AWeDeC - Automated Weed Detection in Crops


A weed is any unwanted plant that has no economic value and naturally encroaches into fields cultivated for various crops. Weeds deprive crop plants of resources necessary for their full potential yield. These are generally resilient plants that spread on their own and compete with less resilient crops mainly for water, light, nutrients, and carbon dioxide. Weeds, if not detected and curbed at appropriate time, are responsible for crop yield losses in the range of 14 - 42%. Each weed plant species has unique characteristics such as leaf shape, leaf color, venation of leaves, and size. There are two major families of weeds: Broadleaf weeds and narrowleaf weeds. A few common weeds belonging to these families are Phalaris minor (Dumbi Sitti), Avena Fatuta (Jangli Javi), Rumex (Palak), Chenopodium Album (Batu), Matri etc.

The automatic detection of weed by means of visual classification of terrain can aid in a focused effort towards its eradication (weed removal spray) and thus in reducing the yield loss. Visual classification is an inherently difficult task because of the diversity in visible characteristics of weeds. The focus of the proposed project will be on weed detection using aerial and land machines equipped with camera systems.

Verantwortliche Mitarbeiter

  • Dr.-Ing. Syed Atif Mehdi


  • FAST – National University of Computer & Emerging Sciences, Lahore, Pakistan