Back to search

Cornell University

WEED IDENTIFICATION USING IMAGE ANALYSIS AND MACHINE LEARNING

Abstract

dc:description.abstract

Increasing cases of herbicide-resistant weeds and the labor-intensive nature of traditional weed control methods necessitate more efficient and accurate solutions in agriculture. This thesis addresses these challenges by evaluating YOLO series algorithms (YOLOv5 to YOLOv10) to determine the most effective model for detecting glyphosate-resistant Palmer amaranth, identifying YOLOv9 as the best performer with the highest mean average precision (mAP). We then compared YOLOv8, YOLOv9, YOLOv10, and Faster R-CNN models for multi-species weed detection and found YOLOv9 to be the most robust model. Finally, the study investigated the use of YOLOv9 algorithm for evaluating the effectiveness of various PRE and POST herbicides for weed control in 2,4-D/glufosinate/glyphosate-resistant (Enlist) soybeans. automates weed control ratings in Enlist soybean fields by correlating weed counts from drone imagery with visual control ratings. Altogether, the findings of this research demonstrated the potential for significantly improving weed management practices through species-specific weed identification using YOLO based object detection models highlighting the transformative impact of advanced machine learning algorithms on developing sustainable and economic weed management practices.

Degree

thesis:*
Name thesis:degree_name
M.S., Soil and Crop Sciences
Level thesis:degree_level
Master of Science
Discipline thesis:degree_discipline
Soil and Crop Sciences
Grantor
Cornell University
Year dc:date.issued
2024

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Sharma, Akhilesh
Chair dc:contributor.chair
  • Kumar, Vipan
Committee members dc:contributor.committeemember
  • Hariharan, Bharath
  • Longchamps, Louis

Subjects

dc:subject × 6

Rights

dc:rights
Statement dc:rights
  • Attribution-NonCommercial-ShareAlike 4.0 International
Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Bibid
16611813
Dc Identifier Other
ProQuest Submission ID: 12226
ProQuest Publication ID: 31488193

Chain of custody

source
Harvested from
Cornell University
Base URL
ecommons.cornell.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Sharma, Akhilesh. WEED IDENTIFICATION USING IMAGE ANALYSIS AND MACHINE LEARNING. Master of Science thesis, Cornell University, 2024. https://hdl.handle.net/1813/116327