Monitoring grassland invasion by spotted knapweed (Centaurea maculosa) with RPAS-acquired multispectral imagery

Abstract

The ability to accurately detect and quantify the presence of invasive plants is integral in their management, treatment, and removal. Remotely piloted aircraft systems (RPASs) are becoming an important remote sensing tool for mapping invasive plants. Spotted knapweed (Centaurea maculosa) is highly invasive in North America. This study developed and evaluated a novel method for analysis of multispectral data to map the relative cover of spotted knapweed in a heterogeneous grassland community. The method developed in this work, termed metapixel-based image analysis, segments the image into a grid of metapixels for which grey level co-occurrence matrix (GLCM)-based statistics can be computed as descriptive features. Using RPAS-acquired multispectral imagery and plant species inventories performed on 1m² quadrats, a random forest classifier was trained to predict the qualitative degree of spotted knapweed ground cover within each metapixel. The best mean cross-validation score achieved was 71.3% when describing relative ground cover of spotted knapweed, with an accuracy of 66.0% when applied to an independent validation dataset. Analysis of the performance of metapixel-based image analysis on this study site suggests that feature optimization, including feature subset selection, and the use of GLCM-based texture features is of critical importance for achieving an accurate classification.

Plant phenotyping relevance

RPASマルチスペクトル画像から植物種の相対被覆度を推定する画像解析法を開発し、独立データで検証しており、植物状態の取得方法が中心的です。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.