Publications
NIBIOs employees contribute to several hundred scientific articles and research reports every year. You can browse or search in our collection which contains references and links to these publications as well as other research and dissemination activities. The collection is continously updated with new and historical material.
2019
Abstract
Advances in techniques for automated classification of point cloud data introduce great opportunities for many new and existing applications. However, with a limited number of labelled points, automated classification by a machine learning model is prone to overfitting and poor generalization. The present paper addresses this problem by inducing controlled noise (on a trained model) generated by invoking conditional random field similarity penalties using nearby features. The method is called Atrous XCRF and works by forcing a trained model to respect the similarity penalties provided by unlabeled data. In a benchmark study carried out using the ISPRS 3D labeling dataset, our technique achieves 85.0% in term of overall accuracy, and 71.1% in term of F1 score. The result is on par with the current best model for the benchmark dataset and has the highest value in term of F1 score. Additionally, transfer learning using the Bergen 2018 dataset, without model retraining, was also performed. Even though our proposal provides a consistent 3% improvement in term of accuracy, more work still needs to be done to alleviate the generalization problem on the domain adaptation and the transfer learning field.
Abstract
No abstract has been registered
Abstract
No abstract has been registered
Abstract
No abstract has been registered
Abstract
No abstract has been registered
Authors
Xiao Huang Höglind Mats Knut Bjørkelo Torben Christensen Kjetil Fadnes Teresa Gómez de la Bárcena Åsa Kasimir Leif Klemedtsson Bjørn Kløve Anders Lyngstad Mikhail Mastepanov Hannu Marttila Marcel van Oijen Peter Petros Ina Pohle Jagadeesh Yeluripati Hanna Marika SilvennoinenAbstract
Cultivated organic soils account for ∼7% of Norway’s agricultural land area, and they are estimated to be a significant source of greenhouse gas (GHG) emissions. The project ‘Climate smart management practices on Norwegian organic soils’ (MYR), commissioned by the Research Council of Norway (decision no. 281109), aims to evaluate GHG (e.g. carbon dioxide, methane and nitrous oxide) emissions and impacts on biomass productivity from three land use types (cultivated, abandoned and restored) on organic soils. At the cultivated sites, impacts of drainage depth and management intensity will be measured. We established experimental sites in Norway covering a broad range of climate and management regimes, which will produce observational data in high spatiotemporal resolution during 2019-2021. Using state-of-the-art modelling techniques, MYR aims to predict the potential GHG mitigation under different scenarios. Four models (BASGRA, DNDC, Coup and ECOSSE) will be further developed according to the soil properties, and then used independently in simulating biogeochemical processes and biomass dynamics in the different land uses. Robust parameterization schemes for each model will be based in the observational data from the project for both soil and crop combinations. Eventually, a multi-model ensemble prediction will be carried out to provide scenario analyses by 2030 and 2050. By integrating experimental results and modelling, the project aims at generating useful information for recommendations on environment-friendly use of Norwegian peatlands.
Authors
Xiao Huang Mats Höglind Knut Bjørkelo Torben Christensen Kjetil Fadnes Teresa Gómez de la Bárcena Åsa Kasimir Leif Klemedtsson Bjørn Kløve Anders Lyngstad Mikhail Mastepanov Hannu Marttila Marcel Van Oijen Peter Petros Ina Pohle Jagadeesh Yeluripati Hanna Marika SilvennoinenAbstract
Cultivated organic soils (7-8% of Norway’s agricultural land area) are economically important sources for forage production in some regions in Norway, but they are also ‘hot spots’ for greenhouse gas (GHG) emissions. The project ‘Climate smart management practices on Norwegian organic soils’ (MYR; funded by the Research Council of Norway, decision no. 281109) will evaluate how water table management and the intensity of other management practices (i.e. tillage and fertilization intensity) affects both GHG emissions and forage’s quality & production. The overall aim of MYR is to generate useful information for recommendations on climate-friendly management of Norwegian peatlands for both policy makers and farmers. For this project, we established two experimental sites on Norwegian peatlands for grass cultivation, of which one in Northern (subarctic, continental climate) and another in Southern (temperate, coastal climate) Norway. Both sites have a water table level (WTL) gradient ranging from low to high. In order to explore the effects of management practices, controlled trials with different fertilization strategies and tillage intensity will be conducted at these sites with WTL gradients considered. Meanwhile, GHG emissions (including carbon dioxide, methane and nitrous oxide), crop-related observations (e.g. phenology, production), and hydrological conditions (e.g. soil moisture, WTL dynamics) will be monitored with high spatiotemporal resolution along the WTL gradients during 2019-2021. Besides, MYR aims at predicting potential GHG mitigation under different scenarios by using state-of-the-art modelling techniques. Four models (BASGRA, Coup, DNDC and ECOSSE), with strengths in predicting grass growth, hydrological processes, soil nitrification-denitrification and carbon decomposition, respectively, will be further developed according to the soil properties. Then these models will be used independently to simulate biogeochemical and agroecological processes in our experimental fields. Robust parameterization schemes will be based on the observational data for both soil and crop combinations. Eventually, a multi-model ensemble prediction will be carried out to provide scenario analyses by 2030 and 2050. We will couple these process-based models with optimization algorithm to explore the potential reduction in GHG emissions with consideration of production sustenance, and upscale our assessment to regional level.
Authors
Valentina Gallina Silvia Torresan Alex Zabeo Jonathan Rizzi Sandro Carniel Mauro Sclavo Lisa Pizzol Antonio Marcomini Andrea CrittoAbstract
No abstract has been registered
Abstract
No abstract has been registered
Authors
Juraj Parajka Nejc Bezak John Burkhart Bjarki Hauksson Ladislav Holko Yeshewa Hundecha Michal Jenicek Pavel Krajčí Walter Mangini Peter Molnar Philippe Riboust Jonathan Rizzi Aynur Sensoy Guillaume Thirel Alberto ViglioneAbstract
No abstract has been registered