Publikasjoner
NIBIOs ansatte publiserer flere hundre vitenskapelige artikler og forskningsrapporter hvert år. Her finner du referanser og lenker til publikasjoner og andre forsknings- og formidlingsaktiviteter. Samlingen oppdateres løpende med både nytt og historisk materiale. For mer informasjon om NIBIOs publikasjoner, besøk NIBIOs bibliotek.
2015
Sammendrag
Vehicles which operate in agricultural row crops, need to strictly follow the established wheel tracks. Errors in navigation where the robot sways of its path with one or more wheels may damage the crop plants. The specific focus of this paper is on an agricultural robot operation in row cultures. The robot performs machine vision detecting weeds within the crop rows and treats the weeds by high precision drop-on-demand application of herbicide. The navigation controller of the robot needs to follow the established wheel tracks and minimize the camera system offset from the seed row. The problem has been formulated as a Nonlinear Model Predictive Control (NMPC) problem with the objective of keeping the vision modules centered over the seed rows, and constraining the wheel motion to the defined Wheel tracks. The system and optimization problem has been implemented in Python using the Casadi framework. The implementation has been evaluated through simulations of the system, and compared with a PD controller. The NMPC approach display advantages and better performance when facing the path constraints of operating in row crops.
Sammendrag
Vehicles which operate in agricultural row crops, need to strictly follow the established wheel tracks. Errors in navigation where the robot sways of its path with one or more wheels may damage the crop plants. The specific focus of this paper is on an agricultural robot operation in row cultures. The robot performs machine vision detecting weeds within the crop rows and treats the weeds by high precision drop-on-demand application of herbicide. The navigation controller of the robot needs to follow the established wheel tracks and minimize the camera system offset from the seed row. The problem has been formulated as a Nonlinear Model Predictive Control (NMPC) problem with the objective of keeping the vision modules centered over the seed rows, and constraining the wheel motion to the defined Wheel tracks. The system and optimization problem has been implemented in Python using the Casadi framework. The implementation has been evaluated through simulations of the system, and compared with a PD controller. The NMPC approach display advantages and better performance when facing the path constraints of operating in row crops.
Sammendrag
The purpose of our study is to explore the possibility to use proximate RGB imagery as basis for site-specific management of perennial weeds in small grain cereals. The targeted species are the broadleaved weeds Cirsium arvense (L.) Scop. (Creeping thistle) and Sonchus arvensis L. (Perennial sowthistle) and the grass weed Elymus repens (L.) Gould. (Common quackgrass). These are the main challenges for perennial weed control in cereals in Norway and temperate zone. The overall idea is to make weed maps based on images acquired during harvest in autumn (August/September) and use these maps for site-specific weed management when these species are normally managed in Norway, i.e. 3-4 weeks after harvest (E. repens) or in the following spring, i.e. late May/early June (C. arvense and S. arvensis). An on-the-go weed detection and glyphosate application in one operation before harvest is also a possible usage of our image-based method where this timing of glyphosate application is allowed. Images were acquired with a consumer grade camera mounted on a 3 m pole and tilted to mimic images acquired from the roof of a combine harvester. Images were acquired few days before harvest, a period where the cereals are yellowish and weed leaves and stalks are still green. Plots, 8 m by 8 m, were established in cereals to cover a wide range in weed pressure- and flora. The four plot corners were marked with white styrofoam balls mounted on sticks prior imaging and recorded with GPS (10 cm accuracy). The machine vision algorithm performs first a geometrical transform to represent the images as pseudo-orthonormal to the ground plane. This transform is aided by white styrofoam balls marking the corners of the plot with known distance. In the intended practical use, the transform can be done by obtaining the camera-angle and heading from inertial and GPS measurements and assuming level ground. The classification algorithm starts by segmenting the image into a class for green parts of the weeds (leaf, stalk), and three classes for flower heads (yellow, white and purple), by using threshold filters in the HSV colour space. A connected components analysis is then performed on each of the binary images, where the very small regions are filtered out. The area and centre of each region is calculated for comparison with ground truth observations. Two types of ground truth data for evaluation of the algorithm are available: Manual assessment of weed coverage from computer display of images and weed maps based on GPS measurements at the time for their management. Machine vision algorithm outputs versus ground truth data will be presented.
Forfattere
Therese With BergeSammendrag
Det er ikke registrert sammendrag
Foredrag – Site-specific weed management (Row crops, non-arable areas, grassland)
Therese With Berge
Forfattere
Therese With BergeSammendrag
Det er ikke registrert sammendrag
Forfattere
Nina TrandemSammendrag
Det er ikke registrert sammendrag
Sammendrag
Det er ikke registrert sammendrag
Forfattere
Nina TrandemSammendrag
Det er ikke registrert sammendrag
Forfattere
E.J. Ahrenfeldt B.K. Klatt J. Arildsen Nina Trandem G.K.S. Andersson T. Tscharntke H.G. Smith L. SigsgaardSammendrag
Det er ikke registrert sammendrag
Sammendrag
Det er ikke registrert sammendrag