Aspen detection in boreal forests: Capturing a key component of biodiversity using airborne hyperspectral, lidar, and UAV data

• DOI: 10.5194/egusphere-egu2020-21268
• Publication Date: 2020-03-10T05:50:03Z
• Authors: Timo Kumpula; Arto Viinikka; Janne Mäyrä; Anton Kuzmin; Pekka Hurskainen; Topi Tanhuanpää; Sonja Kivinen; peter Kullberg; Laura Poikolainen; Pasi Korpelainen; Max Stranden; Aleksi Ritakallio; Petteri Vihervaara
• Keywords: Scots pine

<p>Importance of biodiversity is increasingly highlighted as an essential part sustainable forest management. As direct monitoring not possible, proxy variables have been used to indicate site’s species richness and quality. In boreal forests, European aspen (Populus tremula L.) one the most significant proxies for biodiversity. Aspen a keystone species, hosting range endangered hence having high importance in maintaining Still, reliable fine-scale spatial data on occurrence remains scarce incomprehensive. Although remote sensing-based classification has decades needs forestry, commercially less (e.g., aspen) typically excluded from studies. This creates need developing general methods tree covering also ecologically species.</p><p> </p><p>Our study area, located Evo, Southern Finland, covers approximately 83km<sup>2</sup>, contains both managed protected southern forests. The main area are Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies (L.) Karst), birch (Betula pendula pubescens with relatively sparse scattered aspen. Along thorough field data, airborne hyperspectral LiDAR acquired area. We collected ultra resolution unmanned aerial vehicle (UAV) RGB multispectral sensors.</p><p> </p><p>Our aim gather fundamental classification, that can be utilized produce detailed at large scale. For this, we first analyze detection tree-level. test compare different machine learning (Support Vector Machines, Random Forest, Gradient Boosting Machine) deep (3D convolutional neural networks), specific emphasis accurate feasible detection. results will show, how accurately detected canopy, which bandwidths largest information satellite images scale.</p>