Colour based detection of volunteer potatoes as weeds in sugar beet fields using machine vision

A.T. Nieuwenhuizen, L. Tang, J.W. Hofstee, J. Müller, E.J. van Henten

Research output: Contribution to journalArticleAcademicpeer-review

35 Citations (Scopus)

Abstract

The possible spread of late blight from volunteer potato plants requires the removal of these plants from arable fields. Because of high labour, energy, and chemical demands, a method of automatic detection and removal is needed. The development and comparison of two colour-based machine vision algorithms for in-field volunteer potato plant detection in two sugar beet fields are discussed. Evaluation of the results showed that both methods gave closely matched results within fields, although large differences exist between the fields. At plant level, in one field up to 97% of the volunteer potato plants were correctly classified. In another field, only 49% of the volunteer plants were correctly identified. The differences between the fields were higher than the differences between the methods used for plant classification.
Original languageEnglish
Pages (from-to)267-278
JournalPrecision Agriculture
Volume8
Issue number6
DOIs
Publication statusPublished - 2007

Keywords

  • identification

Fingerprint Dive into the research topics of 'Colour based detection of volunteer potatoes as weeds in sugar beet fields using machine vision'. Together they form a unique fingerprint.

Cite this