Detection of Parthenium Weed (Parthenium hysterophorus L.) and Its Growth Stages Using Artificial Intelligence

Parthenium weed (Parthenium hysterophorus L. (Asteraceae)), native to the Americas, is in top 100 most invasive plant species world. In Australia, it an annual (herb/shrub) of national significance, especially state Queensland where has infested both agricultural and conservation lands, including riparian corridors. Effective control strategies for this (pasture management, biological control, herbicide usage) require populations be detected mapped. However, mapping made difficult due varying nature landscapes (e.g., uneven terrain). This paper proposes a novel method detect map parthenium simulated pastoral environments using Red-Green-Blue (RGB) and/or hyperspectral imagery aided by artificial intelligence. Two datasets were collected environment series naturally co-occurring, non-parthenium (monocot) plants. RGB images processed with YOLOv4 Convolutional Neural Network (CNN) implementation, achieving overall accuracy 95% detection, 86% classification flowering non-flowering stages weed. An XGBoost classifier was used pixel dataset—achieving 99% each growth stage class; all materials received discernible colour mask. When plants artificially combined various permutations, class, again receiving accurate Performance metrics indicate that our proposed processing pipeline can preliminary design detection strategies, extended automated airborne unmanned aerial vehicle (UAV) data. The findings also demonstrate potential controlled, glasshouse field.