Automatic Weed Detection and Killing Robot

Abstract

Though becoming a pillar of the Indian economy, agriculture is currently having a production crisis. The issue of weeds in agriculture is a significant problem that can re- sult in substantial financial losses for farmers. Conventional methods of weed control, such as manual labor or chemical applications, can be time-consuming, demanding, and potentially harmful to the environment. To tackle these challenges, we have developed and implemented an automatic weed detection and killing robot. The robot employs computer vision algorithms to detect and categorize weeds instantly. It comprises a camera and a processing unit that captures images of the crop field and examines them to recognize weeds based on their visual features. The robot then uses a mechanical arm to administer a precise amount of herbicide to the weed, decreasing the use of chemicals and minimizing the possibility of harming non-target crops. This is accomplished by periodically taking pictures of the field, processing them on a Raspberry Pi board and using an image processing algorithm to distinguish the desired plants from the weeds based on a number of characteristics such as the size, colour, and cut. The Raspberry Pi chip sends a signal to the pesticide spraying pump to turn on when the weeds have been accurately discovered and identified through image analysis.