Análisis espectral aplicado a la identificación de estrés vegetal por araña roja en cultivos de aguacate Hass

Fecha
2026
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad de Manizales
Referencia bibliográfica
Resumen
El cultivo de aguacate se ha posicionado como un sector estratégico dentro de la economía del país en los últimos años, representando un incremento significativo en los ingresos per cápita y una reactivación económica en diversas regiones donde esta actividad no era considerada una alternativa de desarrollo. Sin embargo, la expansión de las áreas cultivadas ha propiciado la aparición de plagas que afectan tanto la calidad como la cantidad de la producción. Entre estas, se destaca la araña roja (Tetranychus urticae), una plaga de difícil detección y rápida reproducción que genera pérdidas significativas para los productores. Al ser un organismo fitófago, su acción consiste en la succión de la savia de las plantas, debilitando progresivamente los árboles afectados. Los métodos tradicionales de control suelen ser manuales, aleatorios o programados en intervalos regulares, lo cual no garantiza una detección temprana ni una intervención efectiva.
La presente investigación se desarrolló entre octubre de 2024 y agosto de 2025, en el municipio de Herveo, departamento del Tolima, en la granja Mesones, propiedad de Agrícola Nueva Huatulame; para ella se integró de manera secuencial el uso de imágenes multiespectrales adquiridas desde aeronaves no tripuladas (UAV), procesamiento fotogramétrico y geoespacial. Esta información permitió la generación de índices de vegetación y la extracción de variables espectrales relacionadas con el vigor vegetal. Posteriormente estas variables se analizaron mediante modelos de regresión lineal y algoritmos de aprendizaje automático, permitiendo establecer relaciones cuantitativas entre la respuesta espectral del cultivo y los diferentes niveles de infestación por araña roja. Los resultados permitieron identificar patrones de comportamiento vegetal asociados al estrés causado por la plaga obteniendo modelos con alta capacidad predictiva y una precisión superior en la clasificación de árboles afectados. Esta metodología constituye una herramienta eficiente que permite el monitoreo fitosanitario, optimizando la toma de decisiones oportunas para su control y mitigación.
Avocado cultivation has become a strategic sector within the country’s economy in recent years, representing a significant increase in per capita income and an economic reactivation in various regions where this activity was not previously considered a development alternative. However, the expansion of cultivated areas has led to the emergence of pests that affect both the quality and quantity of production. Among these, the two-spotted spider mite (Tetranychus urticae) stands out—a pest that is difficult to detect and rapidly reproduces, causing significant losses for growers. As a phytophagous organism, its action consists of sucking plant sap, progressively weakening the affected trees. Traditional control methods are often manual, random, or scheduled at regular intervals, which does not guarantee early detection or effective intervention. This research was conducted between October 2024 and August 2025 at the Mesones farm, owned by Agrícola Nueva Huatulame, located in the municipality of Herveo, department of Tolima. The study sequentially integrated the use of multispectral images acquired from unmanned aerial vehicles (UAV), followed by photogrammetric and geospatial processing. This information enabled the generation of vegetation indices and the extraction of spectral variables related to plant vigor. Subsequently, these variables were analyzed using linear regression models and machine learning algorithms, allowing for the establishment of quantitative relationships between the spectral response of the crop and different levels of red spider mite infestation. The results allowed for the identification of plant behavior patterns associated with pest-induced stress, yielding models with high predictive capability and superior accuracy in classifying affected trees. This methodology constitutes an efficient tool for phytosanitary monitoring, optimizing timely decision-making for control and mitigation.
Avocado cultivation has become a strategic sector within the country’s economy in recent years, representing a significant increase in per capita income and an economic reactivation in various regions where this activity was not previously considered a development alternative. However, the expansion of cultivated areas has led to the emergence of pests that affect both the quality and quantity of production. Among these, the two-spotted spider mite (Tetranychus urticae) stands out—a pest that is difficult to detect and rapidly reproduces, causing significant losses for growers. As a phytophagous organism, its action consists of sucking plant sap, progressively weakening the affected trees. Traditional control methods are often manual, random, or scheduled at regular intervals, which does not guarantee early detection or effective intervention. This research was conducted between October 2024 and August 2025 at the Mesones farm, owned by Agrícola Nueva Huatulame, located in the municipality of Herveo, department of Tolima. The study sequentially integrated the use of multispectral images acquired from unmanned aerial vehicles (UAV), followed by photogrammetric and geospatial processing. This information enabled the generation of vegetation indices and the extraction of spectral variables related to plant vigor. Subsequently, these variables were analyzed using linear regression models and machine learning algorithms, allowing for the establishment of quantitative relationships between the spectral response of the crop and different levels of red spider mite infestation. The results allowed for the identification of plant behavior patterns associated with pest-induced stress, yielding models with high predictive capability and superior accuracy in classifying affected trees. This methodology constitutes an efficient tool for phytosanitary monitoring, optimizing timely decision-making for control and mitigation.
Descripción
Palabras clave
Herramientas SIG, Análisis geoespacial, Imágenes multiespectrales, Cultivo de Aguacate Hass, Cartografía digital, Estrés vegetal