<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>221</id>
    <completedYear>2021</completedYear>
    <publishedYear/>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst/>
    <pageLast/>
    <pageNumber>25</pageNumber>
    <edition/>
    <issue>1</issue>
    <volume>14</volume>
    <type>article</type>
    <publisherName>MDPI</publisherName>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>1</belongsToBibliography>
    <completedDate>2022-01-04</completedDate>
    <publishedDate>2021-12-24</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Detection of Windthrown Tree Stems on UAV-Orthomosaics Using U-Net Convolutional Networks</title>
    <abstract language="eng">The increasing number of severe storm events is threatening European forests. Besides the primary damages directly caused by storms, there are secondary damages such as bark beetle outbreaks and tertiary damages due to negative effects on the market. These subsequent damages can be minimized if a detailed overview of the affected area and the amount of damaged wood can be obtained quickly and included in the planning of clearance measures. The present work utilizes UAV-orthophotos and an adaptation of the U-Net architecture for the semantic segmentation and localization of windthrown stems. The network was pre-trained with generic datasets, randomly combining stems and background samples in a copy–paste augmentation, and afterwards trained with a specific dataset of a particular windthrow. The models pre-trained with generic datasets containing 10, 50 and 100 augmentations per annotated windthrown stems achieved F1-scores of 73.9% (S1Mod10), 74.3% (S1Mod50) and 75.6% (S1Mod100), outperforming the baseline model (F1-score 72.6%), which was not pre-trained. These results emphasize the applicability of the method to correctly identify windthrown trees and suggest the collection of training samples from other tree species and windthrow areas to improve the ability to generalize. Further enhancements of the network architecture are considered to improve the classification performance and to minimize the calculative costs.</abstract>
    <parentTitle language="eng">Remote Sensing</parentTitle>
    <identifier type="doi">10.3390/rs14010075</identifier>
    <identifier type="urn">urn:nbn:de:kobv:eb1-opus-2213</identifier>
    <identifier type="issn">2072-4292</identifier>
    <enrichment key="opus.source">publish</enrichment>
    <licence>Creative Commons - CC BY - Namensnennung 4.0 International</licence>
    <author>Stefan Reder</author>
    <author>Jan-Peter Mund</author>
    <author>Nicole Albert</author>
    <author>Lilli Waßermann</author>
    <author>Luis Miranda</author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>forest damage assessment</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>windthrow detection</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>deep learning</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>U-Net</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>UAV</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>semantic segmentation</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>natural disaster analysis</value>
    </subject>
    <collection role="open_access" number="">open_access</collection>
    <collection role="institutes" number="">Fachbereich Wald und Umwelt</collection>
    <collection role="Finanzierung" number=""/>
    <collection role="Hochschulbibliographie" number=""/>
    <collection role="Hochschulbibliographie" number="">Zweitveröffentlichung</collection>
    <collection role="Hochschulbibliographie" number="">Referiert</collection>
    <thesisPublisher>Hochschule für nachhaltige Entwicklung Eberswalde</thesisPublisher>
    <file>https://opus4.kobv.de/opus4-hnee/files/221/remotesensing-14-00075-v4.pdf</file>
  </doc>
  <doc>
    <id>251</id>
    <completedYear>2021</completedYear>
    <publishedYear/>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>71</pageFirst>
    <pageLast>80</pageLast>
    <pageNumber>10</pageNumber>
    <edition/>
    <issue/>
    <volume>35</volume>
    <type>article</type>
    <publisherName>Springer</publisherName>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>1</belongsToBibliography>
    <completedDate>2022-06-17</completedDate>
    <publishedDate>2021-01-27</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Prediction Error-Driven Memory Consolidation for Continual Learning: On the Case of Adaptive Greenhouse Models</title>
    <abstract language="eng">This work presents an adaptive architecture that performs online learning and faces catastrophic forgetting issues by means of an episodic memory system and of prediction-error driven memory consolidation. In line with evidence from brain sciences, memories are retained depending on their congruence with the prior knowledge stored in the system. In this work, congruence is estimated in terms of prediction error resulting from a deep neural model. The proposed AI system is transferred onto an innovative application in the horticulture industry: the learning and transfer of greenhouse models. This work presents models trained on data recorded from research facilities and transferred to a production greenhouse.</abstract>
    <parentTitle language="eng">KI - Künstliche Intelligenz</parentTitle>
    <identifier type="doi">10.1007/s13218-020-00700-8</identifier>
    <identifier type="issn">1610-1987</identifier>
    <identifier type="urn">urn:nbn:de:kobv:eb1-opus-2510</identifier>
    <enrichment key="opus.source">publish</enrichment>
    <licence>Creative Commons - CC BY - Namensnennung 4.0 International</licence>
    <author>Guido Schillaci</author>
    <author>Uwe Schmidt</author>
    <author>Luis Miranda</author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Adaptive models</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Deep neural networks</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Episodic memory</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Memory consolidation</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Greenhouse</value>
    </subject>
    <collection role="ddc" number="006">Spezielle Computerverfahren</collection>
    <collection role="open_access" number="">open_access</collection>
    <collection role="institutes" number="">Fachbereich Wald und Umwelt</collection>
    <collection role="Hochschulbibliographie" number=""/>
    <collection role="Hochschulbibliographie" number="">Zweitveröffentlichung</collection>
    <collection role="Hochschulbibliographie" number="">Referiert</collection>
    <thesisPublisher>Hochschule für nachhaltige Entwicklung Eberswalde</thesisPublisher>
    <file>https://opus4.kobv.de/opus4-hnee/files/251/Schillaci2021_Article.pdf</file>
  </doc>
</export-example>
