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
A novel method for age-independent site index estimation is demonstrated using repeated single-tree airborne laser scanning (ALS) data. A spruce-dominated study area of 114 km2 in southern Norway was covered by single-tree ALS twice, i.e. in 2008 and 2014. We identified top height trees wall-to-wall, and for each of them we derived based on the two heights and the 6-year period length. We reconstructed past, annual height growth in a field campaign on 31 sample trees, and this showed good correspondence with ALS based heights. We found a considerable increase in site index, i.e. about 5 m in the H40 system, from the old site index values. This increase corresponded to a productivity increase of 62%. This increase appeared to mainly represent a real temporal trend caused by changing growing conditions. In addition, the increase could partly result from underestimation in old site index values. The method has the advantages of not requiring tree-age data, of representing current growing conditions, and as well that it is a cost-effective method with wall-towall coverage. In slow-growing forests and short time periods, the method is least reliable due to possible systematic differences in canopy penetration between repeated ALS scans.
Abstract
Agroforestry can be defined as sustainable and multifunctional land-use systems where trees are managed together with agricultural crops or livestock on the same piece of land. In the northern periphery area, agroforestry has a long history with woodland grazing, reindeer husbandry and gathering of different non-wood forest resources as herbs, mushrooms and berries. Traditional agroforestry has gradually disappeared during the 20th century with the intensification of agriculture and forestry. Currently agroforestry systems are gaining new interest, not only from farmers but also from politicians, as this practice can possibly contribute to a more sustainable way of agricultural production. In the northern periphery area, the benefits of agroforestry practices can be manifold not only promoting traditional practices, but also novel systems with the use of new technology. In addition, agroforestry has environmental benefits as a method for conservation and enhancement of biodiversity, improved nutrient cycling, and water quality. Soil humus layer will also increase with several agroforestry systems leading to carbon sequestration. Here we present an overview of agroforestry practices in the Nordic countries and the use of non-wood forest resources with the emphasis on wild berries.
Authors
Stig A. BorgvangAbstract
No abstract has been registered
Authors
Hanno Sandvik Dag Dolmen Reidar Elven Tone Falkenhaug Elisabet Forsgren Haakon Hansen Kristian Hassel Vivian Husa Gaute Kjærstad Frode Ødegaard Hans Christian Pedersen Halvor Solheim Bård Gunnar Stokke Per Arvid Åsen Sandra Charlotte Helene Åström Tor Erik Brandrud Hallvard Elven Anders Endrestøl Anders Gravbrøt Finstad Stein Fredriksen Øivind Gammelmo Jan Ove Gjershaug Bjørn Gulliksen Inger Sofie Hamnes Bjørn Arild Hatteland Hanne Hegre Trygve H. Hesthagen Anders Jelmert Thomas Correll Jensen Stein Ivar Johnsen Egil Karlsbakk Christer Magnusson Kjell Harald Nedreaas Björn Nordén Eivind Oug Oddvar Pedersen Per Anker Pedersen Kjersti Sjøtun Jon Kristian Skei Heidi Solstad Leif Sundheim Jon Swenson Per Ole Syvertsen Venche Talgø Vigdis Vandvik Kristine Bakke Westergaard Rupert Wienerroither Bjørnar Ytrehus Olga Hilmo Snorre Henriksen Lisbeth GederaasAbstract
We present the results of an inventory and status assessment of alien species in Norway. The inventory covered all known multicellular neobiota, 2496 in total, 1039 of which were classified as naturalised. The latter constitute c. 3% of all species known to be stably reproducing in Norway. These figures are higher than expected from Norway’s latitude, which may be due a combination of climatic and historical factors, as well as sampling effort. Most of the naturalised neobiota were plants (71%),followed by animals (21%) and fungi (8%). The main habitat types colonised were open lowlands (79%), urban environments (52%) and woodlands (42%). The main areas of origin were Europe (67%), North America (15%) and Asia (13%). For most taxa, the rate of novel introductions seems to have been increasing during recent decades. Within Norway, the number of alien species recorded per county was negatively correlated with latitude and positively correlated with human population density. In the high-Arctic territories under Norwegian sovereignty, i.e. Svalbard and Jan Mayen, 104 alien species were recorded, of which 5 were naturalised.
Abstract
Sorption could be a way to concentrate nutrients in diluted waste streams to bring more nutrients back to agriculture. However, the sorbed nutrients must be plant available. The aim of this work was to investigate how plant available nitrogen (N) added sorbed to zeolite and is compared to conventionally added N. First, 15N labelled ammonium was sorbed to a sorbent, zeolite, in an aqueous solution. Then, the fertilizer effect was compared to the ammonium fertilizer and added the conventional way, with and without zeolite. A pot experiment with two soil types (chernozem and sandy soil) and wheat as test crop was used. Results indicated that the fertilizer effect of sorbed ammonium in the first growth cycle is about 50% of ammonium added conventionally. The sorbent itself had a positive effect in sandy soil, but not in chernozem. N uptake without added N was higher in chernozem than in sandy soil and more N from fertilizer was left in the soil after the experiment in the chernozem than in the sandy soil. In conclusion, ammonium added sorbed is plant available to some extent, but less so than conventionally added ammonium.
Abstract
Bipolar surface EMG (sEMG) signals of the trapezius muscles bilaterally were recorded continuously with a frequency of 800 Hz during full-shift field-work by a four-channel portable data logger. After recordings of 60 forest machine operators in Finland, Norway and Sweden, we discovered erroneous data. In short of any available procedure to handle these data, a method was developed to automatically discard erroneous data in the raw data reading files (Discarding Erroneous EPOchs (DESEPO) method. The DESEPO method automatically identifies, discards and adjusts the use of signal disturbances in order to achieve the best possible data use. An epoch is a 0.1 s period of raw sEMG signals and makes the basis for the RMS calculations. If erroneous signals constitute more than 30% of the epoch signals, this classifies for discharge of the present epoch. Non-valid epochs have been discarded, as well as all the subsequent epochs. The valid data for further analyses using the automatic detection resulted in an increase of acceptable data from an average of 2.15–6.5 h per day. The combination of long-term full-shift recordings and automatic data reduction procedures made it possible to use large amount of data otherwise discarded for further analyses.
Abstract
No abstract has been registered
Authors
Habtamu AlemAbstract
No abstract has been registered