Tree and Plant Modelling

Image-based plant reconstruction and modeling is an important aspect of digitalization. A digital representation (Digital Twin) of plants, trees, or complete plantations could simulate and manage important agricultural processes. Significant indices (e.g. leaf area, biomass, phenotypical parameters) could be collated automatically. Based on these, yield and growth could be estimated. Additionally, the modeling of planting in infrastructural areas could facilitate the maintenance and repair seriously.

In the Computer Vision and Graphics group, we address the acquisition, reconstruction and modeling of plants based on imagery. Hereby plants and trees are recorded (perspectively using robots and drones). Initially, plants are reconstructed with classic photogrammetrical approaches and reconstructs the images into 3D point clouds. Most challenges arise with the complex structure of trees and the referencing of point clouds from different recording times and growth stages. Therefore, we especially follow procedural modeling and specialized grammars from the biology to represent plants in combination with computer vision concepts.