Sleeping site selection is essential for understanding primate behavioral ecology and survival.Identifying where species sleep helps determine priority areas and critical resources for targeted conservation efforts.Ho...Sleeping site selection is essential for understanding primate behavioral ecology and survival.Identifying where species sleep helps determine priority areas and critical resources for targeted conservation efforts.However,observing sleeping sites at night is challenging,especially for species sensitive to human disturbance.Thermal infrared imaging(TIR)with drones is increasingly used for detecting and counting primates,yet it has not been utilized to investigate ecological strategies.This study investigates the sleeping site selection of the Critically Endangered black-shanked douc langur(Pygathrix nigripes)in Cát Tiên National Park,Vietnam.Our aim is to assess the feasibility of using a TIR drone to test sleeping site selection strategies in non-nesting primates,specifically examining hypotheses related to predation avoidance and food proximity.Between January and April 2023,we conducted 120 drone flights along 22 transects(~1-km long)and identified 114 sleeping sites via thermal imaging.We established 116 forest structure plots along 29 transects in non-selected sites and 65 plots within douc langur sleeping sites.Our observations reveal that douc langurs selected tall and large trees that may provide protection against predators.Additionally,they selected sleeping sites with increased access to food,such as Afzelia xylocarpa,which serves as a preferred food source during the dry season.These results highlight the effective use of TIR drones for studying douc langur sleeping site selection with minimal disturbance.Besides offering valuable insights into habitat selection and behavioral ecology for conservation,TIR drones hold great promise for the noninvasive and long-term monitoring of large-bodied arboreal species.展开更多
Utilizing unmanned aerial vehicle (UAV) photography to timely detect and evaluate potential safety hazards (PSHs) around high-speed rail has great potential to complement and reform the existing manual inspections by ...Utilizing unmanned aerial vehicle (UAV) photography to timely detect and evaluate potential safety hazards (PSHs) around high-speed rail has great potential to complement and reform the existing manual inspections by providing better overhead views and mitigating safety issues. However, UAV inspections based on manual interpretation, which heavily rely on the experience, attention, and judgment of human inspectors, still inevitably suffer from subjectivity and inaccuracy. To address this issue, this study proposes a lightweight hybrid learning algorithm named HDTA (hybrid dual tasks architecture) to automatically and efficiently detect the PSHs of UAV imagery. First, this HDTA architecture seamlessly integrates both detection and segmentation branches within a unified framework. This design enables the model to simultaneously perform PSH detection and railroad parsing, thereby providing comprehensive scene understanding. Such joint learning also lays the foundation for PSH assessment tasks. Second, an innovative lightweight backbone based on the shuffle selective state space model (S^(4)M) is incorporated into HDTA. The state space model approach allows for global contextual information extraction while maintaining linear computational complexity. Furthermore, the incorporation of shuffle operation facilitates more efficient information flow across feature dimensions, enhancing both feature representation and fusion capabilities. Finally, extensive experiments conducted on a railroad environment dataset constructed from UAV imagery demonstrate that the proposed method achieves high detection accuracy while maintaining efficiency and practicality.展开更多
The forest ecosystems of the Lacs 2 commune (South-East, Togo) are undergoing severe degradation, which has not yet been documented. This study is carried out in order to assess and quantify the spatio-temporal dynami...The forest ecosystems of the Lacs 2 commune (South-East, Togo) are undergoing severe degradation, which has not yet been documented. This study is carried out in order to assess and quantify the spatio-temporal dynamics of residual forests and to identify the determinants of deforestation in South East Togo. The methodological approach is based on the use of historical aerial photographs from 1976 and drone images from 2019 in addition to field investigations. Several spatial structure indices were also calculated in order to quantify the fragmentation of classes and of the forest landscape. The results show that the forest landscape is changing. The classes of forests, plantations and palm groves show an annual rate of decline of 7.5%, 0.8% and 9.4% respectively while the classes of savannahs, agglomerations, surface water and swamps increased by 16.4%, 0.4%, 0.7% and 0.1%. The results also reveal a high fragmentation within the forest, plantation, surface water and swamp class and moderate fragmentation for the savannah and palm trees classes. At the landscape level, the savannah class is dominant by more than 70%, thus making the landscape little diversified from an ecological point of view. The main driver of deforestation in the study area remains shifting slash-and-burn agriculture. It is accentuated by the establishment of perennial oil palm crops, which has influenced the annual deforestation rate by 0.72%.展开更多
基金financial support of the Belgian National Fund for Scientific Research(FNRS)the Duesberg Foundation,and the University of Liège.
文摘Sleeping site selection is essential for understanding primate behavioral ecology and survival.Identifying where species sleep helps determine priority areas and critical resources for targeted conservation efforts.However,observing sleeping sites at night is challenging,especially for species sensitive to human disturbance.Thermal infrared imaging(TIR)with drones is increasingly used for detecting and counting primates,yet it has not been utilized to investigate ecological strategies.This study investigates the sleeping site selection of the Critically Endangered black-shanked douc langur(Pygathrix nigripes)in Cát Tiên National Park,Vietnam.Our aim is to assess the feasibility of using a TIR drone to test sleeping site selection strategies in non-nesting primates,specifically examining hypotheses related to predation avoidance and food proximity.Between January and April 2023,we conducted 120 drone flights along 22 transects(~1-km long)and identified 114 sleeping sites via thermal imaging.We established 116 forest structure plots along 29 transects in non-selected sites and 65 plots within douc langur sleeping sites.Our observations reveal that douc langurs selected tall and large trees that may provide protection against predators.Additionally,they selected sleeping sites with increased access to food,such as Afzelia xylocarpa,which serves as a preferred food source during the dry season.These results highlight the effective use of TIR drones for studying douc langur sleeping site selection with minimal disturbance.Besides offering valuable insights into habitat selection and behavioral ecology for conservation,TIR drones hold great promise for the noninvasive and long-term monitoring of large-bodied arboreal species.
基金supported in part by the National Natural Science Foundation of China(grantNo.52362048)in part by Yunnan Fundamental Research Projects(grantNo.202301BE070001-042 and grant No.202401AT070409).
文摘Utilizing unmanned aerial vehicle (UAV) photography to timely detect and evaluate potential safety hazards (PSHs) around high-speed rail has great potential to complement and reform the existing manual inspections by providing better overhead views and mitigating safety issues. However, UAV inspections based on manual interpretation, which heavily rely on the experience, attention, and judgment of human inspectors, still inevitably suffer from subjectivity and inaccuracy. To address this issue, this study proposes a lightweight hybrid learning algorithm named HDTA (hybrid dual tasks architecture) to automatically and efficiently detect the PSHs of UAV imagery. First, this HDTA architecture seamlessly integrates both detection and segmentation branches within a unified framework. This design enables the model to simultaneously perform PSH detection and railroad parsing, thereby providing comprehensive scene understanding. Such joint learning also lays the foundation for PSH assessment tasks. Second, an innovative lightweight backbone based on the shuffle selective state space model (S^(4)M) is incorporated into HDTA. The state space model approach allows for global contextual information extraction while maintaining linear computational complexity. Furthermore, the incorporation of shuffle operation facilitates more efficient information flow across feature dimensions, enhancing both feature representation and fusion capabilities. Finally, extensive experiments conducted on a railroad environment dataset constructed from UAV imagery demonstrate that the proposed method achieves high detection accuracy while maintaining efficiency and practicality.
文摘The forest ecosystems of the Lacs 2 commune (South-East, Togo) are undergoing severe degradation, which has not yet been documented. This study is carried out in order to assess and quantify the spatio-temporal dynamics of residual forests and to identify the determinants of deforestation in South East Togo. The methodological approach is based on the use of historical aerial photographs from 1976 and drone images from 2019 in addition to field investigations. Several spatial structure indices were also calculated in order to quantify the fragmentation of classes and of the forest landscape. The results show that the forest landscape is changing. The classes of forests, plantations and palm groves show an annual rate of decline of 7.5%, 0.8% and 9.4% respectively while the classes of savannahs, agglomerations, surface water and swamps increased by 16.4%, 0.4%, 0.7% and 0.1%. The results also reveal a high fragmentation within the forest, plantation, surface water and swamp class and moderate fragmentation for the savannah and palm trees classes. At the landscape level, the savannah class is dominant by more than 70%, thus making the landscape little diversified from an ecological point of view. The main driver of deforestation in the study area remains shifting slash-and-burn agriculture. It is accentuated by the establishment of perennial oil palm crops, which has influenced the annual deforestation rate by 0.72%.
