Abstract
Precision weeding is one of the most promising applications for autonomous service robots in biological production. Herbicides have been the default weeding solution during the past decades, but there is a growing concern about the environmental impact on drinking water reservoirs etc. The use of computer vision and precision spraying technology makes it possible to significantly reduce the consumption of herbicides.
The work presented here is part of a project with the purpose of performing autonomous precision spraying trials. In this work a novel cell sprayer designed for large scale tests of weed detection algorithms and spraying strategies is developed. The front part of the cell sprayer implement is a camera module mounted in a shock absorbing frame. The camera module is followed by a spraying module containing nozzles with individual valve control. The implement is designed for use with an Armadillo robotic tool carrier consisting of two battery powered track modules mounted on each side of the implement.
This paper focus on the cell sprayer implement design including camera system, sprayer module and integration with the service robot and the robot software. The FroboMind software platform and Armadillo robot is used and it is hypothesized that utilizing FroboMind the cell sprayer can drive smoothly through a test field with a lateral positioning accuracy better than 50 mm.
A precision spraying trial in a 1 Ha maize field using different treatment methods was used for testing the hypothesis. The results substantiates that the cell sprayer is capable of navigating the test field smoothly with a lateral positioning accuracy better than 50 mm. The test results shows that over a distance of approximately 145 m the 95% percentile of the absolute distance to the AB line is 0.0250 m and the 95% percentile of the absolute heading error is 1.17 degrees which is within the requirements of the cell spray implement.
The results from the precision spraying trail with respect to weeding performance will be published in a later publication and the cell sprayer is expected to be utilized in further precision spraying experiments.
The work presented here is part of a project with the purpose of performing autonomous precision spraying trials. In this work a novel cell sprayer designed for large scale tests of weed detection algorithms and spraying strategies is developed. The front part of the cell sprayer implement is a camera module mounted in a shock absorbing frame. The camera module is followed by a spraying module containing nozzles with individual valve control. The implement is designed for use with an Armadillo robotic tool carrier consisting of two battery powered track modules mounted on each side of the implement.
This paper focus on the cell sprayer implement design including camera system, sprayer module and integration with the service robot and the robot software. The FroboMind software platform and Armadillo robot is used and it is hypothesized that utilizing FroboMind the cell sprayer can drive smoothly through a test field with a lateral positioning accuracy better than 50 mm.
A precision spraying trial in a 1 Ha maize field using different treatment methods was used for testing the hypothesis. The results substantiates that the cell sprayer is capable of navigating the test field smoothly with a lateral positioning accuracy better than 50 mm. The test results shows that over a distance of approximately 145 m the 95% percentile of the absolute distance to the AB line is 0.0250 m and the 95% percentile of the absolute heading error is 1.17 degrees which is within the requirements of the cell spray implement.
The results from the precision spraying trail with respect to weeding performance will be published in a later publication and the cell sprayer is expected to be utilized in further precision spraying experiments.
Originalsprog | Dansk |
---|---|
Publikationsdato | 3. jul. 2013 |
Status | Udgivet - 3. jul. 2013 |
Begivenhed | XXXV CIOSTA & CIGR V Conference - Billund, Danmark Varighed: 3. jul. 2013 → 5. jul. 2013 |
Konference
Konference | XXXV CIOSTA & CIGR V Conference |
---|---|
Land/Område | Danmark |
By | Billund |
Periode | 03/07/2013 → 05/07/2013 |