Expert in laser scanning and 3D modeling for roads, industry, maritime, and buildings. Enhancing your projects with precision and innovation.
Question for My laser scanning contacts: How do you identify tree species from your data? 🌳 🌴 🌲 We're encountering challenges in identifying tree species from laser scanning data. In Estonia, surveying regulations mandate the differentiation of trees by species. On drawings, each tree must be distinctly classified as either deciduous, spruce, pine, or another type of conifer. While we can quickly vectorize all our measurement data from point clouds, we hit a snag when it comes to tree identification, requiring us to handle everything manually. However, we cannot overlook this step, as topographic maps must have trees marked accordingly. Does anyone have recommendations or software suggestions that could help?
Artem Mazurok
7mo
We have partners who can identify tree species with over 87% accuracy using airborne laser scanning data (I think, MLS will work too). We prepare scan data for them for tree-by-tree taxation. The requirements are a density of >700 points/m2 and a pre-trained model (manual measurements of 100-200 trees). Now their model is trained to identify 4 species - spruce, pine, birch, aspen. The training was performed at latitudes 61-65 degrees north. Pls message in direct, if it might be useful for you
Operations Manager at GeoTwin
7moA multilevel classification breakdown would be the best in my view to seperate different species of trees by looks- this in turn will separate different looking trees - this could be written in modelling software and base it on layers and height but Inregards to clearly identifying different species you would need an have a dense colour point cloud and a someone having with knowlage of trees looking at the point cloud This is just my opinion