%0 Journal Article
	%A Morten Stigaard Laursen and  Rasmus Nyholm Jørgensen and  Mads Dyrmann and  Robert Poulsen
	%D 2017
	%J International Journal of Agricultural and Biosystems Engineering
	%B World Academy of Science, Engineering and Technology
	%I Open Science Index 124, 2017
	%T RoboWeedSupport-Sub Millimeter Weed Image Acquisition in Cereal Crops with Speeds up till 50 Km/H
	%U https://publications.waset.org/pdf/10007027
	%V 124
	%X For the past three years, the Danish project,
RoboWeedSupport, has sought to bridge the gap between the potential
herbicide savings using a decision support system and the required
weed inspections. In order to automate the weed inspections it is
desired to generate a map of the weed species present within the
field, to generate the map images must be captured with samples
covering the field. This paper investigates the economical cost of
performing this data collection based on a camera system mounted
on a all-terain vehicle (ATV) able to drive and collect data at up to 50
km/h while still maintaining a image quality sufficient for identifying
newly emerged grass weeds. The economical estimates are based on
approximately 100 hectares recorded at three different locations in
Denmark. With an average image density of 99 images per hectare
the ATV had an capacity of 28 ha per hour, which is estimated to cost
6.6 EUR/ha. Alternatively relying on a boom solution for an existing
tracktor it was estimated that a cost of 2.4 EUR/ha is obtainable under
equal conditions.
	%P 317 - 321