Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Plant Recognition through the Fusion of 2D and 3D Images for Robotic Weeding

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2015 ASABE Annual International Meeting  152181371.(doi:10.13031/aim.20152181371)
Authors:   Jingyao Gai, Lie Tang, Brian Steward
Keywords:   Robotic weeding, computer vision, sensor fusion

Abstract. In crop production systems, weed management is vitally important. But both manual weeding and herbicide-based weed controlling are problematic due to concerns in cost, operator health, emergence of herbicide-resistant weed species, and environment impact. Automated robotic weeding offers a possibility of controlling weeds in a precise fashion, particularly for weeds growing near crops or within crop rows. However, identification and localization of plants have not yet been fully automated. The goal of this reported project is to develop a high-throughput plant recognition and localization algorithm by fusing 2D color and textural data with 3D point cloud data. Plant morphological models were developed and applied for plant recognition against different weed species at different growth stages.

(Download PDF)    (Export to EndNotes)