A Simple Drone Survey and Image Processing Approach to Identifying St. John's Wort (<em>Hypericum Perforatum</em>) on Grazing Land in the Hunter Valley NSW, Australia

Abstract

This paper presents a simple drone survey and image processing approach to identifying St. John's wort (Hypericum perforatum) on grazing land in the Hunter Valley NSW, Australia. St John&amp;rsquo;s wort is an invasive species (weed) that competes with pasture, poisons livestock, can downgrade wool with &amp;lsquo;vegetable fault&amp;rsquo;, and decreases property values. Identifying the locations of St John wort from the ground can be difficult due to topography, limited access, and/or larger land areas of mixed vegetation. In this study, a drone was used to survey a 174-ha grazing property in the Hunter Valley NSW (Australia). The images were stitched together using commercially available software. A unique Visual Atmospheric Resistance Index (VARI) attribute was identified and used to highlight the presence of St John&amp;rsquo;s wort in the survey area. These sub-areas were then digitised onto high resolution maps for future planning use by land managers.