home    about    browse    search    latest    help 
Login | Create Account

Vision-based weed identification with farm robots

Swain, Kishore C.; Nørremark, Michael; Bochtis, Dionysis; Sørensen, Claus Grøn and Green, Ole (2010) Vision-based weed identification with farm robots. In: XVIIth World Congress of the International Commission of Agricultural and Biosystems Engineering, Book of Abstracts, CIGR XVIIth World Congress, p. 11.

PDF - English


Robots in agriculture offer new opportunities for real time weed identification and quick removal operations. Weed identification and control remains one of the most challenging task in agriculture, particularly in organic agriculture practices. Considering environmental impacts and food quality, the excess use of chemicals in agriculture for controlling weeds and diseases is decreasing. The cost of herbercides and their field applications must be optimized. As an alternative, a smart weed identification technique followed by the mechanical and thermal weed control can fulfill the organic farmers’ expectations. The smart identification technique works on the concept of ‘shape matching’ and ‘active shape modeling’ of plant and weed leafs. The automated weed detection and control system consists of three major tools. Such as: i) eXcite multispectral camera, ii) LTI image processing library and iii) Hortibot robotic vehicle. The components are combined in Linux interface environment in the eXcite camera associate PC. The laboratory experiments for active shape matching have shown interesting results which will be further enhanced to develop the automated weed detection system. The Hortibot robot will be mounted with the camera unit in the front-end and the mechanical weed remover in the rear-end. The system will be upgraded for intense commercial applications in maize and other row crops.

EPrint Type:Conference paper, poster, etc.
Type of presentation:Paper
Subjects: Crop husbandry > Weed management
Research affiliation: Denmark > DARCOF III (2005-2010) > WEEDS - Control of weeds in organic cropping
Denmark > AU - Aarhus University > AU, DJF - Faculty of Agricultural Sciences
Deposited By: Nørremark, Michael
ID Code:20653
Deposited On:26 Mar 2012 13:18
Last Modified:28 Mar 2012 13:41
Document Language:English
Refereed:Not peer-reviewed

Repository Staff Only: item control page