Urban Tree-Crown Mapping in Troisdorf

Aerial images and airborne laser scanning offer a way to study urban vegetation beyond the trees documented in municipal registers. This interactive website presents research from Troisdorf, Germany, exploring how individual tree crowns can be identified and their three-dimensional structure described.

The website brings together two complementary parts. Tree detection presents the CATCH4D Upscaling experiment, which tested whether adding LiDAR-derived relative height improved detection compared with aerial imagery alone. Visitors can inspect individual crowns, compare model outputs and explore correct detections, missed crowns and false alarms. For this dataset and workflow, the tested height layer did not improve detection.

Crown volume presents the recorded results of Victoria Luise Pissulla’s master’s thesis, covering 70,438 automatically detected tree-like vegetation structures. Visitors can explore selected three-dimensional point clouds, maps and charts of crown height and geometric volume, alongside comparisons with the municipal tree register.

The website explains the methods and limitations and provides links to the accompanying notebooks, data, code and QGIS project for further investigation and reuse.

The original detection and crown-volume work was carried out by Victoria Luise Pissulla at Ruhr University Bochum. The subsequent height study and interactive presentation were developed by Zahra Chehreghan at ILS, supervised by Dr. Shaojuan Xu.