Camera System geometry for site specific treatment inprecision agriculture
- Gonzalo Pajares 1
- José M. Guerrero 1
- Martín Montalvo 1
- Ángela Ribeiro 2
- Pedro Javier Herrera Caro
- María Guijarro 1
- Juan Romeo 1
- 1 University Complutense of Madrid. Facultad Informática. Madrid. Spain
- 2 Centre for Automation and Robotics (UPM-CSIC),
- Pablo Gonzalez-de-Santos (ed. lit.)
- Gilles Rabatel (ed. lit.)
Editorial: S.N.
ISBN: 978-84-615-6184-1
Ano de publicación: 2011
Páxinas: 73-84
Congreso: International Workshop on Robotics and Associated High-technologies and Equipment por Agriculture (1. 2011. Montpellier)
Tipo: Achega congreso
Resumo
A visual perception system, on board a ground mobile unit, is used fordetecting rows in maize crops. Camera-based system geometry, under perspectiveprojection, is exploited to guide the image analysis. The aim is to apply priorknowledge to facilitate the detection and location of weeds in crops. Thisknowledge is used to determine where the crop rows should be located to thesearch for plants (crop and weeds), thus reducing the computational cost from thepoint of view of image processing, i.e. oriented toward real-time image processing.From the point of view of crop/weed identification the most suitable configurationis one that achieves maximum image resolution for an identical area in the field,since with higher resolution better decisions can be made about crop/weeddetection. In this paper we study and compare two camera devices, based onimage geometry, to verify their behaviour with different intrinsic and extrinsicparameters. This allows us to identify the best geometric image configuration fromthe point of view of highest image resolution, making the main contribution of thiswork.