Automatic weed mapping in cereal fields using image processing techniques
Herbicide spraying is necessary to optimise crop production in cereal fields. By using a camera and image processing the weed-infected areas can be detected and only the infected areas are sprayed, thus providing environmental and economic benefits.
It is known that weeds in cereal fields are not evenly distributed, but rather appear in patches. By treating only weed-infected areas, the herbicide usage can be reduced by as much as 35%. In addition to the obvious environmental gain, such a reduction will have a positive effect on the crop production.
In collaboration with DAT AS (Dimensions Agri Technologies), Adigo AS and the The Plant Health and Plant Protection Centre at Norwegian Institute for Agricultural and Environmental Research , SINTEF has developed a camera-based system for on-line mapping of weeds in cereal fields. We have shown that by it is possible to use image-processing techniques to differentiate between plants and soil and between cereals and weeds under variable outdoor conditions. For optimal herbicide treatment of the field, it is also necessary to detect the most harmful weed species. These are species, which require treatment even when they are present at quite low densities. By using contour-based and feature based techniques, we are able to distinguish between harmful species and less harmful ones.
See Adigos System Description for more details.
For further information please contact Kristin Kaspersen
.
Example of an image captured and analysed by the system developed at SINTEF
