3d Reconstruction of a Tree Stem Using Video Images and Pulse Distances

This paper demonstrates how a 3D tree stem model can be reconstructed using video imagery combined with laser pulse distance measurements. Perspective projection is used to place the data collected with the portable video laser-rangefinding device into a real world coordinate system. This hybrid methodology uses a relatively small number of range measurements (compared to laser scanner instruments) and convergent videogrammetry (rather than the traditional stereo approach) along with stem model assumptions to create the stem reconstruction. Multiple video frames are stitched together to increase the accuracy of and to validate the orientation parameters from the multi-sensor device’s inclinometers. The 2D representation of the stem from the image data at a single viewpoint is combined with the additional range data to formulate a 2.5D model. Combining multiple vantage points aids in completing the 3D model and filling the gaps where data are absent.