Scholarlink

Determining the indicators of crop development is an important task in the agricultural sector. This study proposes a method for determining the planting density and the leaf area index of sunflower plants using RGB images and deep learning techniques obtained using unmanned aerial vehicles (UAVs). For image analysis, the instance segmentation task was solved to count the detected objects and highlight their boundaries. Modern architectures such as YOLOv8 and YOLOv9 were considered. YOLOv8x demonstrated better accuracy on the test data. The analysis of the impact of image preprocessing methods such as Histogram Equalization and Contrast Accumulated Histogram Equalization on model performance was carried out. The results of the experiment indicate that the use of these methods can enhance the generalization ability of models.

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

The Use of Neural Networks Based on the YOLO Architecture to Automatically Determine Indicators of Crop Growth Using Photographs Taken from UAVs

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Dmitriy Poleshchenko,
  • Ilia Mikhailov,
  • Vladislav Petrov

摘要

Determining the indicators of crop development is an important task in the agricultural sector. This study proposes a method for determining the planting density and the leaf area index of sunflower plants using RGB images and deep learning techniques obtained using unmanned aerial vehicles (UAVs). For image analysis, the instance segmentation task was solved to count the detected objects and highlight their boundaries. Modern architectures such as YOLOv8 and YOLOv9 were considered. YOLOv8x demonstrated better accuracy on the test data. The analysis of the impact of image preprocessing methods such as Histogram Equalization and Contrast Accumulated Histogram Equalization on model performance was carried out. The results of the experiment indicate that the use of these methods can enhance the generalization ability of models.