Identification of weeds using Hsv color spaces and labelling with machine learning algorithms

Abstract

The goal of this project is to detect the weeds in the farmland, for proper distribution of sparing of herbicides in the farm. The crops are separated from the weeds with their color and feature of their appearance. In that cases the features of the weeds are extracted with HSV color space method, it produces higher accuracy comparing to RGB color space model. The extracted feature is compared with the trained data in Neural Networks for more accurate results comparing to SVM or BP methods. NN is used to divide the images into pixel for more accurate value. It can produce maximum of 95% accuracy comparing to other methods.