Vegetation Mapping by UAV Visible Imagery and Machine Learning

TL;DR

This study develops a semi-automatic method using UAV visible imagery and machine learning to accurately identify and map vegetation species at high resolution, achieving over 90% efficiency at 5m altitude.

Contribution

It introduces a novel integration of UAV imagery and machine learning for detailed vegetation mapping, improving accuracy and efficiency.

Findings

Over 90% identification efficiency at 5m altitude

Effective use of expert masking and hue-filtering for training

Method can be integrated into VRHA tools

Abstract

An experimental field cropped with sugar-beet with a wide spreading of weeds has been used to test vegetation identification from drone visible imagery. Expert masked and hue-filtered pictures have been used to train several Machine Learning algorithms to develop a semi-automatic methodology for identification and mapping species at high resolution. Results show that 5m altitude allows for obtaining maps with an identification efficiency of more than 90%. Such a method can be easily integrated to present VRHA, as much as tools to obtain detailed maps of vegetation.