Neural Network Based Weeding Robot for Crop and Weed Discrimination

Abstract

Weeding is the most difficult process in the field of agriculture, which may consume more manpower, time and money. This paper is designed for detecting out weeds in-between crops and drilling them using blade attached at the end of robotic. Though this robot works autonomously, this does not require more manpower. It offers reduced working time and requires less installation cost. Insufficient labour, improper weeding, pandemic situation like Covid19 are some problems which can be avoided by this robot. This form of weeding is eco-friendly and does not affect the organic nature of the crop. It uses Convolution neural network(CNN) for detecting the image of weed from the crop.