Scholarlink

<p>In 2021, the insect <i>Bactericera cockerelli</i> (Hemiptera: Triozidae), vector of the causal agents of Potato Purple Top Disease (PPTD), was reported for the first time at Colombia, specifically in the Nariño state. Currently there is a great interest in maintaining the phytosanitary problem restricted to this zone or geographic area of the country, due to the yield and economic losses potential in other important producer regions of potato but also on other crops Solanaceae such as tree tomato, chili, pepper and cape gooseberry. This study aimed to design a methodology to detect these diseases in the region to mitigate their spread into the country. The evaluation was conducted in an experimental plot and seven commercial potato farms with presence of the cultivars Diacol Capiro and Superior, using an Unmanned Aerial Vehicle (UAV), equipped with a multispectral camera to capture images of health plants and with mild, intermediate and advanced symptoms of PPTD to estimate vegetation indices. The multispectral information, derived vegetation indices, and field symptom classes were analyzed using supervised machine-learning modeling to evaluate their capacity to discriminate disease-severity levels. Our results conclude that, for the two potato varieties evaluated, there is a decrease in reflectance between 550 and 600&#xa0;nm in affected plants, suggesting an increase in light absorption in this part of the spectrum. Additionally, the Random Forest (RF) model excelled in detecting mild class using indices such as Normalized Difference Vegetation Index (NDVI) and Transformed Vegetation Index (TVI), whereas the Support Vector Machine (SVM) model showed higher sensitivity to detect disease advanced stages, where (TVI) and Green Chlorophyll Index (GCI) were identified as the most influential features in classification.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Precision Agriculture Approach for Potato Purple Top Disease Detection: Utilizing UAV - Derived Multispectral Imagery and Machine Learning for Symptom Characterization

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月25日

查看全文
  • Diego Fernando Sánchez-Vivas,
  • David Rodriguez Puertas,
  • Pedro Pablo Bacca Acosta,
  • Donald Heberth Riascos-Ortiz

摘要

In 2021, the insect Bactericera cockerelli (Hemiptera: Triozidae), vector of the causal agents of Potato Purple Top Disease (PPTD), was reported for the first time at Colombia, specifically in the Nariño state. Currently there is a great interest in maintaining the phytosanitary problem restricted to this zone or geographic area of the country, due to the yield and economic losses potential in other important producer regions of potato but also on other crops Solanaceae such as tree tomato, chili, pepper and cape gooseberry. This study aimed to design a methodology to detect these diseases in the region to mitigate their spread into the country. The evaluation was conducted in an experimental plot and seven commercial potato farms with presence of the cultivars Diacol Capiro and Superior, using an Unmanned Aerial Vehicle (UAV), equipped with a multispectral camera to capture images of health plants and with mild, intermediate and advanced symptoms of PPTD to estimate vegetation indices. The multispectral information, derived vegetation indices, and field symptom classes were analyzed using supervised machine-learning modeling to evaluate their capacity to discriminate disease-severity levels. Our results conclude that, for the two potato varieties evaluated, there is a decrease in reflectance between 550 and 600 nm in affected plants, suggesting an increase in light absorption in this part of the spectrum. Additionally, the Random Forest (RF) model excelled in detecting mild class using indices such as Normalized Difference Vegetation Index (NDVI) and Transformed Vegetation Index (TVI), whereas the Support Vector Machine (SVM) model showed higher sensitivity to detect disease advanced stages, where (TVI) and Green Chlorophyll Index (GCI) were identified as the most influential features in classification.