Scholarlink

<p>The removal of weed (unwanted plant) in the agricultural field is time-consuming and needs the significant resources for it. The existing methods are not much efficient enough. This paper provides the solution for the detection and removal of weed using the image processing techniques. The real time images from the field are captured and the AI model will process it for the detection of weed in it and sends the position of weed to the microcontroller—Jetson Nano which controls the weedicide sprayer movement. The process of identifying the location, existence and the class of the object they belong to using a bounding box is known as object detection. This detection technique creates a bounding box and allocates a label as weed for the object using YOLO v8 technique. The Integration of IoT in this project adds an advantage of monitoring of the real-time data remotely and able to take the actions accordingly. In the agricultural production, the AI-based methodologies made significant progress. It offers various benefits such as, the non-contact detection, increased efficiency &amp; performance, optimizing manpower &amp; cost, this deep learning-based technique becomes the effective solution for weed detection. YOLO represents a streamlined, one-stage technique designed especially for detecting objects. It also significantly improves the speed of the process by the use of CNN (Convolutional Neural Network) architecture, efficiently detects both the locations and the types of objects within the image. YOLO precisely and quickly detected weeds and crops present in the images at the 40 frames per second rate with the detection speed of about 10ms per image and mAP IOU@0.5 of 0.99.</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

A cloud-integrated deep learning system for weed detection using Jetson Nano

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月5日

查看全文
  • N. Dayanand Lal,
  • S. Maheswaran,
  • E. Devika,
  • A. N. Ashwini,
  • N. Indhumathi,
  • Leti Wakuma Turi

摘要

The removal of weed (unwanted plant) in the agricultural field is time-consuming and needs the significant resources for it. The existing methods are not much efficient enough. This paper provides the solution for the detection and removal of weed using the image processing techniques. The real time images from the field are captured and the AI model will process it for the detection of weed in it and sends the position of weed to the microcontroller—Jetson Nano which controls the weedicide sprayer movement. The process of identifying the location, existence and the class of the object they belong to using a bounding box is known as object detection. This detection technique creates a bounding box and allocates a label as weed for the object using YOLO v8 technique. The Integration of IoT in this project adds an advantage of monitoring of the real-time data remotely and able to take the actions accordingly. In the agricultural production, the AI-based methodologies made significant progress. It offers various benefits such as, the non-contact detection, increased efficiency & performance, optimizing manpower & cost, this deep learning-based technique becomes the effective solution for weed detection. YOLO represents a streamlined, one-stage technique designed especially for detecting objects. It also significantly improves the speed of the process by the use of CNN (Convolutional Neural Network) architecture, efficiently detects both the locations and the types of objects within the image. YOLO precisely and quickly detected weeds and crops present in the images at the 40 frames per second rate with the detection speed of about 10ms per image and mAP IOU@0.5 of 0.99.