Weed detection using YOLOv3 and elimination using organic weedicides with Live feed on Web App

Abstract

The article discusses the pressing issues in agri- culture, particularly highlighting the significance of detecting and categorizing weeds. Weeds pose a threat by competing with crops for vital nutrients, traditionally addressed through manual detection and herbicide application. However, recent technological progress has focused on automating weed detection using methods such as YOLOV3, a CNN-based object detection technique. In addition, the article introduces a fresh approach that utilizes linear actuators and organic weedicides for weed control. It evaluates this system’s effectiveness in terms of preci- sion and dynamic intrarow weeding through various analyses and experimental trials, demonstrating high accuracy and efficiency in real field scenarios. The live video footage of weed detection and removal is also showcased on a web application, providing users with information on the number of weeds eliminated. This integration of technological and chemical solutions presents a promising strategy for managing weeds in agriculture.