Scholarlink

<p>Tequila and the agave plant are important symbols of Mexico’s national identity. In the last decade, the care of the Tequilana Weber Blue Agave has improved and the production of tequila has increased. In this paper, we introduce a methodology for the detection and counting of agave plants based on the use of Unmanned Aerial Vehicles (UAV) known as drones, image processing techniques, and machine learning. Several landscape images of agave crops are collected from the Romita region in Guanajuato state. These images are processed to obtain orthomosaics that allow us to study the target region of agave crops. Our proposal is carried out in two stages. In the first stage, we identify agave cultivation rows from other sections of the field to simplify the detection of agave plants using a superpixel approach and the Random Forest Classifier Algorithm (RFCA). In the second stage, we use YOLOv5 to detect agave plants and perform counting through YOLOv5 inference. According to numerical experiments, our approach in the first stage achieves 99% accuracy in identifying agave cultivation rows and 96.60% accuracy in counting agave plants.</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

Integrating remote sensing, image processing and machine learning algorithms for the recognition and counting of agave plants

  • 期刊论文

  • 发表日期：

    2025年7月25日

查看全文
  • Gerardo Asael López-Alfaro,
  • Juan Pablo Serrano-Rubio,
  • Luz María Rodríguez-Vidal,
  • Rafael Herrera Guzmán

摘要

Tequila and the agave plant are important symbols of Mexico’s national identity. In the last decade, the care of the Tequilana Weber Blue Agave has improved and the production of tequila has increased. In this paper, we introduce a methodology for the detection and counting of agave plants based on the use of Unmanned Aerial Vehicles (UAV) known as drones, image processing techniques, and machine learning. Several landscape images of agave crops are collected from the Romita region in Guanajuato state. These images are processed to obtain orthomosaics that allow us to study the target region of agave crops. Our proposal is carried out in two stages. In the first stage, we identify agave cultivation rows from other sections of the field to simplify the detection of agave plants using a superpixel approach and the Random Forest Classifier Algorithm (RFCA). In the second stage, we use YOLOv5 to detect agave plants and perform counting through YOLOv5 inference. According to numerical experiments, our approach in the first stage achieves 99% accuracy in identifying agave cultivation rows and 96.60% accuracy in counting agave plants.