U
Uwe Stilla
Researcher at Technische Universität München
Publications - 465
Citations - 9432
Uwe Stilla is an academic researcher from Technische Universität München. The author has contributed to research in topics: Point cloud & Synthetic aperture radar. The author has an hindex of 39, co-authored 447 publications receiving 7591 citations. Previous affiliations of Uwe Stilla include Karlsruhe University of Applied Sciences & AGH University of Science and Technology.
Papers
More filters
Journal ArticleDOI
Deep Learning Earth Observation Classification Using ImageNet Pretrained Networks
TL;DR: This letter proposes a novel method by considering a pretrained CNN designed for tackling an entirely different classification problem, namely, the ImageNet challenge, and exploit it to extract an initial set of representations, which are then transferred into a supervised CNN classifier, along with their class labels, effectively training the system.
Journal ArticleDOI
Classification With an Edge: Improving Semantic Image Segmentation with Boundary Detection
Dimitrios Marmanis,Dimitrios Marmanis,Konrad Schindler,Jan Dirk Wegner,Silvano Galliani,Mihai Datcu,Uwe Stilla +6 more
TL;DR: This work constructs a comparatively simple, memory-efficient model by adding boundary detection to the Segnet encoder-decoder architecture, and includes boundary detection in FCN-type models and sets up a high-end classifier ensemble, showing that boundary detection significantly improves semantic segmentation with CNNs.
Journal ArticleDOI
3D segmentation of single trees exploiting full waveform LIDAR data
TL;DR: In this paper, a 3D segmentation approach for single trees from LIDAR data is presented, which combines watershed segmentation and normalized cut segmentation to detect small trees in the lower forest layer.
Posted Content
Classification With an Edge: Improving Semantic Image Segmentation with Boundary Detection
Dimitrios Marmanis,Dimitrios Marmanis,Konrad Schindler,Jan Dirk Wegner,Silvano Galliani,Mihai Datcu,Uwe Stilla +6 more
TL;DR: In this article, the authors propose to combine semantic segmentation with semantically informed edge detection, thus making class-boundaries explicit in the model, and achieve state-of-the-art performance on the ISPRS Vaihingen benchmark.
Journal ArticleDOI
Analysis of full waveform LIDAR data for the classification of deciduous and coniferous trees
TL;DR: In this article, a methodology for tree species classification using features that are derived from small-footprint full waveform Light Detection and Ranging (LIDAR) data is described, and tree crowns are delineated from the canopy height model (CHM) using the watershed algorithm.