Logo des Repositoriums
 
Konferenzbeitrag

Digital weed reduction

Vorschaubild nicht verfügbar

Volltext URI

Dokumententyp

Text/Conference Paper

Zusatzinformation

Datum

2023

Zeitschriftentitel

ISSN der Zeitschrift

Bandtitel

Verlag

Gesellschaft für Informatik e.V.

Zusammenfassung

In an effort to reduce the usage of herbicides in herb cultivation, we discuss strategies to detect and eliminate toxic weeds, exemplified by Senecio vulgaris. This paper presents a method to classify plants based on their spectral characteristics utilizing hyperspectral imagery in the range between 400 nm and 1100 nm. We are able to remove background material by masking it automatically using the well-established NDVI and Otsu’s thresholding prior to classification. Based on a neural network, we correctly classify 93% of the leaf area from Senecio vulgaris even with a relatively low spectral resolution of 50 nm. Apart from data collection, all steps were implemented in open source software and the Python language using open source libraries.

Beschreibung

Burkhart, Sebastian; Noack, Patrick (2023): Digital weed reduction. 43. GIL-Jahrestagung, Resiliente Agri-Food-Systeme. Bonn: Gesellschaft für Informatik e.V.. PISSN: 1617-5468. ISBN: 978-3-88579-724-1. pp. 297-302. Osnabrück. 13.-14. Februar 2023

Zitierform

DOI

Tags