI, CIre, RVI, SAVI, VARI, RGBVI, NGRDVI, ExG and the leaf area index were obtained. Subsequently, the correlation matrix was performed and the principal component algorithm was applied creating a database with six principal components, generating the predictive model based on a multilayer perceptron, which showed a good fit between the real and estimated biomass, with a 𝑅2= 92.9% and a mean absolute error of 0.3 kg m-1.
Descripción
Organizado y patrocinado por: Federación iberoamericana de Ingeniería Mecánica y Universidad de Concepción - Chile. Departamento de Mecánica, FeIbIm – FeIbEM
Categorías UNESCO
Palabras clave
dron, inteligencia artificial, rendimiento, agricultura, uav, artificial intelligence, yield, agriculture
Citación
-
Centro
E.T.S. de Ingenieros Industriales
Departamento
Mecánica